diff --git a/.gitignore b/.gitignore new file mode 100644 index 0000000..5b6a065 --- /dev/null +++ b/.gitignore @@ -0,0 +1,4 @@ +.Rproj.user +.Rhistory +.RData +.Ruserdata diff --git a/404.html b/404.html index f619acc..7202086 100644 --- a/404.html +++ b/404.html @@ -23,7 +23,7 @@
404 Page not found - 2021-03-07 + 2023-05-30 @@ -58,7 +58,6 @@

404 NOT FOUND

-
@@ -95,7 +94,6 @@

404 NOT FOUND

} - diff --git a/about-hugo/index.html b/about-hugo/index.html index b92672b..ce18953 100644 --- a/about-hugo/index.html +++ b/about-hugo/index.html @@ -108,41 +108,41 @@

config.toml

By default, your site title will be displayed at the top in Blackletter fonts if your web browser and operating system support Blackletter fonts. Essentially, for every English letter X in your site title, it is substituted with 𝔛, e.g. the Blackletter version of A is 𝔄 (𝔄). Below are the letters from A to Z:

𝔄𝔅ℭ𝔇𝔈𝔉𝔊ℌℑ𝔍𝔎𝔏𝔐𝔑𝔒𝔓𝔔ℜ𝔖𝔗𝔘𝔙𝔚𝔛𝔜ℨ

Alternatively, you can provide a banner image:

-
[params.banner]
-    src = "/path/to/banner.png"
-    alt = "alternative text on image"
-

By default, each summary block on the homepage contains the first 200 letters extracted from all paragraphs of an article. I find Hugo’s built in .Summary often unsatisfactory (e.g. it may contain headings and code blocks, which really should not go to the summary), so I wrote my own version. It works much better and the size of most summary blocks will be the same, unless certain articles are really short. The length 200 can be customized via params.summary_length.

+
[params.banner]
+    src = "/path/to/banner.png"
+    alt = "alternative text on image"
+

By default, each summary block on the homepage contains the first 200 letters extracted from all paragraphs of an article. I find Hugo’s built in .Summary often unsatisfactory (e.g. it may contain headings and code blocks, which really should not go to the summary), so I wrote my own version. It works much better and the size of most summary blocks will be the same, unless certain articles are really short. The length 200 can be customized via params.summary_length.

If you are not satisfied with the automatic summary, you can specify the description option in the (YAML) metadata of your Markdown document, e.g.,

-
title: "My Cool Post"
-description: "Please use this as the summary."
-

Each summary block may contain a thumbnail, which is the first image in an article if exists. You can override it by providing the thumbnail option in the meta data of your Markdown document, e.g.,

-
---
-title: "My Cool Post"
-thumbnail: "/url/of/the/image"
----
-

For each page, this theme adds an edit link to the top-right if a parameter github_edit is provided, so that your reader may easily help you edit a page and submit a pull request on Github.

+
title: "My Cool Post"
+description: "Please use this as the summary."
+

Each summary block may contain a thumbnail, which is the first image in an article if exists. You can override it by providing the thumbnail option in the meta data of your Markdown document, e.g.,

+
---
+title: "My Cool Post"
+thumbnail: "/url/of/the/image"
+---
+

For each page, this theme adds an edit link to the top-right if a parameter github_edit is provided, so that your reader may easily help you edit a page and submit a pull request on Github.

The page footer can be defined in .Params.footer, and the text is treated as Markdown. Below are some sample configurations:

-
[params]
-    summary_length = 200
-    github_edit = "https://github.com/yihui/hugo-xmag/edit/master/exampleSite/content/"
-    footer = "© [Yihui Xie](https://yihui.name) 2017"
-

There are a few phrases that you can “translate” (I didn’t use Hugo’s multi-language feature just because I’m lazy):

-
[params.text]
-    about_author = "About the Author"
-    author_delimiter = ", "
-    back = "Back to Home"
-    edit = "Edit this page"
-    tags = "Tags: "
-    truncated = " …"
-    uncategorized = "Uncategorized"
-

You can define a data file under data/ to store all authors information, e.g., you can use a TOML file data/authors.toml (or YAML/JSON):

-
"Alice Wonder" = "I'm Alice. More about me on [my homepage](http://example.com)."
-"Yihui Xie" = "Hey this is Yihui. You don't want to follow me on Twitter @xieyihui."
+
[params]
+    summary_length = 200
+    github_edit = "https://github.com/yihui/hugo-xmag/edit/master/exampleSite/content/"
+    footer = "© [Yihui Xie](https://yihui.name) 2017"
+
There are a few phrases that you can “translate” (I didn’t use Hugo’s multi-language feature just because I’m lazy):

+
[params.text]
+    about_author = "About the Author"
+    author_delimiter = ", "
+    back = "Back to Home"
+    edit = "Edit this page"
+    tags = "Tags: "
+    truncated = " …"
+    uncategorized = "Uncategorized"
+
You can define a data file under data/ to store all authors information, e.g., you can use a TOML file data/authors.toml (or YAML/JSON):

+
"Alice Wonder" = "I'm Alice. More about me on [my homepage](http://example.com)."
+"Yihui Xie" = "Hey this is Yihui. You don't want to follow me on Twitter @xieyihui."
 
Then for an article with an author name that can be found in data/authors.toml, the author info will be added to the bottom of a page. For example, on this page, you can find information about me. You can change the phrase “About the Author” by changing the parameter about_author in config.toml.

 
To add a table of contents to an article, you can add toc: true to the YAML metadata of the Markdown document.

 
Custom layouts

 
Besides the custom layout^[If this is the first time you have heard about “customizing layouts”, please read the Hugo documentation first: https://gohugo.io/themes/customizing/.] files head_custom.html and foot_custom.html supported in XMin (see documentation), this theme added a few more layout files such as banner.html, comments.html and info.html under layouts/partials/. The first can be used to customize the banner. The second can be used to add a comment section, e.g., if you want to use Hugo’s default Disqus template, just add this to comments.html:

-
{{ template "_internal/disqus.html" . }}
+
{{ template "_internal/disqus.html" . }}
 
You can also append arbitrary text to each article through info.html. For example, you may declare copyrights or briefly introduce your site.

 
There are other partial templates in this theme and I encourage you to read the source code to figure out what they do.

 
Final words

@@ -195,10 +195,9 @@ 
Final words

 }
 
 
-
   

 
 
-
 
 
 
diff --git a/about-me/index.html b/about-me/index.html
index 68f181e..8bef16a 100644
--- a/about-me/index.html
+++ b/about-me/index.html
@@ -105,10 +105,9 @@ 
Yang Liu

 }
 
 
-
   

 
 
-
 
 
 
diff --git a/categories/catalog/index.html b/categories/catalog/index.html
index f69f560..3e54eab 100644
--- a/categories/catalog/index.html
+++ b/categories/catalog/index.html
@@ -23,7 +23,7 @@
       

       
         Catalog
-        2021-03-07
+        2023-05-30
         
         
         
@@ -72,9 +72,7 @@ 
Catalog of my old blog<
 
 
 
-
+
 

 
 
@@ -96,7 +94,6 @@ 
Catalog of my old blog<
 
 
 
-
 

   
   
@@ -133,7 +130,6 @@ 
Catalog of my old blog<
 }
 
 
-
 
 
 
diff --git a/categories/catalog/index.xml b/categories/catalog/index.xml
index 2346ec5..a83bca6 100644
--- a/categories/catalog/index.xml
+++ b/categories/catalog/index.xml
@@ -14,7 +14,7 @@
       
       https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/
       Introducing my new blog written solely in R Markdown
When I realized it was so convenient to write blog directly using R Markdown, I searched if there is a specific tool for it. And I found Blogdown, an R package developed by Yihui Xie, who also developed R Markdown.
-
Catalog
As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018.
+Catalog
As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018.
     
     
   
diff --git a/categories/catalog/page/1/index.html b/categories/catalog/page/1/index.html
index c58e182..38f159e 100644
--- a/categories/catalog/page/1/index.html
+++ b/categories/catalog/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/catalog/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/catalog/
+    
+    
+    
+    
+  
+
diff --git a/categories/data-engineering/index.html b/categories/data-engineering/index.html
index bb96269..eb9520c 100644
--- a/categories/data-engineering/index.html
+++ b/categories/data-engineering/index.html
@@ -1,83 +1,81 @@
-
-
-  
-    
-    
-    
-    Data Engineering | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-

-

-
-
-
-
-

-  

-  
-  
-    Data Engineering
-  
-  
-  

-  
Working with 3D array as long-format data in R

-  

-    Yang Liu
- / 
-    2020-01-11
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
-    check.and.install.pkgs <- function(pkgs){
new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d"))
Happy New…
-    
-  

-

-
-

-
-

-
-
-
-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/data-engineering/index.xml b/categories/data-engineering/index.xml
index 12e80f3..e0c352f 100644
--- a/categories/data-engineering/index.xml
+++ b/categories/data-engineering/index.xml
@@ -13,9 +13,9 @@
       Sat, 11 Jan 2020 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/
-      Example: a 3-D array with dimension 4x3x2
Melt into long
Recover the array


check.and.install.pkgs &lt;- function(pkgs){
new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;))
Happy New Year!
+      Example: a 3-D array with dimension 4x3x2
Melt into long
Recover the array
check.and.install.pkgs &lt;- function(pkgs){
new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;))
Happy New Year!
 Recently I spent some time working with array in R.
-I believe it is a bad idea to work with array using for loop, which is both slow and error-prone.
+I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. We can just melt it into a long data, do the work, and arrange back into array in the end if needed.
     
     
   
diff --git a/categories/data-engineering/page/1/index.html b/categories/data-engineering/page/1/index.html
index 882f69d..5fb5b0a 100644
--- a/categories/data-engineering/page/1/index.html
+++ b/categories/data-engineering/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/data-engineering/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/data-engineering/
+    
+    
+    
+    
+  
+
diff --git a/categories/data-visualization/index.html b/categories/data-visualization/index.html
index ebf8848..ee7c601 100644
--- a/categories/data-visualization/index.html
+++ b/categories/data-visualization/index.html
@@ -1,272 +1,265 @@
-
-
-  
-    
-    
-    
-    Data Visualization | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-

-

-
-
-
-
-

-  

-  
-  
-    Data Visualization
-  
-  
-  

-  
US and China Admin1 COVID19 mortality and incidence rate

-  

-    Yang Liu
- / 
-    2020-06-14
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
+
+
+  
+    
+    
+    
+    Data Visualization | Welcome to my blog
+    
+    
+    
+
+  
+
+  
+    
+
+
+

+

+
+
+
+
+

+  

+  
+  
+    Data Visualization
+  
+  
+  

+  
Plot multiple countries on the world map

+  

+    Yang Liu
+ / 
+    2023-05-29
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
+    This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+[…] (You may download the world map from Github)
+suppressPackageStartupMessages({
library("data.table")
library("ggplot2")…
+    
+  

+

+
+

+  

+  
+  
+    Data Visualization
+  
+  
+  

+  
US and China Admin1 COVID19 mortality and incidence rate

+  

+    Yang Liu
+ / 
+    2020-06-14
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
     Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
 Code hosted on my Github repo.
-[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here("Code/COVID_make_map.R"))
# US
dt_JUH_US…
-    
-  

-

-
-

-  

-  
-  
-    Data Visualization
-  
-  
-  

-  
ggplot US state and China province heatmap

-  

-    Yang Liu
- / 
-    2020-06-12
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
+[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here("Code/COVID_make_map.R"))
# US
dt_JUH_US…
+    
+  

+

+
+

+  

+  
+  
+    Data Visualization
+  
+  
+  

+  
ggplot US state and China province heatmap

+  

+    Yang Liu
+ / 
+    2020-06-12
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
     (Updated on 2020-06-12, First posted on 2019-04-17)
 It sounds easy but turned out not as straight-forward as I thought.
I will show:
 […] 50-state (including Alaska and Hawaii) United States thematic map, with map scale and with state abbreviations
 […] China thematic map, with map scale, with names of provinces in either English or Chinese
-[…] The dt1 below can be used by both…
-    
-  

-

-
-

-  

-  
-  
-    R package
-  
-  
-  

-  
SHAP for XGBoost in R: SHAPforxgboost

-  

-    Yang Liu
- / 
-    2019-07-18
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
+[…] The dt1 below can be used by both…
+    
+  

+

+
+

+  

+  
+  
+    R package
+  
+  
+  

+  
SHAP for XGBoost in R: SHAPforxgboost

+  

+    Yang Liu
+ / 
+    2019-07-18
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
     I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
-    
-  

-

-
-

-  

-  
-  
-    Data Visualization
-  
-  
-  

-  
Shiny in Blogdown

-  

-    Yang Liu
- / 
-    2019-02-24
-  

-  

-    
-    
-    
-    
-    
-    
-      
-    
+install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
+    
+  

+

+
+

+  

+  
+  
+    Data Visualization
+  
+  
+  

+  
Shiny in Blogdown

+  

+    Yang Liu
+ / 
+    2019-02-24
+  

+  

+    
+    
+    
+    
+    
+    
+      
+    
     Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
 The only way to do it is using iframe and write outside the chunk:
 <iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"></iframe>
-There is also a built-in function in knitr to do the same thing and write in the chunk. The only…
-    
-  

-

-
-

-  

-  
-  
-    Machine Learning
-  
-  
-  

-  
SHAP Visualization in R (first post)

-  

-    Yang Liu
- / 
-    2018-10-14
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
+There is also a built-in function in knitr to do the same thing and write in the chunk. The only…
+    
+  

+

+
+

+  

+  
+  
+    Machine Learning
+  
+  
+  

+  
SHAP Visualization in R (first post)

+  

+    Yang Liu
+ / 
+    2018-10-14
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
     Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
 […] This is the example I used in the package SHAPforxgboost
-# Example use iris
suppressPackageStartupMessages({…
-    
-  

-

-
-

-  

-  
-  
-    Data Visualization
-  
-  
-  

-  
Spatial data in R: Dividing raster layers into equal-area rings

-  

-    Yang Liu
- / 
-    2018-07-20
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
-    This data visualization example include:
-* Import .img file as a raster
-* Turn raster into a data.frame of points (coordinates) and values
-* Dividing the points into 100 equal-area rings
-* Calculate Built-up Area/Urban Extent for each ring
-* Turn dataframe into raster
-* Plot multiple figures on the same color scale
-[…] […] list.of.packages <- c("raster",…
-    
-  

-

-
-

+# Example use iris
suppressPackageStartupMessages({…
+    
+  

+

+
+

 
-

-
-
-
-

+
+
+
+

 
 
+
+
+
+
+

+  
+  
+  
@2021 Yang Liu github.com/liuyanguu. Powered by Github.

+  
+

+

+
+
 
 
 
-
-

-  
-  
-  
@2021 Yang Liu github.com/liuyanguu. Powered by Github.

-  
-

-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/data-visualization/index.xml b/categories/data-visualization/index.xml
index 352822f..5308a15 100644
--- a/categories/data-visualization/index.xml
+++ b/categories/data-visualization/index.xml
@@ -6,14 +6,25 @@
     Recent content in Data Visualization on Welcome to my blog
     Hugo -- gohugo.io
     en-us
-    Sun, 14 Jun 2020 00:00:00 +0000
+    Mon, 29 May 2023 00:00:00 +0000
+    
+      Plot multiple countries on the world map
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      Mon, 29 May 2023 00:00:00 +0000
+      
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+The most straightforward way
(You may download the world map from Github)
+suppressPackageStartupMessages({
library(&quot;data.table&quot;)
library(&quot;ggplot2&quot;)
library(&quot;rgdal&quot;)
library(&quot;raster&quot;)
library(&quot;rgeos&quot;)
library(&quot;here&quot;)
library(&quot;ggthemes&quot;)
})
# download data from GADM directly and row-bind the spatial polygons data frame
cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;)
download.
+    
+    
     
       US and China Admin1 COVID19 mortality and incidence rate
       https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/
       Sun, 14 Jun 2020 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/
-      US
China
Compared to lung cancer mortality


Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
+      US
China
Compared to lung cancer mortality
Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
 Code hosted on my Github repo.
 US
# major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here(&quot;Code/COVID_make_map.R&quot;))
# US
dt_JUH_US &lt;- get.JHU.us.state()
make_heatmap(data = dt_JUH_US, geo_data = get_state_name(),
state_var = &quot;Province_State&quot;, fill_var = &quot;Mortality_Rate&quot;, label_var = &quot;abb&quot;)
us_maps &lt;- lapply(c(&quot;Mortality_Rate&quot;, &quot;Incident_Rate&quot;, &quot;Testing_Rate&quot;, &quot;Hospitalization_Rate&quot;),
make_heatmap, data = dt_JUH_US, geo_data = get_state_name(),
state_var = &quot;Province_State&quot;, label_var = &quot;abb&quot;)
plot_grid &lt;- gridExtra::grid.
     
@@ -33,9 +44,9 @@ US
# major function, can download from Github repo Blogdown/hugo-xmag/Code
sourc
       Thu, 18 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/
-      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance

SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot

References


The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
+      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance
SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot
References
The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
+install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
     
     
     
@@ -44,10 +55,10 @@ install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_gith
       Sun, 24 Feb 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/
-      How to embed Shiny
My Shiny app example


How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
+      How to embed Shiny
My Shiny app example
How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
 The only way to do it is using iframe and write outside the chunk:
 &lt;iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"&gt;&lt;/iframe&gt;
-There is also a built-in function in knitr to do the same thing and write in the chunk.
+There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it.
     
     
     
@@ -56,7 +67,7 @@ There is also a built-in function in knitr to do the same thing and write in the
       Sun, 14 Oct 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/
-      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot

Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups



Update 19/07/21:
+      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot
Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups
Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
     
     
@@ -66,7 +77,7 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post
       Fri, 20 Jul 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/
-      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code


This data visualization example include:
+      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code
This data visualization example include:
 * Import .img file as a raster
 * Turn raster into a data.frame of points (coordinates) and values
 * Dividing the points into 100 equal-area rings
@@ -80,10 +91,10 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post
       Mon, 16 Jul 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/
-      Data
Version 1: Color both the branches and labels
Version 2: color only the labels.
Version 3: If there is no color, and we do not reorder the branches


This data visualization example include:
+      Data
Version 1: Color both the branches and labels
Version 2: color only the labels.
Version 3: If there is no color, and we do not reorder the branches
This data visualization example include:
 * Hierarchical clustering, dendrogram and heat map based on normalized odds ratios
 * The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend
-* Heatmap is made by heatmap.
+* Heatmap is made by heatmap.2 from gplots using the built dendrogram
     
     
   
diff --git a/categories/data-visualization/page/1/index.html b/categories/data-visualization/page/1/index.html
index 97985ba..c5e921b 100644
--- a/categories/data-visualization/page/1/index.html
+++ b/categories/data-visualization/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/data-visualization/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/data-visualization/
+    
+    
+    
+    
+  
+
diff --git a/categories/data-visualization/page/2/index.html b/categories/data-visualization/page/2/index.html
index 252ba5a..c90a974 100644
--- a/categories/data-visualization/page/2/index.html
+++ b/categories/data-visualization/page/2/index.html
@@ -23,7 +23,7 @@
       

       
         Data Visualization
-        2021-03-07
+        2023-05-30
         
         
         
@@ -39,6 +39,43 @@
 
 
 
+

+  

+  
+  
+    Data Visualization
+  
+  
+  

+  
Spatial data in R: Dividing raster layers into equal-area rings

+  

+    Yang Liu
+ / 
+    2018-07-20
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
+    This data visualization example include:
+* Import .img file as a raster
+* Turn raster into a data.frame of points (coordinates) and values
+* Dividing the points into 100 equal-area rings
+* Calculate Built-up Area/Urban Extent for each ring
+* Turn dataframe into raster
+* Plot multiple figures on the same color scale
+[…] […] list.of.packages <- c("raster",…
+    
+  

+

+
 

   

   
@@ -75,28 +112,26 @@ 
How
 
 
 
+    
 

 
 
@@ -118,7 +153,6 @@ 
How
 
 
 
-
 

   
   
@@ -155,7 +189,6 @@ 
How
 }
 
 
-
 
 
 
diff --git a/categories/github/index.html b/categories/github/index.html
index 27a42de..75b5e6d 100644
--- a/categories/github/index.html
+++ b/categories/github/index.html
@@ -1,85 +1,83 @@
-
-
-  
-    
-    
-    
-    Github | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-

-

-
-
-
-
-

-  

-  
-  
-    Github
-  
-  
-  

-  
RMarkdown to Github Pages

-  

-    Yang Liu
- / 
-    2021-01-06
-  

-  

-    
-    
-    
-    
-    
Thumbnail

-    
-    
-    
-      
-    
+
+
+  
+    
+    
+    
+    Github | Welcome to my blog
+    
+    
+    
+
+  
+
+  
+    
+
+
+

+

+
+
+
+
+

+  

+  
+  
+    Github
+  
+  
+  

+  
RMarkdown to Github Pages

+  

+    Yang Liu
+ / 
+    2021-01-06
+  

+  

+    
+    
+    
+    
+    
Thumbnail

+    
+    
+    
+      
+    
     Example of one of my Github pages:
You might have read this GitHub and RStudio tutorial by searching this topic. It is quite long and confusing as it tries to teach Git at the same time. You don’t need any of those branch operations. And the example yaml code is not indented correctly (corrected below).
 Here is the short version:
-[…] Make sure the yaml format is indented correct. Most…
-    
-  

-

-
-

-
-

-
-
-
-

+[…] Make sure the yaml format is indented correct. Most…
+    
+  

+

+
+

+
+

+
+
+
+

 
 
+
+
+
+
+

+  
+  
+  
@2021 Yang Liu github.com/liuyanguu. Powered by Github.

+  
+

+

+
+
 
 
 
-
-

-  
-  
-  
@2021 Yang Liu github.com/liuyanguu. Powered by Github.

-  
-

-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/github/page/1/index.html b/categories/github/page/1/index.html
index d86807f..53a3132 100644
--- a/categories/github/page/1/index.html
+++ b/categories/github/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/github/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/github/
+    
+    
+    
+    
+  
+
diff --git a/categories/index.html b/categories/index.html
index 8c200f4..f8e8999 100644
--- a/categories/index.html
+++ b/categories/index.html
@@ -23,7 +23,7 @@
       

       
         Categories
-        2021-03-07
+        2023-05-30
         
         
         
@@ -58,7 +58,7 @@ 
Categories

     
       data-visualization
     
-    (7)
+    (8)
   
   
   
  • 
@@ -72,7 +72,7 @@ 
    Categories
    
     
       machine-learning
     
-    (4)
+    (3)
   
    • 
   
   
  • 
@@ -100,7 +100,7 @@ 
    Categories
    
     
       spatial-data
     
-    (1)
+    (2)
   
    • 
   
 
@@ -127,7 +127,6 @@ 
    Categories
    
 
 
 
-
 
    
   
   
@@ -164,7 +163,6 @@ 
    Categories
    
 }
 
 
-
 
 
 
diff --git a/categories/index.xml b/categories/index.xml
index 16987d8..128f667 100644
--- a/categories/index.xml
+++ b/categories/index.xml
@@ -6,22 +6,31 @@
     Recent content in Categories on Welcome to my blog
     Hugo -- gohugo.io
     en-us
-    Wed, 06 Jan 2021 00:00:00 +0000
+    Mon, 29 May 2023 00:00:00 +0000
     
-      Github
-      https://liuyanguu.github.io/categories/github/
-      Wed, 06 Jan 2021 00:00:00 +0000
+      Data Visualization
+      https://liuyanguu.github.io/categories/data-visualization/
+      Mon, 29 May 2023 00:00:00 +0000
       
-      https://liuyanguu.github.io/categories/github/
+      https://liuyanguu.github.io/categories/data-visualization/
       
     
     
     
-      Data Visualization
-      https://liuyanguu.github.io/categories/data-visualization/
-      Sun, 14 Jun 2020 00:00:00 +0000
+      Spatial Data
+      https://liuyanguu.github.io/categories/spatial-data/
+      Mon, 29 May 2023 00:00:00 +0000
       
-      https://liuyanguu.github.io/categories/data-visualization/
+      https://liuyanguu.github.io/categories/spatial-data/
+      
+    
+    
+    
+      Github
+      https://liuyanguu.github.io/categories/github/
+      Wed, 06 Jan 2021 00:00:00 +0000
+      
+      https://liuyanguu.github.io/categories/github/
       
     
     
@@ -70,15 +79,6 @@
       
     
     
-    
-      Spatial Data
-      https://liuyanguu.github.io/categories/spatial-data/
-      Fri, 20 Jul 2018 00:00:00 +0000
-      
-      https://liuyanguu.github.io/categories/spatial-data/
-      
-    
-    
     
       Catalog
       https://liuyanguu.github.io/categories/catalog/
diff --git a/categories/machine-learning/index.html b/categories/machine-learning/index.html
index 75e6ad3..8d7d907 100644
--- a/categories/machine-learning/index.html
+++ b/categories/machine-learning/index.html
@@ -1,177 +1,145 @@
-
-
-  
-    
-    
-    
-    Machine Learning | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-
    
-
    
-
-
-
-
-
    
-  
    
-  
-  
-    R package
-  
-  
-  
    
-  
    SHAP for XGBoost in R: SHAPforxgboost
    
-  
    
-    Yang Liu
- / 
-    2019-07-18
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+
+
+  
+    
+    
+    
+    Machine Learning | Welcome to my blog
+    
+    
+    
+
+  
+
+  
+        
+
+
+
    
+
    
+
+
+
+
+
    
+  
    
+  
+  
+    R package
+  
+  
+  
    
+  
    SHAP for XGBoost in R: SHAPforxgboost
    
+  
    
+    Yang Liu
+ / 
+    2019-07-18
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Machine Learning
-  
-  
-  
    
-  
    Study shrinkage and DART in xgboost modeling using a simple dataset
    
-  
    
-    Yang Liu
- / 
-    2018-11-15
-  
    
-  
    
-    
-    
-    
-    
-    
-    
-      
-    
+install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
+    
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Machine Learning
+  
+  
+  
    
+  
    Study shrinkage and DART in xgboost modeling using a simple dataset
    
+  
    
+    Yang Liu
+ / 
+    2018-11-15
+  
    
+  
    
+    
+    
+    
+    
+    
+    
+      
+    
     It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
-[…] set.seed(123)
n0 <- 5
X <- data.frame(x1 = runif(n0), x2 = runif(n0))
Y <- c(1, 5, 20, 50, 100)
cbind(X,…
-    
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Machine Learning
-  
-  
-  
    
-  
    SHAP Visualization in R (first post)
    
-  
    
-    Yang Liu
- / 
-    2018-10-14
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…] set.seed(123)
n0 <- 5
X <- data.frame(x1 = runif(n0), x2 = runif(n0))
Y <- c(1, 5, 20, 50, 100)
cbind(X,…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Machine Learning
+  
+  
+  
    
+  
    SHAP Visualization in R (first post)
    
+  
    
+    Yang Liu
+ / 
+    2018-10-14
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
 […] This is the example I used in the package SHAPforxgboost
-# Example use iris
suppressPackageStartupMessages({…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Machine Learning
-  
-  
-  
    
-  
    eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting
    
-  
    
-    Yang Liu
- / 
-    2018-07-09
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
-    I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset.
The ensemble method is powerful as it combines the predictions from multiple machine learning algorithms together to…
-        
-  
    
-
    
-
-
    
-
-
    
-
-
-
-
    
+# Example use iris
suppressPackageStartupMessages({…
+    
+  
    
+
    
+
+
    
+
+
    
+
+
+
+
    
 
 
+
+
+
+
+
    
+  
+  
+  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
+  
+
    
+
    
+
+
 
 
 
-
-
    
-  
-  
-  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
-  
-
    
-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/machine-learning/index.xml b/categories/machine-learning/index.xml
index 8f95e47..8779422 100644
--- a/categories/machine-learning/index.xml
+++ b/categories/machine-learning/index.xml
@@ -13,9 +13,9 @@
       Thu, 18 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/
-      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance

SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot

References


The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
+      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance
SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot
References
The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
+install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
         
     
     
@@ -24,8 +24,8 @@ install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_gith
       Thu, 15 Nov 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/
-      Data
Shrinkage
DART: Dropout - MART
skip_drop
rate_drop
one_drop



It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
-Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.
+      Data
Shrinkage
DART: Dropout - MART
skip_drop
rate_drop
one_drop
It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
+Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.frame(x1 = runif(n0), x2 = runif(n0))
Y &lt;- c(1, 5, 20, 50, 100)
cbind(X, Y)
## x1 x2 Y
## 1 0.
     
     
     
@@ -34,18 +34,9 @@ Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.
       Sun, 14 Oct 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/
-      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot

Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups



Update 19/07/21:
+      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot
Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups
Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
     
     
-    
-      eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting
-      https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/
-      Mon, 09 Jul 2018 00:00:00 +0000
-      
-      https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/
-      Overview
About the Data
1. Extreme Gradient Boosting
2. Gradient boosting
3. Random Forest
4. Lasso
5. Best Subset
Compare MSEs
Original Code are stored on github


Overview
I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset.
-    
-    
   
 
diff --git a/categories/machine-learning/page/1/index.html b/categories/machine-learning/page/1/index.html
index db090bd..b933a4f 100644
--- a/categories/machine-learning/page/1/index.html
+++ b/categories/machine-learning/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/machine-learning/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/machine-learning/
+    
+    
+    
+    
+  
+
diff --git a/categories/page/1/index.html b/categories/page/1/index.html
index a66c110..191a923 100644
--- a/categories/page/1/index.html
+++ b/categories/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/
+    
+    
+    
+    
+  
+
diff --git a/categories/page/2/index.html b/categories/page/2/index.html
index 358f935..16e1a76 100644
--- a/categories/page/2/index.html
+++ b/categories/page/2/index.html
@@ -23,7 +23,7 @@
       
    
       
         Categories
-        2021-03-07
+        2023-05-30
         
         
         
@@ -58,7 +58,7 @@ 
    Categories
    
     
       data-visualization
     
-    (7)
+    (8)
   
   
   
  • 
@@ -72,7 +72,7 @@ 
    Categories
    
     
       machine-learning
     
-    (4)
+    (3)
   
    • 
   
   
  • 
@@ -100,7 +100,7 @@ 
    Categories
    
     
       spatial-data
     
-    (1)
+    (2)
   
    • 
   
 
@@ -127,7 +127,6 @@ 
    Categories
    
 
 
 
-
 
    
   
   
@@ -164,7 +163,6 @@ 
    Categories
    
 }
 
 
-
 
 
 
diff --git a/categories/r-package/index.html b/categories/r-package/index.html
index b5ec1fb..a9772ad 100644
--- a/categories/r-package/index.html
+++ b/categories/r-package/index.html
@@ -1,118 +1,116 @@
-
-
-  
-    
-    
-    
-    R package | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-
    
-
    
-
-
-
-
-
    
-  
    
-  
-  
-    R package
-  
-  
-  
    
-  
    Notes on writing an R package
    
-  
    
-    Yang Liu
- / 
-    2019-07-28
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+
+
+  
+    
+    
+    
+    R package | Welcome to my blog
+    
+    
+    
+
+  
+
+  
+        
+
+
+
    
+
    
+
+
+
+
+
    
+  
    
+  
+  
+    R package
+  
+  
+  
    
+  
    Notes on writing an R package
    
+  
    
+    Yang Liu
+ / 
+    2019-07-28
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
 install.packages("SHAPforxgboost")
or
 devtools::install_github("liuyanguu/SHAPforxgboost")
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.
-[…]…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    R package
-  
-  
-  
    
-  
    SHAP for XGBoost in R: SHAPforxgboost
    
-  
    
-    Yang Liu
- / 
-    2019-07-18
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…]…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    R package
+  
+  
+  
    
+  
    SHAP for XGBoost in R: SHAPforxgboost
    
+  
    
+    Yang Liu
+ / 
+    2019-07-18
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
-        
-  
    
-
    
-
-
    
-
-
    
-
-
-
-
    
+install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
+    
+  
    
+
    
+
+
    
+
+
    
+
+
+
+
    
 
 
+
+
+
+
+
    
+  
+  
+  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
+  
+
    
+
    
+
+
 
 
 
-
-
    
-  
-  
-  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
-  
-
    
-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/r-package/index.xml b/categories/r-package/index.xml
index 528ca96..f343891 100644
--- a/categories/r-package/index.xml
+++ b/categories/r-package/index.xml
@@ -13,8 +13,9 @@
       Sun, 28 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/
-      Some of my own experience
On description
Namespace
load vs attach

Documentation
Potential problems when checking the package

Some nice suggestions from the CRAN team when submitting the package


Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
-install.packages(&quot;SHAPforxgboost&quot;)
or
+      Some of my own experience
On description
Namespace
load vs attach
Documentation
Potential problems when checking the package
Some nice suggestions from the CRAN team when submitting the package
Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
+install.packages(&quot;SHAPforxgboost&quot;)
or
+devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.
     
     
     
@@ -23,9 +24,9 @@ install.packages(&quot;SHAPforxgboost&quot;)
or
       Thu, 18 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/
-      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance

SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot

References


The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
+      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance
SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot
References
The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
+install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
     
     
   
diff --git a/categories/r-package/page/1/index.html b/categories/r-package/page/1/index.html
index 7785f41..67becbe 100644
--- a/categories/r-package/page/1/index.html
+++ b/categories/r-package/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/r-package/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/r-package/
+    
+    
+    
+    
+  
+
diff --git a/categories/reproducible-workflow/index.html b/categories/reproducible-workflow/index.html
index c1fa905..aa86279 100644
--- a/categories/reproducible-workflow/index.html
+++ b/categories/reproducible-workflow/index.html
@@ -23,7 +23,7 @@
       
    
       
         Reproducible Workflow
-        2021-03-07
+        2023-05-30
         
         
         
@@ -71,9 +71,7 @@ 
    
@@ -95,7 +93,6 @@ 
    
   
   
@@ -132,7 +129,6 @@ 
    
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/reproducible-workflow/
+    
+    
+    
+    
+  
+
diff --git a/categories/shiny/index.html b/categories/shiny/index.html
index 2dd11ba..0d02e08 100644
--- a/categories/shiny/index.html
+++ b/categories/shiny/index.html
@@ -23,7 +23,7 @@
       
    
       
         Shiny
-        2021-03-07
+        2023-05-30
         
         
         
@@ -71,9 +71,7 @@ 
    Shiny in Blogdown
    
 
    
 
 
-
+
 
    
 
 
@@ -95,7 +93,6 @@ 
    Shiny in Blogdown
    
 
 
 
-
 
    
   
   
@@ -132,7 +129,6 @@ 
    Shiny in Blogdown
    
 }
 
 
-
 
 
 
diff --git a/categories/shiny/index.xml b/categories/shiny/index.xml
index 5a10c7a..a91b2cd 100644
--- a/categories/shiny/index.xml
+++ b/categories/shiny/index.xml
@@ -13,10 +13,10 @@
       Sun, 24 Feb 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/
-      How to embed Shiny
My Shiny app example


How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
+      How to embed Shiny
My Shiny app example
How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
 The only way to do it is using iframe and write outside the chunk:
 &lt;iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"&gt;&lt;/iframe&gt;
-There is also a built-in function in knitr to do the same thing and write in the chunk.
+There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it.
     
     
   
diff --git a/categories/shiny/page/1/index.html b/categories/shiny/page/1/index.html
index 5cf36d6..66c39d0 100644
--- a/categories/shiny/page/1/index.html
+++ b/categories/shiny/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/shiny/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/shiny/
+    
+    
+    
+    
+  
+
diff --git a/categories/spatial-data/index.html b/categories/spatial-data/index.html
index f45eefb..d3124da 100644
--- a/categories/spatial-data/index.html
+++ b/categories/spatial-data/index.html
@@ -1,69 +1,101 @@
-
-
-  
-    
-    
-    
-    Spatial Data | Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-
    
-
    
-
-
-
-
-
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    Spatial data in R: Dividing raster layers into equal-area rings
    
-  
    
-    Yang Liu
- / 
-    2018-07-20
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+
+
+  
+    
+    
+    
+    Spatial Data | Welcome to my blog
+    
+    
+    
+
+  
+
+  
+        
+
+
+
    
+
    
+
+
+
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    Plot multiple countries on the world map
    
+  
    
+    Yang Liu
+ / 
+    2023-05-29
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
+    This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+[…] (You may download the world map from Github)
+suppressPackageStartupMessages({
library("data.table")
library("ggplot2")…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    Spatial data in R: Dividing raster layers into equal-area rings
    
+  
    
+    Yang Liu
+ / 
+    2018-07-20
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     This data visualization example include:
 * Import .img file as a raster
 * Turn raster into a data.frame of points (coordinates) and values
@@ -71,20 +103,18 @@ 
    
-
-
    
+
    
+
+
    
+
+
    
+
+
+
+
    
 
 
-
-
-
-
-
    
-  
-  
-  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
-  
-
    
-
    
-
-
-
-
-
+
+
+
+
+
    
+  
+  
+  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
+  
+
    
+
+
+
+
+
+
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/categories/spatial-data/index.xml b/categories/spatial-data/index.xml
index dcbb83f..ad6e259 100644
--- a/categories/spatial-data/index.xml
+++ b/categories/spatial-data/index.xml
@@ -6,14 +6,25 @@
     Recent content in Spatial Data on Welcome to my blog
     Hugo -- gohugo.io
     en-us
-    Fri, 20 Jul 2018 00:00:00 +0000
+    Mon, 29 May 2023 00:00:00 +0000
+    
+      Plot multiple countries on the world map
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      Mon, 29 May 2023 00:00:00 +0000
+      
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+The most straightforward way
(You may download the world map from Github)
+suppressPackageStartupMessages({
library(&quot;data.table&quot;)
library(&quot;ggplot2&quot;)
library(&quot;rgdal&quot;)
library(&quot;raster&quot;)
library(&quot;rgeos&quot;)
library(&quot;here&quot;)
library(&quot;ggthemes&quot;)
})
# download data from GADM directly and row-bind the spatial polygons data frame
cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;)
download.
+    
+    
     
       Spatial data in R: Dividing raster layers into equal-area rings
       https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/
       Fri, 20 Jul 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/
-      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code


This data visualization example include:
+      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code
This data visualization example include:
 * Import .img file as a raster
 * Turn raster into a data.frame of points (coordinates) and values
 * Dividing the points into 100 equal-area rings
diff --git a/categories/spatial-data/page/1/index.html b/categories/spatial-data/page/1/index.html
index c96a6fe..57bbe97 100644
--- a/categories/spatial-data/page/1/index.html
+++ b/categories/spatial-data/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/categories/spatial-data/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/categories/spatial-data/
+    
+    
+    
+    
+  
+
diff --git a/index.html b/index.html
index 08485f2..a385cb5 100644
--- a/index.html
+++ b/index.html
@@ -1,269 +1,266 @@
-
-
+
+
   
-	
-    
-    
-    
-    Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-
    
-
    
-
-
-
-
-
    
-  
    
-  
-  
-    Github
-  
-  
-  
    
-  
    RMarkdown to Github Pages
    
-  
    
-    Yang Liu
- / 
-    2021-01-06
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+	
+    
+    
+    
+    Welcome to my blog
+    
+    
+    
+
+  
+
+  
+        
+
+
+
    
+
    
+
+
+
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    Plot multiple countries on the world map
    
+  
    
+    Yang Liu
+ / 
+    2023-05-29
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
+    This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+[…] (You may download the world map from Github)
+suppressPackageStartupMessages({
library("data.table")
library("ggplot2")…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Github
+  
+  
+  
    
+  
    RMarkdown to Github Pages
    
+  
    
+    Yang Liu
+ / 
+    2021-01-06
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     Example of one of my Github pages:
You might have read this GitHub and RStudio tutorial by searching this topic. It is quite long and confusing as it tries to teach Git at the same time. You don’t need any of those branch operations. And the example yaml code is not indented correctly (corrected below).
 Here is the short version:
-[…] Make sure the yaml format is indented correct. Most…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    US and China Admin1 COVID19 mortality and incidence rate
    
-  
    
-    Yang Liu
- / 
-    2020-06-14
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…] Make sure the yaml format is indented correct. Most…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    US and China Admin1 COVID19 mortality and incidence rate
    
+  
    
+    Yang Liu
+ / 
+    2020-06-14
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
 Code hosted on my Github repo.
-[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here("Code/COVID_make_map.R"))
# US
dt_JUH_US…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    ggplot US state and China province heatmap
    
-  
    
-    Yang Liu
- / 
-    2020-06-12
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here("Code/COVID_make_map.R"))
# US
dt_JUH_US…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    ggplot US state and China province heatmap
    
+  
    
+    Yang Liu
+ / 
+    2020-06-12
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     (Updated on 2020-06-12, First posted on 2019-04-17)
 It sounds easy but turned out not as straight-forward as I thought.
I will show:
 […] 50-state (including Alaska and Hawaii) United States thematic map, with map scale and with state abbreviations
 […] China thematic map, with map scale, with names of provinces in either English or Chinese
-[…] The dt1 below can be used by both…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Data Engineering
-  
-  
-  
    
-  
    Working with 3D array as long-format data in R
    
-  
    
-    Yang Liu
- / 
-    2020-01-11
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
-    check.and.install.pkgs <- function(pkgs){
new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d"))
Happy New…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Reproducible Workflow
-  
-  
-  
    
-  
    Drake: powerful tool for automatic reproducible workflow
    
-  
    
-    Yang Liu
- / 
-    2019-09-15
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…] The dt1 below can be used by both…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Engineering
+  
+  
+  
    
+  
    Working with 3D array as long-format data in R
    
+  
    
+    Yang Liu
+ / 
+    2020-01-11
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
+    check.and.install.pkgs <- function(pkgs){
new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d"))
Happy New…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Reproducible Workflow
+  
+  
+  
    
+  
    Drake: powerful tool for automatic reproducible workflow
    
+  
    
+    Yang Liu
+ / 
+    2019-09-15
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     drake is a powerful tool for automatic reproducible workflow. I found it a perfect match when used together with RMarkdown. There are great documentations online for drake thus here I only show a simple example working with RMarkdown.
-RMarkdown file could be very slow to generate if lots of calculations are involved. Any small revise makes you rerun everything. When we use drake we can do all the…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    R package
-  
-  
-  
    
-  
    Notes on writing an R package
    
-  
    
-    Yang Liu
- / 
-    2019-07-28
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
-    Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
-install.packages("SHAPforxgboost")
or
-devtools::install_github("liuyanguu/SHAPforxgboost")
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.
-[…]…
-        
-  
    
-
    
-
-
    
+RMarkdown file could be very slow to generate if lots of calculations are involved. Any small revise makes you rerun everything. When we use drake we can do all the…
+    
+  
    
+
    
+
+
    
 
-
    
-
-
-
-
    
+
+
+
+
    
 
 
+
+
+
+
+
    
+  
+  
+  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
+  
+
    
+
    
+
+
 
 
 
-
-
    
-  
-  
-  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
-  
-
    
-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/index.xml b/index.xml
index 95a1f8f..b4b3cf9 100644
--- a/index.xml
+++ b/index.xml
@@ -6,7 +6,18 @@
     Recent content on Welcome to my blog
     Hugo -- gohugo.io
     en-us
-    Wed, 06 Jan 2021 00:00:00 +0000
+    Mon, 29 May 2023 00:00:00 +0000
+    
+      Plot multiple countries on the world map
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      Mon, 29 May 2023 00:00:00 +0000
+      
+      https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/
+      This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map.
I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!
+The most straightforward way
(You may download the world map from Github)
+suppressPackageStartupMessages({
library(&quot;data.table&quot;)
library(&quot;ggplot2&quot;)
library(&quot;rgdal&quot;)
library(&quot;raster&quot;)
library(&quot;rgeos&quot;)
library(&quot;here&quot;)
library(&quot;ggthemes&quot;)
})
# download data from GADM directly and row-bind the spatial polygons data frame
cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;)
download.
+    
+    
     
       RMarkdown to Github Pages
       https://liuyanguu.github.io/post/2021/01/06/rmarkdown-to-github-pages/
@@ -24,7 +35,7 @@ Create in the root directory a yaml file: "
       Sun, 14 Jun 2020 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/
-      US
China
Compared to lung cancer mortality


Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
+      US
China
Compared to lung cancer mortality
Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable.
 Code hosted on my Github repo.
 US
# major function, can download from Github repo Blogdown/hugo-xmag/Code
source(here::here(&quot;Code/COVID_make_map.R&quot;))
# US
dt_JUH_US &lt;- get.JHU.us.state()
make_heatmap(data = dt_JUH_US, geo_data = get_state_name(),
state_var = &quot;Province_State&quot;, fill_var = &quot;Mortality_Rate&quot;, label_var = &quot;abb&quot;)
us_maps &lt;- lapply(c(&quot;Mortality_Rate&quot;, &quot;Incident_Rate&quot;, &quot;Testing_Rate&quot;, &quot;Hospitalization_Rate&quot;),
make_heatmap, data = dt_JUH_US, geo_data = get_state_name(),
state_var = &quot;Province_State&quot;, label_var = &quot;abb&quot;)
plot_grid &lt;- gridExtra::grid.
     
@@ -44,9 +55,9 @@ US
# major function, can download from Github repo Blogdown/hugo-xmag/Code
sourc
       Sat, 11 Jan 2020 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/
-      Example: a 3-D array with dimension 4x3x2
Melt into long
Recover the array


check.and.install.pkgs &lt;- function(pkgs){
new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;))
Happy New Year!
+      Example: a 3-D array with dimension 4x3x2
Melt into long
Recover the array
check.and.install.pkgs &lt;- function(pkgs){
new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]]
if(length(new.packages)) install.packages(new.packages, dependencies = TRUE)
suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE)))
}
check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;))
Happy New Year!
 Recently I spent some time working with array in R.
-I believe it is a bad idea to work with array using for loop, which is both slow and error-prone.
+I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. We can just melt it into a long data, do the work, and arrange back into array in the end if needed.
     
     
     
@@ -65,8 +76,9 @@ RMarkdown file could be very slow to generate if lots of calculations are involv
       Sun, 28 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/
-      Some of my own experience
On description
Namespace
load vs attach

Documentation
Potential problems when checking the package

Some nice suggestions from the CRAN team when submitting the package


Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
-install.packages(&quot;SHAPforxgboost&quot;)
or
+      Some of my own experience
On description
Namespace
load vs attach
Documentation
Potential problems when checking the package
Some nice suggestions from the CRAN team when submitting the package
Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
+install.packages(&quot;SHAPforxgboost&quot;)
or
+devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.
     
     
     
@@ -75,9 +87,9 @@ install.packages(&quot;SHAPforxgboost&quot;)
or
       Thu, 18 Jul 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/
-      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance

SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot

References


The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
+      The SHAPforxgboost package
Why SHAP values
Local explanation
Consistency in global feature importance
SHAP plots
Summary plot
Dependence plot
Interaction effects
SHAP force plot
References
The SHAPforxgboost package
I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
+install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;)
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model structure.
     
     
     
@@ -86,10 +98,10 @@ install.packages(&quot;SHAPforxgboost&quot;)
# or devtools::install_gith
       Sun, 24 Feb 2019 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/
-      How to embed Shiny
My Shiny app example


How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
+      How to embed Shiny
My Shiny app example
How to embed Shiny
Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
 The only way to do it is using iframe and write outside the chunk:
 &lt;iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"&gt;&lt;/iframe&gt;
-There is also a built-in function in knitr to do the same thing and write in the chunk.
+There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it.
     
     
     
@@ -98,8 +110,8 @@ There is also a built-in function in knitr to do the same thing and write in the
       Thu, 15 Nov 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/
-      Data
Shrinkage
DART: Dropout - MART
skip_drop
rate_drop
one_drop



It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
-Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.
+      Data
Shrinkage
DART: Dropout - MART
skip_drop
rate_drop
one_drop
It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
+Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.frame(x1 = runif(n0), x2 = runif(n0))
Y &lt;- c(1, 5, 20, 50, 100)
cbind(X, Y)
## x1 x2 Y
## 1 0.
     
     
     
@@ -108,7 +120,7 @@ Data
set.seed(123)
n0 &lt;- 5
X &lt;- data.
       Sun, 14 Oct 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/
-      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot

Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups



Update 19/07/21:
+      Example 1
SHAP summary plot
Alternative ways:
SHAP dependence plot
SHAP interaction effect plot
SHAP force plot
Example 2
Summary plot
Dependence plot for each feature
Force plot
Stack plot by clustering groups
Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
     
     
@@ -118,7 +130,7 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post
       Fri, 20 Jul 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/
-      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code


This data visualization example include:
+      Saturation in ten cities with equal-area rings
R Code for one city
Results for the ring saturations
Average saturation in each ring
Methodology
Original Code
This data visualization example include:
 * Import .img file as a raster
 * Turn raster into a data.frame of points (coordinates) and values
 * Dividing the points into 100 equal-area rings
@@ -132,19 +144,10 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post
       Mon, 16 Jul 2018 00:00:00 +0000
       
       https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/
-      Data
Version 1: Color both the branches and labels
Version 2: color only the labels.
Version 3: If there is no color, and we do not reorder the branches


This data visualization example include:
+      Data
Version 1: Color both the branches and labels
Version 2: color only the labels.
Version 3: If there is no color, and we do not reorder the branches
This data visualization example include:
 * Hierarchical clustering, dendrogram and heat map based on normalized odds ratios
 * The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend
-* Heatmap is made by heatmap.
-    
-    
-    
-      eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting
-      https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/
-      Mon, 09 Jul 2018 00:00:00 +0000
-      
-      https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/
-      Overview
About the Data
1. Extreme Gradient Boosting
2. Gradient boosting
3. Random Forest
4. Lasso
5. Best Subset
Compare MSEs
Original Code are stored on github


Overview
I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset.
+* Heatmap is made by heatmap.2 from gplots using the built dendrogram
     
     
     
@@ -154,7 +157,7 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post
       
       https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/
       Introducing my new blog written solely in R Markdown
When I realized it was so convenient to write blog directly using R Markdown, I searched if there is a specific tool for it. And I found Blogdown, an R package developed by Yihui Xie, who also developed R Markdown.
-
Catalog
As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018.
+Catalog
As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018.
     
     
     
diff --git a/liuyanguu.github.io.Rproj b/liuyanguu.github.io.Rproj
new file mode 100644
index 0000000..e83436a
--- /dev/null
+++ b/liuyanguu.github.io.Rproj
@@ -0,0 +1,16 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: Yes
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX
+
+AutoAppendNewline: Yes
+StripTrailingWhitespace: Yes
diff --git a/page/1/index.html b/page/1/index.html
index f01d785..05fee2b 100644
--- a/page/1/index.html
+++ b/page/1/index.html
@@ -1 +1,10 @@
-https://liuyanguu.github.io/
\ No newline at end of file
+
+
+  
+    https://liuyanguu.github.io/
+    
+    
+    
+    
+  
+
diff --git a/page/2/index.html b/page/2/index.html
index 0b02dc3..db07454 100644
--- a/page/2/index.html
+++ b/page/2/index.html
@@ -1,195 +1,228 @@
-
-
+
+
   
-	
-    
-    
-    
-    Welcome to my blog
-    
-    
-    
-
-  
-
-  
-    
-
-
-
    
-
    
-
-
-
-
-
    
-  
    
-  
-  
-    R package
-  
-  
-  
    
-  
    SHAP for XGBoost in R: SHAPforxgboost
    
-  
    
-    Yang Liu
- / 
-    2019-07-18
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+	
+    
+    
+    
+    Welcome to my blog
+    
+    
+    
+
+  
+
+  
+        
+
+
+
    
+
    
+
+
+
+
+
    
+  
    
+  
+  
+    R package
+  
+  
+  
    
+  
    Notes on writing an R package
    
+  
    
+    Yang Liu
+ / 
+    2019-07-28
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
+    Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either
+install.packages("SHAPforxgboost")
or
+devtools::install_github("liuyanguu/SHAPforxgboost")
Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package.
+[…]…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    R package
+  
+  
+  
    
+  
    SHAP for XGBoost in R: SHAPforxgboost
    
+  
    
+    Yang Liu
+ / 
+    2019-07-18
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package.
 Please install from CRAN or Github.
-install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")

Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    Shiny in Blogdown
    
-  
    
-    Yang Liu
- / 
-    2019-02-24
-  
    
-  
    
-    
-    
-    
-    
-    
-    
-      
-    
+install.packages("SHAPforxgboost")
# or devtools::install_github("liuyanguu/SHAPforxgboost")
Why SHAP values
SHAP’s main advantages are local explanation and consistency in global model…
+    
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    Shiny in Blogdown
    
+  
    
+    Yang Liu
+ / 
+    2019-02-24
+  
    
+  
    
+    
+    
+    
+    
+    
+    
+      
+    
     Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here.
 The only way to do it is using iframe and write outside the chunk:
 <iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"></iframe>
-There is also a built-in function in knitr to do the same thing and write in the chunk. The only…
-    
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Machine Learning
-  
-  
-  
    
-  
    Study shrinkage and DART in xgboost modeling using a simple dataset
    
-  
    
-    Yang Liu
- / 
-    2018-11-15
-  
    
-  
    
-    
-    
-    
-    
-    
-    
-      
-    
+There is also a built-in function in knitr to do the same thing and write in the chunk. The only…
+    
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Machine Learning
+  
+  
+  
    
+  
    Study shrinkage and DART in xgboost modeling using a simple dataset
    
+  
    
+    Yang Liu
+ / 
+    2018-11-15
+  
    
+  
    
+    
+    
+    
+    
+    
+    
+      
+    
     It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model.
-[…] set.seed(123)
n0 <- 5
X <- data.frame(x1 = runif(n0), x2 = runif(n0))
Y <- c(1, 5, 20, 50, 100)
cbind(X,…
-    
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Machine Learning
-  
-  
-  
    
-  
    SHAP Visualization in R (first post)
    
-  
    
-    Yang Liu
- / 
-    2018-10-14
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+[…] set.seed(123)
n0 <- 5
X <- data.frame(x1 = runif(n0), x2 = runif(n0))
Y <- c(1, 5, 20, 50, 100)
cbind(X,…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Machine Learning
+  
+  
+  
    
+  
    SHAP Visualization in R (first post)
    
+  
    
+    Yang Liu
+ / 
+    2018-10-14
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     Update 19/07/21:
 Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R
 […] This is the example I used in the package SHAPforxgboost
-# Example use iris
suppressPackageStartupMessages({…
-        
-  
    
-
    
-
-
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    Spatial data in R: Dividing raster layers into equal-area rings
    
-  
    
-    Yang Liu
- / 
-    2018-07-20
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
+# Example use iris
suppressPackageStartupMessages({…
+        
+  
    
+
    
+
+
    
+  
    
+  
+  
+    Data Visualization
+  
+  
+  
    
+  
    Spatial data in R: Dividing raster layers into equal-area rings
    
+  
    
+    Yang Liu
+ / 
+    2018-07-20
+  
    
+  
    
+    
+    
+    
+    
+    
    Thumbnail
    
+    
+    
+    
+      
+    
     This data visualization example include:
 * Import .img file as a raster
 * Turn raster into a data.frame of points (coordinates) and values
@@ -197,77 +230,41 @@ 
    
-  
    
-  
-  
-    Data Visualization
-  
-  
-  
    
-  
    How to Draw Heatmap with Colorful Dendrogram
    
-  
    
-    Yang Liu
- / 
-    2018-07-16
-  
    
-  
    
-    
-    
-    
-    
-    
    Thumbnail
    
-    
-    
-    
-      
-    
-    This data visualization example include:
-* Hierarchical clustering, dendrogram and heat map based on normalized odds ratios
-* The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend
-* Heatmap is made by heatmap.2 from gplots using the built dendrogram
-* The rows are sorted by means from highest to lowest, it can be done in either the…
-        
-  
    
-
    
-
-
    
+[…] […] list.of.packages <- c("raster",…
+    
+  
    
+
    
+
+
    
 
-
    
-
-
-
-
    
+
+
+
+
    
 
 
+
+
+
+
+
    
+  
+  
+  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
+  
+
    
+
    
+
+
 
 
 
-
-
    
-  
-  
-  
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    
-  
-
    
-
-
-
-
-
-
 
 
-
-
-
-
-
+
+
+
+
+
diff --git a/page/3/index.html b/page/3/index.html
index 0328ecf..7a4bd2a 100644
--- a/page/3/index.html
+++ b/page/3/index.html
@@ -1,7 +1,7 @@
 
 
   
-	
+	
     
     
     
@@ -24,7 +24,7 @@
       
    
       
         Welcome to my blog
-        2021-03-07
+        2023-05-30
         
         
         
@@ -44,28 +44,32 @@
   
    
   
   
-    Machine Learning
+    Data Visualization
   
   
   
    
-  
    eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting
    
+  
    How to Draw Heatmap with Colorful Dendrogram
    
   
    
     Yang Liu
  / 
-    2018-07-09
+    2018-07-16
   
    
   
    
-    
+    
     
     
     
-    
    Thumbnail
    
+    
    Thumbnail
    
     
     
     
       
     
-    I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset.
The ensemble method is powerful as it combines the predictions from multiple machine learning algorithms together to…
+    This data visualization example include:
+* Hierarchical clustering, dendrogram and heat map based on normalized odds ratios
+* The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend
+* Heatmap is made by heatmap.2 from gplots using the built dendrogram
+* The rows are sorted by means from highest to lowest, it can be done in either the…
         
   
    
 
@@ -104,31 +108,29 @@ 
    Catalog of my old blog<
 
 
 
+    
 
    
 
 
@@ -147,9 +149,8 @@ 
    Catalog of my old blog<
 	ga('send', 'pageview');
 }
 
-
-
-
+
+
 
 
    
   
   
@@ -187,7 +188,6 @@ 
    Catalog of my old blog<
 }
 
 
-
 
 
 
diff --git a/post/2018-07-09-extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting.html b/post/2018-07-09-extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting.html
new file mode 100644
index 0000000..58af17f
--- /dev/null
+++ b/post/2018-07-09-extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting.html
@@ -0,0 +1,301 @@
+
+
+
+
+
+
+
+
+
    
+
    Overview
    
+
    I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset. The ensemble method is powerful as it combines the predictions from multiple machine learning algorithms together to make more accurate predictions than an individual model. Random Forest is among the most famous ones and it is easy to use. Random Forest is based on bagging (bootstrap aggregation) which averages the results over many decision trees from sub-samples. It further limits its search to only 1/3 of the features (in regression) to fit each tree, weakening the correlations among decision trees. It is easy to get a lower testing error compared to linear models.
    
+
    Boosting takes slower steps, making predictors sequentially instead of independently. It repetitively leverages the patterns in residuals, strengthens the model with weak predictions, and make it better. By combining the advantages from both random forest and gradient boosting, XGBoost gave the a prediction error ten times lower than boosting or random forest in my case.
    
+
    (Correction) 18.10.3. When I further test this dataset I realized there was a mistake. There is indeed no magic in machine learning. When I got a testing error ten-time smaller than other methods, I should question if this is a mistake. In the corrected result XGBoost still gave the lowest testing RMSE but was close to other two methods.
    
+
    Link to the earlier version: Model Selection using Lasso and Best Subset
    
+
    
+
    
+
    About the Data
    
+
    In sub-Saharan Africa, the region where deprivations in terms of living conditions are the most severe, slum dwellers represent an estimated 56% of the regions’ urban population (UN Habitat, 2016). Measuring informal settlements in a reliable way is a critical challenge for the United Nations to monitor the Sustainable Development Goals (SDGs) towards its 2030 Agenda for Sustainable Development. This data was collected by Slum Dwellers International (SDI), which was nominated for the Nobel Peace Prize in 2014. By extracting valuable information from the data, We will help the UN to create better SDG indicators by making full use of available data.
    
+
    In this exercise, we only model Share_Temporary: Share of Temporary Structure in Slums as the dependent variable. The independent variables are monitoring indicators like water, sanitation, housing conditions and overcrowding in African slum settlements. Dataset dimension is 973 x 153.
    
+
    
+
    
+
    1. Extreme Gradient Boosting
    
+
      
+
    • Random search: randomized parameters and update the record with best ones.
      • 
+
    • It turns out to be a very interesting method to scan for hyperparameters. It will take a while for 100 iterations.
      • 
+
    • The package xgboost is really fast.
      • 
+
    
+
    library(xgboost)
+# Randomize and bound
+best_param <- list()
+best_seednumber <- 1234
+best_rmse <- Inf
+best_rmse_index <- 0
+
+set.seed(1234)
+# In reality, might need 100 or 200 iterations
+for (iter in 1:10) {
+  param <- list(objective = "reg:squarederror",  # For regression
+                eval_metric = "rmse",      # rmse is used for regression
+                max_depth = sample(6:10, 1),
+                eta = runif(1, .01, .1),   # Learning rate, default: 0.3
+                subsample = runif(1, .6, .9),
+                colsample_bytree = runif(1, .5, .8), 
+                min_child_weight = sample(5:10, 1), # These two are important
+                max_delta_step = sample(5:10, 1)    # Can help to focus error
+                                                    # into a small range.
+  )
+  cv.nround <-  1000
+  cv.nfold <-  5 # 5-fold cross-validation
+  seed.number  <-  sample.int(10000, 1) # set seed for the cv
+  set.seed(seed.number)
+  mdcv <- xgb.cv(data = dtrain, params = param,  
+                 nfold = cv.nfold, nrounds = cv.nround,
+                 verbose = F, early_stopping_rounds = 8, maximize = FALSE)
+  
+  min_rmse_index  <-  mdcv$best_iteration
+  min_rmse <-  mdcv$evaluation_log[min_rmse_index]$test_rmse_mean
+  
+  if (min_rmse < best_rmse) {
+    best_rmse <- min_rmse
+    best_rmse_index <- min_rmse_index
+    best_seednumber <- seed.number
+    best_param <- param
+  }
+}
    
+
      
+
    • The best tuning parameters
      
+
      • 
+
    
+
    ##          objective eval_metric max_depth     eta subsample colsample_bytree
+## 1 reg:squarederror        rmse         9 0.09822      0.64           0.6853
+##   min_child_weight max_delta_step best_rmse_index best_rmse best_seednumber
+## 1                6              8              56    0.2102            3660
    
+
      
+
    • MSE
      
+
      • 
+
    
+
    ## [1] 0.04237
    
+
      
+
    • Feature Importance
      
+
      • 
+
    
+
    importance_matrix <- xgb.importance(feature_names = colnames(X_train), 
+                                    model = xg_mod)
+# Use `xgb.plot.importance`, which create a _barplot_ or use `xgb.ggplot.importance`
+library(Ckmeans.1d.dp) # for xgb.ggplot.importance
+xgb.ggplot.importance(importance_matrix, top_n = 15, measure = "Gain")
    
+
    
+
      
+
    • Plot only 2 trees as an example (use trees= 1)
      
+
      • 
+
    
+
    library("DiagrammeR")
+xgb.plot.tree(model = xg_mod, trees = 1, feature_names = colnames(X_train))
    
+
    
+
+
      
+
    • Plot all trees on one tree and plot it: A huge plot
      • 
+
    
+
    xgb.plot.multi.trees(model = xg_mod, n_first_tree = 1, feature_names = colnames(X_train))
    
+
    
+
+
    
+
    
+
    2. Gradient boosting
    
+
      
+
    • Use library gbm
      
+
      • 
+
    • Tuning Method: use train function from caret to scan a grid of parameters.
      • 
+
    
+
    library(gbm)   # for Gradient boosting
+library(caret) # scan the parameter grid using `train` function
    
+
    # time_now <- Sys.time()
+para_grid <- expand.grid(n.trees = (20*c(50:100)), 
+                         shrinkage = c(0.1, 0.05, 0.01), 
+                         interaction.depth = c(1,3,5),
+                         n.minobsinnode = 10)
+trainControl <- trainControl(method = "cv", number = 10)
+set.seed(123)
+gbm_caret <- train(Share_Temporary ~ ., mydata[train_idx,], 
+                   distribution = "gaussian", method = "gbm",
+                   trControl = trainControl, verbose = FALSE, 
+                   tuneGrid = para_grid, metric = "RMSE", bag.fraction = 0.75)  
+
+# Sys.time() - time_now 
+## Time difference of 2.283 mins
    
+
      
+
    • The tuning parameters that give the lowest MSE in training set CV.
      • 
+
    
+
    ##    n.trees interaction.depth shrinkage n.minobsinnode
+## 36    1700                 1      0.01             10
    
+
      
+
    • MSE
      • 
+
    
+
    ## [1] 0.04838
    
+
    
+
    
+
    3. Random Forest
    
+
      
+
    • Use library randomForest.
      
+
      • 
+
    
+
    library(randomForest)
+rf.fit <- randomForest(Share_Temporary ~ ., data = mydata2, subset = train_idx)
+# Test on test data: mydata[-train_idx,]
+yhat_bag <- predict(rf.fit, newdata = mydata2[-train_idx,])
    
+
      
+
    • MSE on the testing dataset:
      • 
+
    
+
    ## [1] 0.04359
    
+
      
+
    • Feature Importance (showing top 15)
      
+
        
+
      • The variables high on rank show the relative importance of features in the tree model
        • 
+
      • For example, Monthly Water Cost, Resettled Housing, and Population Estimate are the most influential features.
        • 
+
      • 
+
    
+
    varImpPlot(rf.fit, n.var=15)
    
+
    
+
    
+
    
+
    4. Lasso
    
+
      
+
    • Use library glmnet.
      
+Lasso is a shrinkage approach for feature selection. The tuning parameter lambda is the magnitudes of penalty. A increasing penalty shrinks coefficients towards zero. The advantage of a linear model is that the result is highly interpretable.
      • 
+
    • We use cross-validation to choose the lambda and corresponding features
      
+
      • 
+
    • The dotted line on the left is lambda.min, the lambda that generates the lowest MSE in the testing dataset. The dotted line on the right is lambda.1se, its corresponding MSE is not the lowest but acceptable, and it has even fewer features in the model. We use lambda.1se in our case.
      
+
      • 
+
    
+
    # Use cross-validation to select the lambda
+cv_lasso = cv.glmnet(X_train, Y_train, alpha=1) # Lasso regression
+plot(cv_lasso)
    
+
    
+
    # lambda selected by 1se rule
+(best_lam <- cv_lasso$lambda.1se)
    
+
    ## [1] 0.03845
    
+
      
+
    • MSE
      
+
      • 
+
    
+
    # Check prediction error in the testing dataset
+lasso_pred <- predict(lasso_mod, s = best_lam, newx = X_test)
+# The Mean squared error (MSE)
+(MSE_Lasso <- mean((lasso_pred - Y_test)^2))
    
+
    ## [1] 0.06751
    
+
      
+
    • The regression model for the selected lambda (lasso). We extract the coefficients from the selected model and run a linear regression.
      • 
+
    • The model has used 17 variables.
      • 
+
    • The most useful predictors selected by lasso include Water_MonthlyCost, Water_Sources: shared_taps, Resettled Housing and Eviction Threats. For these variables, higher values or binary variables being Yes are associated with fewer temporary structures in slums.
      • 
+
    • Relative importance of coefficients by showing standardized regression coefficients in decreasing order of their absolute values.
      
+
      • 
+
    
+
    coef_table2 <- data.frame(reg_lasso_summary$coefficients, stb = c(0, lm.beta(reg_lasso_mod)))
+coef_table2[order(abs(coef_table2$stb), decreasing = T),]
    
+
    ##                                           Estimate Std..Error t.value  Pr...t..
+## B14__resettled                          -1.500e-01  3.232e-02  -4.641 4.404e-06
+## DD1_Location_Problemscanal               1.896e-01  3.278e-02   5.785 1.261e-08
+## FF11_Water_MonthlyCost                  -3.354e-06  6.940e-07  -4.832 1.788e-06
+## FF1_8_Water_Sourceswells                -1.146e-01  2.294e-02  -4.995 8.084e-07
+## Eviction_Threats                         1.053e-01  2.428e-02   4.337 1.736e-05
+## B14__declared_legal_protected            8.688e-02  2.967e-02   2.928 3.561e-03
+## DD1_Location_Problemsslope               8.235e-02  2.366e-02   3.481 5.422e-04
+## EE2B_Current_Eviction_Seriousnessmedium -2.149e-01  6.549e-02  -3.282 1.101e-03
+## GG1_Sewer_Line                           7.801e-02  2.500e-02   3.120 1.908e-03
+## FF1_8_Water_Sourceswater_tankers        -1.203e-01  3.949e-02  -3.046 2.434e-03
+## GG7_Managerprivate                      -5.983e-02  2.417e-02  -2.475 1.363e-02
+## DD1_Location_Problemsflood_prone_area    4.860e-02  2.254e-02   2.156 3.152e-02
+## FF1_8_Water_Sourcescommunity_taps        4.666e-02  2.929e-02   1.593 1.118e-01
+## (Intercept)                              3.949e-01  3.957e-02   9.979 1.479e-21
+##                                              stb
+## B14__resettled                          -0.19503
+## DD1_Location_Problemscanal               0.18221
+## FF11_Water_MonthlyCost                  -0.17854
+## FF1_8_Water_Sourceswells                -0.15557
+## Eviction_Threats                         0.14234
+## B14__declared_legal_protected            0.11390
+## DD1_Location_Problemsslope               0.10926
+## EE2B_Current_Eviction_Seriousnessmedium -0.10030
+## GG1_Sewer_Line                           0.09478
+## FF1_8_Water_Sourceswater_tankers        -0.09144
+## GG7_Managerprivate                      -0.07741
+## DD1_Location_Problemsflood_prone_area    0.06608
+## FF1_8_Water_Sourcescommunity_taps        0.05158
+## (Intercept)                              0.00000
    
+
    
+
    
+
    5. Best Subset
    
+
      
+
    • Use library leaps.
      
+Best subset is a subset selection approach for feature selection. Not like stepwise or forward selection, best subset check all the possible feature combinations in theory. Since I select from 49 predictors but set the maximum size of subsets to be 25, there are C(49,25) + C(49,24) + …+ C(49,0) = 345 trillion models to check. As I discussed in my post, it won’t be possible to scan all of them. Both R and SAS use the branch and bound algorithm to speed up the calculation.
      • 
+
    • If without cross-validation we can use the traditional way to choose model: Adjusted R-squared, Cp(AIC), or BIC.
      • 
+
    • The turning parameter is to decide how many predictors to use. The selected number of feature also happens to be 17.
      
+Cross-validation selects more features than BIC but fewer than Adj Rsq or Cp(AIC).
      • 
+
    • The regression model selected and Standardized parameter estimates showing relative feature importance in decreasing order.
      • 
+
    
+
    ##                                       b.Estimate b.Std..Error b.t.value
+## B14__resettled                        -2.275e-01    4.111e-02   -5.5342
+## DD1_Location_Problemscanal             2.076e-01    4.235e-02    4.9029
+## FF11_Water_MonthlyCost                -3.553e-06    9.722e-07   -3.6545
+## Eviction_Threats                       1.227e-01    4.541e-02    2.7018
+## B14__declared_legal_protected          1.203e-01    4.067e-02    2.9583
+## FF1_8_Water_Sourceswater_tankers      -1.820e-01    5.533e-02   -3.2893
+## FF1_8_Water_Sourcesshared_taps        -1.117e-01    4.894e-02   -2.2831
+## DD1_Location_Problemsflood_prone_area  6.877e-02    3.054e-02    2.2516
+## GG7_10_Toilet_Typesindividual_toilets -6.214e-02    4.268e-02   -1.4561
+## FF1_8_Water_Sourcessprings            -5.563e-02    4.252e-02   -1.3085
+## JJ1_Electricity_Availableyes           5.644e-02    4.699e-02    1.2012
+## DD1_Location_Problemsgarbage_dump     -3.611e-02    3.347e-02   -1.0791
+## EE2A_Current_Eviction_Threat           2.456e-02    4.528e-02    0.5425
+## FF1_8_Water_Sourcesrivers             -4.108e-02    5.889e-02   -0.6976
+## FF1_8_Water_Sourcesdams               -2.846e-02    7.258e-02   -0.3921
+## FF12_Water_CollectionTime30_minutes    1.275e-02    4.478e-02    0.2847
+## DD1_Location_Problemsroad_side        -1.899e-03    3.160e-02   -0.0601
+## (Intercept)                            3.932e-01    7.437e-02    5.2871
+##                                       b.Pr...t..       stb
+## B14__resettled                         6.856e-08 -0.294884
+## DD1_Location_Problemscanal             1.554e-06  0.205337
+## FF11_Water_MonthlyCost                 3.044e-04 -0.185969
+## Eviction_Threats                       7.293e-03  0.164439
+## B14__declared_legal_protected          3.341e-03  0.156849
+## FF1_8_Water_Sourceswater_tankers       1.125e-03 -0.135517
+## FF1_8_Water_Sourcesshared_taps         2.313e-02 -0.096201
+## DD1_Location_Problemsflood_prone_area  2.508e-02  0.093098
+## GG7_10_Toilet_Typesindividual_toilets  1.464e-01 -0.058293
+## FF1_8_Water_Sourcessprings             1.917e-01 -0.056320
+## JJ1_Electricity_Availableyes           2.306e-01  0.050303
+## DD1_Location_Problemsgarbage_dump      2.814e-01 -0.044732
+## EE2A_Current_Eviction_Threat           5.879e-01  0.031196
+## FF1_8_Water_Sourcesrivers              4.860e-01 -0.027726
+## FF1_8_Water_Sourcesdams                6.953e-01 -0.015868
+## FF12_Water_CollectionTime30_minutes    7.760e-01  0.011364
+## DD1_Location_Problemsroad_side         9.521e-01 -0.002456
+## (Intercept)                            2.408e-07  0.000000
    
+
      
+
    • MSE
      • 
+
    
+
    ## [1] 0.06979
    
+
    
+
    
+
    Compare MSE
    
+
      
+
    • XGBoost has the lowest mean squared error
      
+
      • 
+
    • The real advantages of XGBoost include its speed and the ability to handle missing values
      • 
+
    
+
    ##   MSE_xgb MSE_boost MSE_Lasso MSE_rForest MSE_best.subset
+## 1 0.04237   0.04838   0.06751     0.04359         0.06979
    
+
    Original code is saved on github
    
+
    
diff --git a/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-11-1.png b/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-11-1.png
index d0b2158..26ff095 100644
Binary files a/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-11-1.png and b/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-11-1.png differ
diff --git a/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-2-1.png b/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-2-1.png
index acc6b16..409f019 100644
Binary files a/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-2-1.png and b/post/2018-10-05-shap-visualization-for-xgboost_files/figure-html/unnamed-chunk-2-1.png differ
diff --git a/post/2018/07/01/catalogue-of-my-old-blog/index.html b/post/2018/07/01/catalogue-of-my-old-blog/index.html
index 279d69c..c4e79e4 100644
--- a/post/2018/07/01/catalogue-of-my-old-blog/index.html
+++ b/post/2018/07/01/catalogue-of-my-old-blog/index.html
@@ -149,7 +149,7 @@ 
    Catalog
    
 
 
 
 
@@ -184,10 +184,9 @@ 
    Catalog
    
 }
 
 
-
   
    
 
 
-
 
 
 
diff --git a/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/index.html b/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/index.html
index 7b93f11..372cff6 100644
--- a/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/index.html
+++ b/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/index.html
@@ -251,7 +251,7 @@ 
    Version 3: If there is no color, and we do not reorder the branches
    
 
 
 
 
@@ -285,10 +285,9 @@ 
    Version 3: If there is no color, and we do not reorder the branches
    
 }
 
 
-
   
    
 
 
-
 
 
 
diff --git a/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/index.html b/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/index.html
index 2c164bd..c943949 100644
--- a/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/index.html
+++ b/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/index.html
@@ -469,10 +469,9 @@ 
    Original Code
    
 }
 
 
-
   
    
 
 
-
 
 
 
diff --git a/post/2018/10/14/shap-visualization-for-xgboost/index.html b/post/2018/10/14/shap-visualization-for-xgboost/index.html
index 4420969..d4a4661 100644
--- a/post/2018/10/14/shap-visualization-for-xgboost/index.html
+++ b/post/2018/10/14/shap-visualization-for-xgboost/index.html
@@ -69,17 +69,20 @@ 
    Yang Liu
 
 
    
 
+
 
 
    
 
    
+
    ## [23:14:47] WARNING: amalgamation/../src/objective/regression_obj.cu:174: reg:linear is now deprecated in favor of reg:squarederror.
    
 
    shap.plot.summary.wrap1(mod1, X_train, top_n = 10)

    Dependence plot for each feature

      -
    • Here we choose to show top 6 features ranked by mean|SHAP|
      • +
    • Here we choose to show top 6 features ranked by mean|SHAP|
      +
    data_long <-  shap.prep(mod1, X_train = X_train)
 shap_values <- shap.values(mod1, X_train)
@@ -223,7 +228,8 @@ 
    Dependence plot for each feature
    
 gridExtra::grid.arrange(grobs = fig_list, ncol = 2)

      -
    • If Use the built-in xgb.shap.plot function
      • +
    • If Use the built-in xgb.shap.plot function
      +
    xgboost::xgb.plot.shap(data = X_train, model = mod1, top_n = 4, n_col = 2)

    @@ -300,10 +306,9 @@

    Stack plot by clustering groups

    } -
    - diff --git a/post/2018/11/15/xgboost-dart-example/index.html b/post/2018/11/15/xgboost-dart-example/index.html index 485410f..ec5a820 100644 --- a/post/2018/11/15/xgboost-dart-example/index.html +++ b/post/2018/11/15/xgboost-dart-example/index.html @@ -370,10 +370,9 @@

    one_drop

    } -
    - diff --git a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-2-1.png b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-2-1.png index b3af423..0398b17 100644 Binary files a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-2-1.png and b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-2-1.png differ diff --git a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-3-1.png b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-3-1.png index 8a3fa50..10ad675 100644 Binary files a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-3-1.png and b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-3-1.png differ diff --git a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-1.png b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-1.png index cdd82cf..0bb29ab 100644 Binary files a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-1.png and b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-1.png differ diff --git a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-2.png b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-2.png index 5755c95..59539c6 100644 Binary files a/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-2.png and b/post/2019-04-17-ggplot-heatmap-us-50-states-map-and-china-province-map_files/figure-html/unnamed-chunk-4-2.png differ diff --git a/post/2019/02/24/shiny-in-blogdown/index.html b/post/2019/02/24/shiny-in-blogdown/index.html index f6856a3..2306aa1 100644 --- a/post/2019/02/24/shiny-in-blogdown/index.html +++ b/post/2019/02/24/shiny-in-blogdown/index.html @@ -149,10 +149,9 @@

    My Shiny app example

    } -
    - diff --git a/post/2019/07/18/visualization-of-shap-for-xgboost/index.html b/post/2019/07/18/visualization-of-shap-for-xgboost/index.html index 1097683..060042d 100644 --- a/post/2019/07/18/visualization-of-shap-for-xgboost/index.html +++ b/post/2019/07/18/visualization-of-shap-for-xgboost/index.html @@ -362,10 +362,9 @@

    References

    } -
    - diff --git a/post/2019/07/28/some-experience-on-writing-r-package/index.html b/post/2019/07/28/some-experience-on-writing-r-package/index.html index a6466ec..5e1e1e0 100644 --- a/post/2019/07/28/some-experience-on-writing-r-package/index.html +++ b/post/2019/07/28/some-experience-on-writing-r-package/index.html @@ -197,10 +197,9 @@

    Some nice suggestions from the CRAN team when submitting the package

    } -
    - diff --git a/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/index.html b/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/index.html index 8227203..f927d22 100644 --- a/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/index.html +++ b/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/index.html @@ -190,10 +190,9 @@

    Yang Liu } -
    - diff --git a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-1-1.png b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-1-1.png index ba471c7..e4bc720 100644 Binary files a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-1-1.png and b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-1-1.png differ diff --git a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-2-1.png b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-2-1.png index 5a45592..a5d750d 100644 Binary files a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-2-1.png and b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-2-1.png differ diff --git a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-1.png b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-1.png index b58086c..d292c46 100644 Binary files a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-1.png and b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-1.png differ diff --git a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-2.png b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-2.png index e4e571e..faf3374 100644 Binary files a/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-2.png and b/post/2020-06-15-us-and-china-admin1-covid-19-mortality-and-incidence-rate_files/figure-html/unnamed-chunk-3-2.png differ diff --git a/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/index.html b/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/index.html index 01726cc..5862a42 100644 --- a/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/index.html +++ b/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/index.html @@ -266,10 +266,9 @@

    Recover the array

    } -
    - diff --git a/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/index.html b/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/index.html index 613d5d7..da41485 100644 --- a/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/index.html +++ b/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/index.html @@ -74,20 +74,22 @@

    Yang Liu - + @@ -101,7 +103,7 @@

    Yang Liu

    1. US Map by state

    -

    The dt1 below can be used by both methods, but to use usmap only need two variables: state names and value to plot.

    +

    The dt1 below can be used by both methods, but using usmap only need two variables: state names and value to plot.

    suppressPackageStartupMessages({
 library(ggplot2)
 library(maps)
@@ -260,14 +262,10 @@ 
    Method 1. China map by province using downloaded shap files
    
 china_map_data <- china_map_data[!is.na(province)]
 china_map_data <- china_map_data[AREA > 0.1]
 head(china_map_data, 3)
    -
    ##       AREA PERIMETER BOU2_4M_ BOU2_4M_ID ADCODE93 ADCODE99
-## 1:  54.447    68.489        2         23   230000   230000
-## 2: 129.113   129.933        3         15   150000   150000
-## 3: 175.591    84.905        4         65   650000   650000
-##                                                            province id
-## 1:                                 <U+9ED1><U+9F99><U+6C5F><U+7701>  0
-## 2:                 <U+5185><U+8499><U+53E4><U+81EA><U+6CBB><U+533A>  1
-## 3: <U+65B0><U+7586><U+7EF4><U+543E><U+5C14><U+81EA><U+6CBB><U+533A>  2
    +
    ##       AREA PERIMETER BOU2_4M_ BOU2_4M_ID ADCODE93 ADCODE99         province id
+## 1:  54.447    68.489        2         23   230000   230000         黑龙江省  0
+## 2: 129.113   129.933        3         15   150000   150000     内蒙古自治区  1
+## 3: 175.591    84.905        4         65   650000   650000 新疆维吾尔自治区  2
    # 
 dt_china = setDT(fortify(china_map))
 head(dt_china, 3)
    @@ -396,10 +394,9 @@

    Method 2. Using geojsonMap (leaflet)

    } -
    - diff --git a/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/index.html b/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/index.html index c4a652d..d6612ae 100644 --- a/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/index.html +++ b/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/index.html @@ -70,9 +70,9 @@

    Yang Liu @@ -105,10 +105,10 @@

    China

    dt_JUH <- get.JHU.daily() dt_CN <- dt_JUH[Country_Region %in% c("China", "Taiwan*"),] dt_CN[Country_Region=="Taiwan*", Province_State:= "Taiwan"] -dt_china_map <- readRDS("D:/liuyanguu/Blogdown/Data/dt_china.rds") +dt_china_map <- readRDS(here::here("Data/dt_china.rds")) setnames(dt_china_map, "province_EN", "state") unique(dt_CN$Province_State[!dt_CN$Province_State%in%dt_china_map$state])     -
    ## [1] "Hong Kong" "Macau"
    +
    ## [1] "Hong Kong" "Macau"     "Unknown"
    make_heatmap(data = dt_CN, 
              geo_data = dt_china_map,
              state_var = "Province_State", fill_var = "Case-Fatality_Ratio")
    @@ -175,10 +175,9 @@

    Compared to lung cancer mortality

    } -
    - diff --git a/post/2021/01/06/rmarkdown-to-github-pages/index.html b/post/2021/01/06/rmarkdown-to-github-pages/index.html index 5529733..a5f758a 100644 --- a/post/2021/01/06/rmarkdown-to-github-pages/index.html +++ b/post/2021/01/06/rmarkdown-to-github-pages/index.html @@ -147,7 +147,7 @@

    RMarkdown into Github Pages

    @@ -182,10 +182,9 @@

    RMarkdown into Github Pages

    } -
    - diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_AGO_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_AGO_1_sp.rds new file mode 100644 index 0000000..9117da9 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_AGO_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_GHA_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_GHA_1_sp.rds new file mode 100644 index 0000000..b383ab7 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_GHA_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_HTI_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_HTI_1_sp.rds new file mode 100644 index 0000000..e644a17 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_HTI_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_TGO_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_TGO_1_sp.rds new file mode 100644 index 0000000..82fb09f Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_TGO_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_UGA_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_UGA_1_sp.rds new file mode 100644 index 0000000..c4ec10e Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_UGA_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_ZAF_1_sp.rds b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_ZAF_1_sp.rds new file mode 100644 index 0000000..06b8bd8 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/gadm36_ZAF_1_sp.rds differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/index.html b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index.html new file mode 100644 index 0000000..1798370 --- /dev/null +++ b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index.html @@ -0,0 +1,305 @@ + + + + + + + Plot multiple countries on the world map | Welcome to my blog + + + + + + + + + +
    +
    +
    + +
    + + Data Visualization + + ⊹ Spatial Data + +
    + +

    Plot multiple countries on the world map

    + + +

    Yang Liu +

    + + + +

    Tags: Geodata; Geographic +

    + + + +
    + + + +
    + + + +

    This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. +I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries!

    +
    +

    The most straightforward way

    +

    (You may download the world map from Github)

    +
    suppressPackageStartupMessages({
+library("data.table")
+library("ggplot2")
+library("rgdal")
+library("raster")
+library("rgeos")
+library("here")
+library("ggthemes")
+})
+
+# download data from GADM directly and row-bind the spatial polygons data frame
+cnames <- c("Haiti", "Togo", "Uganda", "Ghana", "South Africa", "Angola")
+download.GADM <- function(cname0) raster::getData("GADM", country = cname0, level = 1)
+dt_geos <- raster::bind(lapply(cnames, download.GADM))
+dt_geos <- sp::spTransform(dt_geos, CRS("+proj=robin")) # robin transformation
+dt_geos_df <- broom::tidy(dt_geos, region = "GID_1")
+
+  x_min <- min(dt_geos_df$long)*1.2
+  x_max <- max(dt_geos_df$long)*1.2
+  y_min <- min(dt_geos_df$lat)* 1.2
+  y_max <- max(dt_geos_df$lat)* 1.2
+  
+  map_theme <- ggthemes::theme_map() +
+    theme(legend.position = "bottom", legend.direction = "horizontal", legend.justification = c("center"))
+  
+  shp_world_robin <- readRDS(here::here("../Data/World.shp/sp.world.robin.rds")) # this has to be sourced locally
+  ggplot() +
+    geom_polygon(data = shp_world_robin, aes(x = long, y = lat, group = group), fill="lightgray", 
+                 colour = "white", size=0.05) +
+    geom_polygon(data = dt_geos, aes(x = long, y = lat, group = group), 
+                 color = "red", size=0.05, fill="#ffc069")  +
+    # if want to crop map:
+    coord_fixed(xlim=c(x_min, x_max), ylim=c(y_min, y_max)) +
+    map_theme +
+    guides(fill = guide_legend(nrow = 1, title.position = "top"))
    +

    +
    +
    +

    A more complicated example

    +

    Now we will also plot some values on the map. These values need to be merged directly to the spatial polygons sp data frame, or the corresponding data frame. Here I show the second approach: transforming sp data frame into a data frame first. And we can supply either the sp data frame or the data frame to ggplot2::geom_polygon.

    +

    Note that in GADM admin 1 shape files, GID_1 is a unique identifier even after binding multiple countries, here we intentionally use region names (the NAME_1 column from GADM file) as the region identifier to show a bug. And in my case we indeed have to use the region names to merge with estimates.

    +
    # `region` will become the id/group used to identify each area 
+# Either broom::tidy or ggplot2::fortify can work:
+# dt_geos_data <- broom::tidy(dt_geos, region = "NAME_1")
+# dt_geos_data <- ggplot2::fortify(dt_geos, region = "NAME_1")
+
+download.GADM.df <- function(cname0){
+  dfsp <- raster::getData("GADM", country = cname0, level = 1) # download data from GADM
+  dfsp <- sp::spTransform(dfsp, CRS("+proj=robin")) # robin transformation
+  dfsf <- broom::tidy(dfsp, region = "NAME_1") # get df
+  dfsf$country <- cname0 # add a country identifier 
+  return(setDT(dfsf))
+}
+
+dt_geos_data <- rbindlist(lapply(cnames, download.GADM.df))
+
+# imaging we want to plot some estimates, not only the map
+dt_admin1 <- unique(setDT(dt_geos_data)[,.(country, id)])
+set.seed(1234)
+dt_admin1$value <- rgamma(nrow(dt_admin1), shape = 4, scale = 15) # some random values
+
+setkey(dt_geos_data, country, id)
+setkey(dt_admin1, country, id)
+dt_geos_data_value <- dt_geos_data[dt_admin1]
+
+legend_break <- c(0, 25, 50, 75,  100, 150, 200, 300, 500)
+legend_label <- c("≤25", "25 to 50", "50 to 75", "75 to 100", "100 to 150","150 to 200","200 to 300",">300")
+legend_color <- c("#80BD41", "#CFF4FF","#feec9f","#ffc069","#fa8c16","#d46b08","#ad4e00","#612500")
+dt_geos_data_value$col <- cut(dt_geos_data_value$value, breaks = legend_break, labels = legend_label)
    +
    +
    +

    The bug: strange connecting lines

    +
    ggplot() +
+  geom_polygon(data = shp_world_robin, aes(x = long, y = lat, group = group), fill="lightgray", 
+               colour = "white", size=0.05) +
+  geom_polygon(data = dt_geos_data_value, aes(x = long, y = lat, group = group, fill= col), 
+               color = "red", size=0.05)  +
+  coord_fixed(xlim=c(x_min, x_max), ylim=c(y_min, y_max)) + # crop map
+  scale_fill_manual("Some random values for example", values = legend_color, drop = FALSE) +# Keep all legend item
+  map_theme + 
+  guides(fill = guide_legend(nrow = 1, title.position = "top"))
    +

    +

    The reason is some countries have shared admin names, which comes from NAME_1 in the GADM file (during the broom::tidy or ggplot2::fortify step).
    +Here, Haiti and Togo have these shared admin 1 names: “Centre.1” and “Centre.2” (see the table below).
    +The bug is solved by setting group to a unique identifier in the ggplot2::geom_polygon:

    +
    dt_geos_data_value[, county_group := paste(country, group, sep = "_")]
+unique(dt_geos_data_value[group %in% c("Centre.1"),.(country, id, group, county_group, value)])
    +
    ##    country     id    group   county_group    value
+## 1:   Haiti Centre Centre.1 Haiti_Centre.1 24.09058
+## 2:    Togo Centre Centre.1  Togo_Centre.1 54.32445
    +
    ggplot() +
+    geom_polygon(data = shp_world_robin, aes(x = long, y = lat, group = group), fill="lightgray", 
+                 colour = "white", size=0.05) +
+    geom_polygon(data = dt_geos_data_value, aes(x = long, y = lat, group = county_group, fill= col), 
+                 # it's important to set `group = county_group` instead of `group = group`
+                 color = "red", size=0.05)  +
+    coord_fixed(xlim=c(x_min, x_max), ylim=c(y_min, y_max))  +
+    scale_fill_manual("Some random values for example", values = legend_color, drop = FALSE) +# Keep all legend item
+    map_theme + 
+    guides(fill = guide_legend(nrow = 1, title.position = "top"))
    +

    +

    Hope this post is helpful to people who meet similar issues!

    +
    + +
    + + + + + + + + + + + + + + + + + +
    + + +
    +
    + + + +
    + + + +
    + + +comments powered by Disqus + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + + + + + + + + + + + + + + + + + diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-1-1.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-1-1.png new file mode 100644 index 0000000..660cd05 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-1-1.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-2-1.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-2-1.png new file mode 100644 index 0000000..7847d49 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-2-1.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-3-1.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-3-1.png new file mode 100644 index 0000000..7847d49 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-3-1.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-4-1.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-4-1.png new file mode 100644 index 0000000..8efd73b Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/index_files/figure-html/unnamed-chunk-4-1.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map.png new file mode 100644 index 0000000..6b0aefb Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp1.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp1.png new file mode 100644 index 0000000..3e67f34 Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp1.png differ diff --git a/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp2.png b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp2.png new file mode 100644 index 0000000..a582d6a Binary files /dev/null and b/post/2023/05/29/plot-multiple-countries-on-the-world-map/temp/World map_exp2.png differ diff --git a/post/index.html b/post/index.html index 3767d9b..caa47a0 100644 --- a/post/index.html +++ b/post/index.html @@ -1,268 +1,265 @@ - - - - - - - Posts | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Github - - -
    -

    RMarkdown to Github Pages

    -
    - Yang Liu - / - 2021-01-06 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + Posts | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + Data Visualization + + +
    +

    Plot multiple countries on the world map

    +
    + Yang Liu + / + 2023-05-29 +
    +
    + + + + +
    Thumbnail
    + + + + + + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +[…] (You may download the world map from Github) +suppressPackageStartupMessages({ library("data.table") library("ggplot2")… +     +
    +
    + +
    +
    + + + Github + + +
    +

    RMarkdown to Github Pages

    +
    + Yang Liu + / + 2021-01-06 +
    +
    + + + + +
    Thumbnail
    + + + + + Example of one of my Github pages: You might have read this GitHub and RStudio tutorial by searching this topic. It is quite long and confusing as it tries to teach Git at the same time. You don’t need any of those branch operations. And the example yaml code is not indented correctly (corrected below). Here is the short version: -[…] Make sure the yaml format is indented correct. Most… -     -
    -
    - -
    -
    - - - Data Visualization - - -
    -

    US and China Admin1 COVID19 mortality and incidence rate

    -
    - Yang Liu - / - 2020-06-14 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] Make sure the yaml format is indented correct. Most… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    US and China Admin1 COVID19 mortality and incidence rate

    +
    + Yang Liu + / + 2020-06-14 +
    +
    + + + + +
    Thumbnail
    + + + + + Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. Code hosted on my Github repo. -[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here("Code/COVID_make_map.R")) # US dt_JUH_US… -     -
    -
    - -
    -
    - - - Data Visualization - - -
    -

    ggplot US state and China province heatmap

    -
    - Yang Liu - / - 2020-06-12 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here("Code/COVID_make_map.R")) # US dt_JUH_US… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    ggplot US state and China province heatmap

    +
    + Yang Liu + / + 2020-06-12 +
    +
    + + + + +
    Thumbnail
    + + + + + (Updated on 2020-06-12, First posted on 2019-04-17) It sounds easy but turned out not as straight-forward as I thought. I will show: […] 50-state (including Alaska and Hawaii) United States thematic map, with map scale and with state abbreviations […] China thematic map, with map scale, with names of provinces in either English or Chinese -[…] The dt1 below can be used by both… -     -
    -
    - -
    -
    - - - Data Engineering - - -
    -

    Working with 3D array as long-format data in R

    -
    - Yang Liu - / - 2020-01-11 -
    -
    - - - - -
    Thumbnail
    - - - - - - check.and.install.pkgs <- function(pkgs){ new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d")) Happy New… -     -
    -
    - -
    -
    - - - Reproducible Workflow - - -
    -

    Drake: powerful tool for automatic reproducible workflow

    -
    - Yang Liu - / - 2019-09-15 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] The dt1 below can be used by both… +     +
    +
    + +
    +
    + + + Data Engineering + + +
    +

    Working with 3D array as long-format data in R

    +
    + Yang Liu + / + 2020-01-11 +
    +
    + + + + +
    Thumbnail
    + + + + + + check.and.install.pkgs <- function(pkgs){ new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d")) Happy New… +     +
    +
    + +
    +
    + + + Reproducible Workflow + + +
    +

    Drake: powerful tool for automatic reproducible workflow

    +
    + Yang Liu + / + 2019-09-15 +
    +
    + + + + +
    Thumbnail
    + + + + + drake is a powerful tool for automatic reproducible workflow. I found it a perfect match when used together with RMarkdown. There are great documentations online for drake thus here I only show a simple example working with RMarkdown. -RMarkdown file could be very slow to generate if lots of calculations are involved. Any small revise makes you rerun everything. When we use drake we can do all the… -     -
    -
    - -
    -
    - - - R package - - -
    -

    Notes on writing an R package

    -
    - Yang Liu - / - 2019-07-28 -
    -
    - - - - -
    Thumbnail
    - - - - - - Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either -install.packages("SHAPforxgboost") or -devtools::install_github("liuyanguu/SHAPforxgboost") Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package. -[…]… -     -
    -
    - -
    +RMarkdown file could be very slow to generate if lots of calculations are involved. Any small revise makes you rerun everything. When we use drake we can do all the… + +
    +
    + +
    -
    - - - -
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/post/index.xml b/post/index.xml index 140b723..121fe92 100644 --- a/post/index.xml +++ b/post/index.xml @@ -6,7 +6,18 @@ Recent content in Posts on Welcome to my blog Hugo -- gohugo.io en-us - Wed, 06 Jan 2021 00:00:00 +0000 + Mon, 29 May 2023 00:00:00 +0000 + + Plot multiple countries on the world map + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + Mon, 29 May 2023 00:00:00 +0000 + + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +The most straightforward way (You may download the world map from Github) +suppressPackageStartupMessages({ library(&quot;data.table&quot;) library(&quot;ggplot2&quot;) library(&quot;rgdal&quot;) library(&quot;raster&quot;) library(&quot;rgeos&quot;) library(&quot;here&quot;) library(&quot;ggthemes&quot;) }) # download data from GADM directly and row-bind the spatial polygons data frame cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;) download. + + RMarkdown to Github Pages https://liuyanguu.github.io/post/2021/01/06/rmarkdown-to-github-pages/ @@ -24,7 +35,7 @@ Create in the root directory a yaml file: " Sun, 14 Jun 2020 00:00:00 +0000 https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/ - US China Compared to lung cancer mortality Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. + US China Compared to lung cancer mortality Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. Code hosted on my Github repo. US # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here(&quot;Code/COVID_make_map.R&quot;)) # US dt_JUH_US &lt;- get.JHU.us.state() make_heatmap(data = dt_JUH_US, geo_data = get_state_name(), state_var = &quot;Province_State&quot;, fill_var = &quot;Mortality_Rate&quot;, label_var = &quot;abb&quot;) us_maps &lt;- lapply(c(&quot;Mortality_Rate&quot;, &quot;Incident_Rate&quot;, &quot;Testing_Rate&quot;, &quot;Hospitalization_Rate&quot;), make_heatmap, data = dt_JUH_US, geo_data = get_state_name(), state_var = &quot;Province_State&quot;, label_var = &quot;abb&quot;) plot_grid &lt;- gridExtra::grid. @@ -44,9 +55,9 @@ US # major function, can download from Github repo Blogdown/hugo-xmag/Code sourc Sat, 11 Jan 2020 00:00:00 +0000 https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/ - Example: a 3-D array with dimension 4x3x2 Melt into long Recover the array check.and.install.pkgs &lt;- function(pkgs){ new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;)) Happy New Year! + Example: a 3-D array with dimension 4x3x2 Melt into long Recover the array check.and.install.pkgs &lt;- function(pkgs){ new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;)) Happy New Year! Recently I spent some time working with array in R. -I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. +I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. We can just melt it into a long data, do the work, and arrange back into array in the end if needed. @@ -65,8 +76,9 @@ RMarkdown file could be very slow to generate if lots of calculations are involv Sun, 28 Jul 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/ - Some of my own experience On description Namespace load vs attach Documentation Potential problems when checking the package Some nice suggestions from the CRAN team when submitting the package Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either -install.packages(&quot;SHAPforxgboost&quot;) or + Some of my own experience On description Namespace load vs attach Documentation Potential problems when checking the package Some nice suggestions from the CRAN team when submitting the package Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either +install.packages(&quot;SHAPforxgboost&quot;) or +devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package. @@ -75,9 +87,9 @@ install.packages(&quot;SHAPforxgboost&quot;) or Thu, 18 Jul 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/ - The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. + The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. +install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. @@ -86,10 +98,10 @@ install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_gith Sun, 24 Feb 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/ - How to embed Shiny My Shiny app example How to embed Shiny Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here. + How to embed Shiny My Shiny app example How to embed Shiny Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here. The only way to do it is using iframe and write outside the chunk: &lt;iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"&gt;&lt;/iframe&gt; -There is also a built-in function in knitr to do the same thing and write in the chunk. +There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it. @@ -98,8 +110,8 @@ There is also a built-in function in knitr to do the same thing and write in the Thu, 15 Nov 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/ - Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -Data set.seed(123) n0 &lt;- 5 X &lt;- data. + Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. +Data set.seed(123) n0 &lt;- 5 X &lt;- data.frame(x1 = runif(n0), x2 = runif(n0)) Y &lt;- c(1, 5, 20, 50, 100) cbind(X, Y) ## x1 x2 Y ## 1 0. @@ -108,7 +120,7 @@ Data set.seed(123) n0 &lt;- 5 X &lt;- data. Sun, 14 Oct 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/ - Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: + Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R @@ -118,7 +130,7 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post Fri, 20 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/ - Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: + Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: * Import .img file as a raster * Turn raster into a data.frame of points (coordinates) and values * Dividing the points into 100 equal-area rings @@ -132,19 +144,10 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post Mon, 16 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/ - Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: + Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: * Hierarchical clustering, dendrogram and heat map based on normalized odds ratios * The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend -* Heatmap is made by heatmap. - - - - eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting - https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/ - Mon, 09 Jul 2018 00:00:00 +0000 - - https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/ - Overview About the Data 1. Extreme Gradient Boosting 2. Gradient boosting 3. Random Forest 4. Lasso 5. Best Subset Compare MSEs Original Code are stored on github Overview I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset. +* Heatmap is made by heatmap.2 from gplots using the built dendrogram @@ -154,7 +157,7 @@ Since my R Package SHAPforxgboost has been released on CRAN, I updated this post https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/ Introducing my new blog written solely in R Markdown When I realized it was so convenient to write blog directly using R Markdown, I searched if there is a specific tool for it. And I found Blogdown, an R package developed by Yihui Xie, who also developed R Markdown. - Catalog As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018. +Catalog As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018. diff --git a/post/page/1/index.html b/post/page/1/index.html index 5bc707b..c055f15 100644 --- a/post/page/1/index.html +++ b/post/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/post/ \ No newline at end of file + + + + https://liuyanguu.github.io/post/ + + + + + + diff --git a/post/page/2/index.html b/post/page/2/index.html index a6166a7..a568b74 100644 --- a/post/page/2/index.html +++ b/post/page/2/index.html @@ -1,194 +1,227 @@ - - - - - - - Posts | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - R package - - -
    -

    SHAP for XGBoost in R: SHAPforxgboost

    -
    - Yang Liu - / - 2019-07-18 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + Posts | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + R package + + +
    +

    Notes on writing an R package

    +
    + Yang Liu + / + 2019-07-28 +
    +
    + + + + +
    Thumbnail
    + + + + + + Although ‘SHAPforxgboost’ is not a package too complicated, it took me some time to get the package pass all the cran check. Now (Aug.03,2019) it is available on cran. Install by either +install.packages("SHAPforxgboost") or +devtools::install_github("liuyanguu/SHAPforxgboost") Use the ‘usethis’ package https://usethis.r-lib.org/ to set up the structure of the package. +[…]… +     +
    +
    + +
    +
    + + + R package + + +
    +

    SHAP for XGBoost in R: SHAPforxgboost

    +
    + Yang Liu + / + 2019-07-18 +
    +
    + + + + +
    Thumbnail
    + + + + + I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… -     -
    -
    - -
    -
    - - - Data Visualization - - -
    -

    Shiny in Blogdown

    -
    - Yang Liu - / - 2019-02-24 -
    -
    - - - - - - - - +install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… + +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    Shiny in Blogdown

    +
    + Yang Liu + / + 2019-02-24 +
    +
    + + + + + + + + Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here. The only way to do it is using iframe and write outside the chunk: <iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"></iframe> -There is also a built-in function in knitr to do the same thing and write in the chunk. The only… - -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    Study shrinkage and DART in xgboost modeling using a simple dataset

    -
    - Yang Liu - / - 2018-11-15 -
    -
    - - - - - - - - +There is also a built-in function in knitr to do the same thing and write in the chunk. The only… + +
    +
    + +
    +
    + + + Machine Learning + + +
    +

    Study shrinkage and DART in xgboost modeling using a simple dataset

    +
    + Yang Liu + / + 2018-11-15 +
    +
    + + + + + + + + It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… - -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    SHAP Visualization in R (first post)

    -
    - Yang Liu - / - 2018-10-14 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… +     +
    +
    + +
    +
    + + + Machine Learning + + +
    +

    SHAP Visualization in R (first post)

    +
    + Yang Liu + / + 2018-10-14 +
    +
    + + + + +
    Thumbnail
    + + + + + Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R […] This is the example I used in the package SHAPforxgboost -# Example use iris suppressPackageStartupMessages({… -     -
    -
    - -
    -
    - - - Data Visualization - - -
    -

    Spatial data in R: Dividing raster layers into equal-area rings

    -
    - Yang Liu - / - 2018-07-20 -
    -
    - - - - -
    Thumbnail
    - - - - - +# Example use iris suppressPackageStartupMessages({… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    Spatial data in R: Dividing raster layers into equal-area rings

    +
    + Yang Liu + / + 2018-07-20 +
    +
    + + + + +
    Thumbnail
    + + + + + This data visualization example include: * Import .img file as a raster * Turn raster into a data.frame of points (coordinates) and values @@ -196,77 +229,41 @@

    -
    - - - Data Visualization - - -
    -

    How to Draw Heatmap with Colorful Dendrogram

    -
    - Yang Liu - / - 2018-07-16 -
    -
    - - - - -
    Thumbnail
    - - - - - - This data visualization example include: -* Hierarchical clustering, dendrogram and heat map based on normalized odds ratios -* The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend -* Heatmap is made by heatmap.2 from gplots using the built dendrogram -* The rows are sorted by means from highest to lowest, it can be done in either the… -     -
    -
    - -
    +[…] […] list.of.packages <- c("raster",… + +
    +
    + +
    -
    - - - -

    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/post/page/3/index.html b/post/page/3/index.html index 37511fc..5806834 100644 --- a/post/page/3/index.html +++ b/post/page/3/index.html @@ -23,7 +23,7 @@
    Posts - 2021-03-07 + 2023-05-30 @@ -43,28 +43,32 @@
    - Machine Learning + Data Visualization
    -

    eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting

    +

    How to Draw Heatmap with Colorful Dendrogram

    Yang Liu / - 2018-07-09 + 2018-07-16
    - + -
    Thumbnail
    +
    Thumbnail
    - I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset. The ensemble method is powerful as it combines the predictions from multiple machine learning algorithms together to… + This data visualization example include: +* Hierarchical clustering, dendrogram and heat map based on normalized odds ratios +* The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend +* Heatmap is made by heatmap.2 from gplots using the built dendrogram +* The rows are sorted by means from highest to lowest, it can be done in either the…
    @@ -103,31 +107,29 @@

    Catalog of my old blog< +

    @@ -146,9 +148,8 @@

    Catalog of my old blog< ga('send', 'pageview'); } - - - + +
    @@ -186,7 +187,6 @@

    Catalog of my old blog< } - diff --git a/rmarkdown-libs/htmlwidgets/htmlwidgets.js b/rmarkdown-libs/htmlwidgets/htmlwidgets.js index 6f3d672..da8b236 100644 --- a/rmarkdown-libs/htmlwidgets/htmlwidgets.js +++ b/rmarkdown-libs/htmlwidgets/htmlwidgets.js @@ -249,13 +249,13 @@ function tryEval(code) { var result = null; try { - result = eval(code); + result = eval("(" + code + ")"); } catch(error) { - if (!error instanceof SyntaxError) { + if (!(error instanceof SyntaxError)) { throw error; } try { - result = eval("(" + code + ")"); + result = eval(code); } catch(e) { if (e instanceof SyntaxError) { throw error; diff --git a/rmarkdown-libs/leaflet-binding/leaflet.js b/rmarkdown-libs/leaflet-binding/leaflet.js index c6c233b..1eafcbe 100644 --- a/rmarkdown-libs/leaflet-binding/leaflet.js +++ b/rmarkdown-libs/leaflet-binding/leaflet.js @@ -4,14 +4,17 @@ Object.defineProperty(exports, "__esModule", { value: true }); - -var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }(); +exports["default"] = undefined; var _util = require("./util"); function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } } -var ClusterLayerStore = function () { +function _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } + +function _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; } + +var ClusterLayerStore = /*#__PURE__*/function () { function ClusterLayerStore(group) { _classCallCheck(this, ClusterLayerStore); @@ -26,8 +29,10 @@ var ClusterLayerStore = function () { if (this._layers[id]) { this._group.removeLayer(this._layers[id]); } + this._layers[id] = layer; } + this._group.addLayer(layer); } }, { @@ -38,9 +43,11 @@ var ClusterLayerStore = function () { } id = (0, _util.asArray)(id); + for (var i = 0; i < id.length; i++) { if (this._layers[id[i]]) { this._group.removeLayer(this._layers[id[i]]); + delete this._layers[id[i]]; } } @@ -49,6 +56,7 @@ var ClusterLayerStore = function () { key: "clear", value: function clear() { this._layers = {}; + this._group.clearLayers(); } }]); @@ -56,7 +64,7 @@ var ClusterLayerStore = function () { return ClusterLayerStore; }(); -exports.default = ClusterLayerStore; +exports["default"] = ClusterLayerStore; },{"./util":17}],2:[function(require,module,exports){ @@ -66,11 +74,13 @@ Object.defineProperty(exports, "__esModule", { value: true }); -var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }(); - function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } } -var ControlStore = function () { +function _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } + +function _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; } + +var ControlStore = /*#__PURE__*/function () { function ControlStore(map) { _classCallCheck(this, ControlStore); @@ -86,19 +96,23 @@ var ControlStore = function () { if (this._controlsById[id]) { this._map.removeControl(this._controlsById[id]); } + this._controlsById[id] = control; } else { this._controlsNoId.push(control); } + this._map.addControl(control); } }, { key: "get", value: function get(id) { var control = null; + if (this._controlsById[id]) { control = this._controlsById[id]; } + return control; } }, { @@ -106,7 +120,9 @@ var ControlStore = function () { value: function remove(id) { if (this._controlsById[id]) { var control = this._controlsById[id]; + this._map.removeControl(control); + delete this._controlsById[id]; } } @@ -115,14 +131,18 @@ var ControlStore = function () { value: function clear() { for (var i = 0; i < this._controlsNoId.length; i++) { var control = this._controlsNoId[i]; + this._map.removeControl(control); } + this._controlsNoId = []; for (var key in this._controlsById) { var _control = this._controlsById[key]; + this._map.removeControl(_control); } + this._controlsById = {}; } }]); @@ -130,7 +150,7 @@ var ControlStore = function () { return ControlStore; }(); -exports.default = ControlStore; +exports["default"] = ControlStore; },{}],3:[function(require,module,exports){ @@ -149,46 +169,56 @@ var _proj4leaflet = require("./global/proj4leaflet"); var _proj4leaflet2 = _interopRequireDefault(_proj4leaflet); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } // Helper function to instanciate a ICRS instance. function getCRS(crsOptions) { - var crs = _leaflet2.default.CRS.EPSG3857; // Default Spherical Mercator + var crs = _leaflet2["default"].CRS.EPSG3857; // Default Spherical Mercator switch (crsOptions.crsClass) { case "L.CRS.EPSG3857": - crs = _leaflet2.default.CRS.EPSG3857; + crs = _leaflet2["default"].CRS.EPSG3857; break; + case "L.CRS.EPSG4326": - crs = _leaflet2.default.CRS.EPSG4326; + crs = _leaflet2["default"].CRS.EPSG4326; break; + case "L.CRS.EPSG3395": - crs = _leaflet2.default.CRS.EPSG3395; + crs = _leaflet2["default"].CRS.EPSG3395; break; + case "L.CRS.Simple": - crs = _leaflet2.default.CRS.Simple; + crs = _leaflet2["default"].CRS.Simple; break; + case "L.Proj.CRS": if (crsOptions.options && crsOptions.options.bounds) { - crsOptions.options.bounds = _leaflet2.default.bounds(crsOptions.options.bounds); + crsOptions.options.bounds = _leaflet2["default"].bounds(crsOptions.options.bounds); } + if (crsOptions.options && crsOptions.options.transformation) { - crsOptions.options.transformation = new _leaflet2.default.Transformation(crsOptions.options.transformation[0], crsOptions.options.transformation[1], crsOptions.options.transformation[2], crsOptions.options.transformation[3]); + crsOptions.options.transformation = new _leaflet2["default"].Transformation(crsOptions.options.transformation[0], crsOptions.options.transformation[1], crsOptions.options.transformation[2], crsOptions.options.transformation[3]); } - crs = new _proj4leaflet2.default.CRS(crsOptions.code, crsOptions.proj4def, crsOptions.options); + + crs = new _proj4leaflet2["default"].CRS(crsOptions.code, crsOptions.proj4def, crsOptions.options); break; + case "L.Proj.CRS.TMS": if (crsOptions.options && crsOptions.options.bounds) { - crsOptions.options.bounds = _leaflet2.default.bounds(crsOptions.options.bounds); + crsOptions.options.bounds = _leaflet2["default"].bounds(crsOptions.options.bounds); } + if (crsOptions.options && crsOptions.options.transformation) { - crsOptions.options.transformation = _leaflet2.default.Transformation(crsOptions.options.transformation[0], crsOptions.options.transformation[1], crsOptions.options.transformation[2], crsOptions.options.transformation[3]); - } - // L.Proj.CRS.TMS is deprecated as of Leaflet 1.x, fall back to L.Proj.CRS + crsOptions.options.transformation = _leaflet2["default"].Transformation(crsOptions.options.transformation[0], crsOptions.options.transformation[1], crsOptions.options.transformation[2], crsOptions.options.transformation[3]); + } // L.Proj.CRS.TMS is deprecated as of Leaflet 1.x, fall back to L.Proj.CRS //crs = new Proj4Leaflet.CRS.TMS(crsOptions.code, crsOptions.proj4def, crsOptions.projectedBounds, crsOptions.options); - crs = new _proj4leaflet2.default.CRS(crsOptions.code, crsOptions.proj4def, crsOptions.options); + + + crs = new _proj4leaflet2["default"].CRS(crsOptions.code, crsOptions.proj4def, crsOptions.options); break; } + return crs; } @@ -199,21 +229,23 @@ function getCRS(crsOptions) { Object.defineProperty(exports, "__esModule", { value: true }); - -var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }(); +exports["default"] = undefined; var _util = require("./util"); function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } } -var DataFrame = function () { +function _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } + +function _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; } + +var DataFrame = /*#__PURE__*/function () { function DataFrame() { _classCallCheck(this, DataFrame); this.columns = []; this.colnames = []; this.colstrict = []; - this.effectiveLength = 0; this.colindices = {}; } @@ -225,7 +257,6 @@ var DataFrame = function () { this.effectiveLength = 0; this.colindices = {}; - this.columns.forEach(function (column, i) { _this.effectiveLength = Math.max(_this.effectiveLength, column.length); _this.colindices[_this.colnames[i]] = i; @@ -253,10 +284,9 @@ var DataFrame = function () { index = this.colnames.length; this.colnames.push(name); } - this.columns[index] = (0, _util.asArray)(values); - this.colstrict[index] = !!strict; - // TODO: Validate strictness (ensure lengths match up with other stricts) + this.columns[index] = (0, _util.asArray)(values); + this.colstrict[index] = !!strict; // TODO: Validate strictness (ensure lengths match up with other stricts) this._updateCachedProperties(); @@ -269,6 +299,7 @@ var DataFrame = function () { Object.keys(obj).forEach(function (name) { var coldata = obj[name]; + _this2.col(name, coldata); }); return this; @@ -279,8 +310,8 @@ var DataFrame = function () { var _this3 = this; if (row > this.effectiveLength) throw new Error("Row argument was out of bounds: " + row + " > " + this.effectiveLength); - var colIndex = -1; + if (typeof col === "undefined") { var rowData = {}; this.colnames.forEach(function (name, i) { @@ -292,6 +323,7 @@ var DataFrame = function () { } else if (typeof col === "number") { colIndex = col; } + if (colIndex < 0 || colIndex > this.columns.length) { if (missingOK) return void 0;else throw new Error("Unknown column index: " + col); } @@ -308,7 +340,7 @@ var DataFrame = function () { return DataFrame; }(); -exports.default = DataFrame; +exports["default"] = DataFrame; },{"./util":17}],5:[function(require,module,exports){ @@ -318,34 +350,13 @@ var _leaflet = require("./global/leaflet"); var _leaflet2 = _interopRequireDefault(_leaflet); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } // In RMarkdown's self-contained mode, we don't have a way to carry around the -// images that Leaflet needs but doesn't load into the page. Instead, we'll set -// data URIs for the default marker, and let any others be loaded via CDN. -if (typeof _leaflet2.default.Icon.Default.imagePath === "undefined") { - // if in a local file, support http - switch (window.location.protocol) { - case "http:": - // don't force http site to be done with https - _leaflet2.default.Icon.Default.imagePath = "http://cdn.leafletjs.com/leaflet/v1.3.1/images/"; - break; - default: - // file - // https - // otherwise use https as it works on files and https - _leaflet2.default.Icon.Default.imagePath = "https://unpkg.com/leaflet@1.3.1/dist/images/"; - break; - } - // don't know how to make this dataURI work since - // will be appended to Defaul.imagePath above - /* - if (L.Browser.retina) { - L.Icon.Default.prototype.options.iconUrl = "data:image/png;base64,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"; - } else { - L.Icon.Default.prototype.options.iconUrl = "data:image/png;base64,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"; - } - */ +// images that Leaflet needs but doesn't load into the page. Instead, we'll use +// the unpkg CDN. +if (typeof _leaflet2["default"].Icon.Default.imagePath === "undefined") { + _leaflet2["default"].Icon.Default.imagePath = "https://unpkg.com/leaflet@1.3.1/dist/images/"; } @@ -356,22 +367,21 @@ var _leaflet = require("./global/leaflet"); var _leaflet2 = _interopRequireDefault(_leaflet); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } // add texxtsize, textOnly, and style -_leaflet2.default.Tooltip.prototype.options.textsize = "10px"; -_leaflet2.default.Tooltip.prototype.options.textOnly = false; -_leaflet2.default.Tooltip.prototype.options.style = null; +_leaflet2["default"].Tooltip.prototype.options.textsize = "10px"; +_leaflet2["default"].Tooltip.prototype.options.textOnly = false; +_leaflet2["default"].Tooltip.prototype.options.style = null; // copy original layout to not completely stomp it. -// copy original layout to not completely stomp it. -var initLayoutOriginal = _leaflet2.default.Tooltip.prototype._initLayout; +var initLayoutOriginal = _leaflet2["default"].Tooltip.prototype._initLayout; -_leaflet2.default.Tooltip.prototype._initLayout = function () { +_leaflet2["default"].Tooltip.prototype._initLayout = function () { initLayoutOriginal.call(this); this._container.style.fontSize = this.options.textsize; if (this.options.textOnly) { - _leaflet2.default.DomUtil.addClass(this._container, "leaflet-tooltip-text-only"); + _leaflet2["default"].DomUtil.addClass(this._container, "leaflet-tooltip-text-only"); } if (this.options.style) { @@ -389,9 +399,10 @@ var _leaflet = require("./global/leaflet"); var _leaflet2 = _interopRequireDefault(_leaflet); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } var protocolRegex = /^\/\//; + var upgrade_protocol = function upgrade_protocol(urlTemplate) { if (protocolRegex.test(urlTemplate)) { if (window.location.protocol === "file:") { @@ -400,77 +411,80 @@ var upgrade_protocol = function upgrade_protocol(urlTemplate) { urlTemplate = "http:" + urlTemplate; } } + return urlTemplate; }; -var originalLTileLayerInitialize = _leaflet2.default.TileLayer.prototype.initialize; -_leaflet2.default.TileLayer.prototype.initialize = function (urlTemplate, options) { +var originalLTileLayerInitialize = _leaflet2["default"].TileLayer.prototype.initialize; + +_leaflet2["default"].TileLayer.prototype.initialize = function (urlTemplate, options) { urlTemplate = upgrade_protocol(urlTemplate); originalLTileLayerInitialize.call(this, urlTemplate, options); }; -var originalLTileLayerWMSInitialize = _leaflet2.default.TileLayer.WMS.prototype.initialize; -_leaflet2.default.TileLayer.WMS.prototype.initialize = function (urlTemplate, options) { +var originalLTileLayerWMSInitialize = _leaflet2["default"].TileLayer.WMS.prototype.initialize; + +_leaflet2["default"].TileLayer.WMS.prototype.initialize = function (urlTemplate, options) { urlTemplate = upgrade_protocol(urlTemplate); originalLTileLayerWMSInitialize.call(this, urlTemplate, options); }; },{"./global/leaflet":10}],8:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); -exports.default = global.HTMLWidgets; +exports["default"] = global.HTMLWidgets; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{}],9:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); -exports.default = global.jQuery; +exports["default"] = global.jQuery; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{}],10:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); -exports.default = global.L; +exports["default"] = global.L; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{}],11:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); -exports.default = global.L.Proj; +exports["default"] = global.L.Proj; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{}],12:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); -exports.default = global.Shiny; +exports["default"] = global.Shiny; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{}],13:[function(require,module,exports){ "use strict"; @@ -520,38 +534,40 @@ var _clusterLayerStore = require("./cluster-layer-store"); var _clusterLayerStore2 = _interopRequireDefault(_clusterLayerStore); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } window.LeafletWidget = {}; window.LeafletWidget.utils = {}; -var methods = window.LeafletWidget.methods = _jquery2.default.extend({}, _methods2.default); -window.LeafletWidget.DataFrame = _dataframe2.default; -window.LeafletWidget.ClusterLayerStore = _clusterLayerStore2.default; -window.LeafletWidget.utils.getCRS = _crs_utils.getCRS; -// Send updated bounds back to app. Takes a leaflet event object as input. +var methods = window.LeafletWidget.methods = _jquery2["default"].extend({}, _methods2["default"]); + +window.LeafletWidget.DataFrame = _dataframe2["default"]; +window.LeafletWidget.ClusterLayerStore = _clusterLayerStore2["default"]; +window.LeafletWidget.utils.getCRS = _crs_utils.getCRS; // Send updated bounds back to app. Takes a leaflet event object as input. + function updateBounds(map) { var id = map.getContainer().id; var bounds = map.getBounds(); - _shiny2.default.onInputChange(id + "_bounds", { + _shiny2["default"].onInputChange(id + "_bounds", { north: bounds.getNorthEast().lat, east: bounds.getNorthEast().lng, south: bounds.getSouthWest().lat, west: bounds.getSouthWest().lng }); - _shiny2.default.onInputChange(id + "_center", { + + _shiny2["default"].onInputChange(id + "_center", { lng: map.getCenter().lng, lat: map.getCenter().lat }); - _shiny2.default.onInputChange(id + "_zoom", map.getZoom()); + + _shiny2["default"].onInputChange(id + "_zoom", map.getZoom()); } function preventUnintendedZoomOnScroll(map) { // Prevent unwanted scroll capturing. Similar in purpose to // https://github.com/CliffCloud/Leaflet.Sleep but with a // different set of heuristics. - // The basic idea is that when a mousewheel/DOMMouseScroll // event is seen, we disable scroll wheel zooming until the // user moves their mouse cursor or clicks on the map. This @@ -561,96 +577,96 @@ function preventUnintendedZoomOnScroll(map) { // handle this by examining the mousemove event's screenX // and screenY properties; if they change, we know it's a // "true" move. - // lastScreen can never be null, but its x and y can. - var lastScreen = { x: null, y: null }; - (0, _jquery2.default)(document).on("mousewheel DOMMouseScroll", "*", function (e) { + var lastScreen = { + x: null, + y: null + }; + (0, _jquery2["default"])(document).on("mousewheel DOMMouseScroll", "*", function (e) { // Disable zooming (until the mouse moves or click) - map.scrollWheelZoom.disable(); - // Any mousemove events at this screen position will be ignored. - lastScreen = { x: e.originalEvent.screenX, y: e.originalEvent.screenY }; + map.scrollWheelZoom.disable(); // Any mousemove events at this screen position will be ignored. + + lastScreen = { + x: e.originalEvent.screenX, + y: e.originalEvent.screenY + }; }); - (0, _jquery2.default)(document).on("mousemove", "*", function (e) { + (0, _jquery2["default"])(document).on("mousemove", "*", function (e) { // Did the mouse really move? if (lastScreen.x !== null && e.screenX !== lastScreen.x || e.screenY !== lastScreen.y) { // It really moved. Enable zooming. map.scrollWheelZoom.enable(); - lastScreen = { x: null, y: null }; + lastScreen = { + x: null, + y: null + }; } }); - (0, _jquery2.default)(document).on("mousedown", ".leaflet", function (e) { + (0, _jquery2["default"])(document).on("mousedown", ".leaflet", function (e) { // Clicking always enables zooming. map.scrollWheelZoom.enable(); - lastScreen = { x: null, y: null }; + lastScreen = { + x: null, + y: null + }; }); } -_htmlwidgets2.default.widget({ - +_htmlwidgets2["default"].widget({ name: "leaflet", type: "output", factory: function factory(el, width, height) { - var map = null; - return { - // we need to store our map in our returned object. getMap: function getMap() { return map; }, - renderValue: function renderValue(data) { - // Create an appropriate CRS Object if specified - if (data && data.options && data.options.crs) { data.options.crs = (0, _crs_utils.getCRS)(data.options.crs); - } + } // As per https://github.com/rstudio/leaflet/pull/294#discussion_r79584810 + - // As per https://github.com/rstudio/leaflet/pull/294#discussion_r79584810 if (map) { map.remove(); + map = function () { return; }(); // undefine map + } if (data.options.mapFactory && typeof data.options.mapFactory === "function") { map = data.options.mapFactory(el, data.options); } else { - map = _leaflet2.default.map(el, data.options); + map = _leaflet2["default"].map(el, data.options); } - preventUnintendedZoomOnScroll(map); + preventUnintendedZoomOnScroll(map); // Store some state in the map object - // Store some state in the map object map.leafletr = { // Has the map ever rendered successfully? hasRendered: false, // Data to be rendered when resize is called with area != 0 pendingRenderData: null - }; - - // Check if the map is rendered statically (no output binding) - if (_htmlwidgets2.default.shinyMode && /\bshiny-bound-output\b/.test(el.className)) { + }; // Check if the map is rendered statically (no output binding) - map.id = el.id; + if (_htmlwidgets2["default"].shinyMode && /\bshiny-bound-output\b/.test(el.className)) { + map.id = el.id; // Store the map on the element so we can find it later by ID - // Store the map on the element so we can find it later by ID - (0, _jquery2.default)(el).data("leaflet-map", map); + (0, _jquery2["default"])(el).data("leaflet-map", map); // When the map is clicked, send the coordinates back to the app - // When the map is clicked, send the coordinates back to the app map.on("click", function (e) { - _shiny2.default.onInputChange(map.id + "_click", { + _shiny2["default"].onInputChange(map.id + "_click", { lat: e.latlng.lat, lng: e.latlng.lng, ".nonce": Math.random() // Force reactivity if lat/lng hasn't changed + }); }); - var groupTimerId = null; - map.on("moveend", function (e) { updateBounds(e.target); }).on("layeradd layerremove", function (e) { @@ -662,13 +678,16 @@ _htmlwidgets2.default.widget({ clearTimeout(groupTimerId); groupTimerId = null; } + groupTimerId = setTimeout(function () { groupTimerId = null; - _shiny2.default.onInputChange(map.id + "_groups", map.layerManager.getVisibleGroups()); + + _shiny2["default"].onInputChange(map.id + "_groups", map.layerManager.getVisibleGroups()); }, 100); } }); } + this.doRenderValue(data, map); }, doRenderValue: function doRenderValue(data, map) { @@ -678,75 +697,83 @@ _htmlwidgets2.default.widget({ // are off. Therefore we wait until the map is actually showing to // render the value (we rely on the resize() callback being invoked // at the appropriate time). - // - // There may be an issue with leafletProxy() calls being made while - // the map is not being viewed--not sure what the right solution is - // there. if (el.offsetWidth === 0 || el.offsetHeight === 0) { map.leafletr.pendingRenderData = data; return; } - map.leafletr.pendingRenderData = null; - // Merge data options into defaults - var options = _jquery2.default.extend({ zoomToLimits: "always" }, data.options); + map.leafletr.pendingRenderData = null; // Merge data options into defaults + + var options = _jquery2["default"].extend({ + zoomToLimits: "always" + }, data.options); if (!map.layerManager) { - map.controls = new _controlStore2.default(map); - map.layerManager = new _layerManager2.default(map); + map.controls = new _controlStore2["default"](map); + map.layerManager = new _layerManager2["default"](map); } else { map.controls.clear(); map.layerManager.clear(); } var explicitView = false; + if (data.setView) { explicitView = true; map.setView.apply(map, data.setView); } + if (data.fitBounds) { explicitView = true; methods.fitBounds.apply(map, data.fitBounds); } + if (data.flyTo) { if (!explicitView && !map.leafletr.hasRendered) { // must be done to give a initial starting point map.fitWorld(); } + explicitView = true; map.flyTo.apply(map, data.flyTo); } + if (data.flyToBounds) { if (!explicitView && !map.leafletr.hasRendered) { // must be done to give a initial starting point map.fitWorld(); } + explicitView = true; methods.flyToBounds.apply(map, data.flyToBounds); } + if (data.options.center) { explicitView = true; - } - - // Returns true if the zoomToLimits option says that the map should be + } // Returns true if the zoomToLimits option says that the map should be // zoomed to map elements. + + function needsZoom() { return options.zoomToLimits === "always" || options.zoomToLimits === "first" && !map.leafletr.hasRendered; } if (!explicitView && needsZoom() && !map.getZoom()) { - if (data.limits && !_jquery2.default.isEmptyObject(data.limits)) { + if (data.limits && !_jquery2["default"].isEmptyObject(data.limits)) { // Use the natural limits of what's being drawn on the map // If the size of the bounding box is 0, leaflet gets all weird var pad = 0.006; + if (data.limits.lat[0] === data.limits.lat[1]) { data.limits.lat[0] = data.limits.lat[0] - pad; data.limits.lat[1] = data.limits.lat[1] + pad; } + if (data.limits.lng[0] === data.limits.lng[1]) { data.limits.lng[0] = data.limits.lng[0] - pad; data.limits.lng[1] = data.limits.lng[1] + pad; } + map.fitBounds([[data.limits.lat[0], data.limits.lng[0]], [data.limits.lat[1], data.limits.lng[1]]]); } else { map.fitWorld(); @@ -760,7 +787,7 @@ _htmlwidgets2.default.widget({ map.leafletr.hasRendered = true; - if (_htmlwidgets2.default.shinyMode) { + if (_htmlwidgets2["default"].shinyMode) { setTimeout(function () { updateBounds(map); }, 1); @@ -769,6 +796,7 @@ _htmlwidgets2.default.widget({ resize: function resize(width, height) { if (map) { map.invalidateSize(); + if (map.leafletr.pendingRenderData) { this.doRenderValue(map.leafletr.pendingRenderData, map); } @@ -778,36 +806,52 @@ _htmlwidgets2.default.widget({ } }); -if (_htmlwidgets2.default.shinyMode) { - _shiny2.default.addCustomMessageHandler("leaflet-calls", function (data) { +if (_htmlwidgets2["default"].shinyMode) { + _shiny2["default"].addCustomMessageHandler("leaflet-calls", function (data) { var id = data.id; var el = document.getElementById(id); - var map = el ? (0, _jquery2.default)(el).data("leaflet-map") : null; + var map = el ? (0, _jquery2["default"])(el).data("leaflet-map") : null; + if (!map) { (0, _util.log)("Couldn't find map with id " + id); return; + } // If the map has not rendered, stash the proposed `leafletProxy()` calls + // in `pendingRenderData.calls` to be run on display via `doRenderValue()`. + // This is necessary if the map has not been rendered. + // If new pendingRenderData is set via a new `leaflet()`, the previous calls will be discarded. + + + if (!map.leafletr.hasRendered) { + map.leafletr.pendingRenderData.calls = map.leafletr.pendingRenderData.calls.concat(data.calls); + return; } for (var i = 0; i < data.calls.length; i++) { var call = data.calls[i]; + var args = call.args; + + for (var _i = 0; _i < call.evals.length; _i++) { + window.HTMLWidgets.evaluateStringMember(args, call.evals[_i]); + } + if (call.dependencies) { - _shiny2.default.renderDependencies(call.dependencies); + _shiny2["default"].renderDependencies(call.dependencies); } - if (methods[call.method]) methods[call.method].apply(map, call.args);else (0, _util.log)("Unknown method " + call.method); + + if (methods[call.method]) methods[call.method].apply(map, args);else (0, _util.log)("Unknown method " + call.method); } }); } },{"./cluster-layer-store":1,"./control-store":2,"./crs_utils":3,"./dataframe":4,"./fixup-default-icon":5,"./fixup-default-tooltip":6,"./fixup-url-protocol":7,"./global/htmlwidgets":8,"./global/jquery":9,"./global/leaflet":10,"./global/shiny":12,"./layer-manager":14,"./methods":15,"./util":17}],14:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { value: true }); - -var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }(); +exports["default"] = undefined; var _jquery = require("./global/jquery"); @@ -819,40 +863,40 @@ var _leaflet2 = _interopRequireDefault(_leaflet); var _util = require("./util"); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } } -var LayerManager = function () { +function _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } + +function _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; } + +var LayerManager = /*#__PURE__*/function () { function LayerManager(map) { _classCallCheck(this, LayerManager); - this._map = map; + this._map = map; // BEGIN layer indices + // {: {: layer}} - // BEGIN layer indices + this._byGroup = {}; // {: {: layer}} - // {: {: layer}} - this._byGroup = {}; - // {: {: layer}} - this._byCategory = {}; - // {: layer} - this._byLayerId = {}; - // {: { + this._byCategory = {}; // {: layer} + + this._byLayerId = {}; // {: { // "group": , // "layerId": , // "category": , // "container": // } // } - this._byStamp = {}; - // {: {: [, , ...], ...}} - this._byCrosstalkGroup = {}; - // END layer indices + this._byStamp = {}; // {: {: [, , ...], ...}} + this._byCrosstalkGroup = {}; // END layer indices // {: L.layerGroup} - this._categoryContainers = {}; - // {: L.layerGroup} + + this._categoryContainers = {}; // {: L.layerGroup} + this._groupContainers = {}; } @@ -864,45 +908,44 @@ var LayerManager = function () { // Was a group provided? var hasId = typeof layerId === "string"; var grouped = typeof group === "string"; - - var stamp = _leaflet2.default.Util.stamp(layer) + ""; - - // This will be the default layer group to add the layer to. + var stamp = _leaflet2["default"].Util.stamp(layer) + ""; // This will be the default layer group to add the layer to. // We may overwrite this let before using it (i.e. if a group is assigned). // This one liner creates the _categoryContainers[category] entry if it // doesn't already exist. - var container = this._categoryContainers[category] = this._categoryContainers[category] || _leaflet2.default.layerGroup().addTo(this._map); + + var container = this._categoryContainers[category] = this._categoryContainers[category] || _leaflet2["default"].layerGroup().addTo(this._map); var oldLayer = null; + if (hasId) { // First, remove any layer with the same category and layerId var prefixedLayerId = this._layerIdKey(category, layerId); + oldLayer = this._byLayerId[prefixedLayerId]; + if (oldLayer) { this._removeLayer(oldLayer); - } + } // Update layerId index + - // Update layerId index this._byLayerId[prefixedLayerId] = layer; - } + } // Update group index + - // Update group index if (grouped) { this._byGroup[group] = this._byGroup[group] || {}; - this._byGroup[group][stamp] = layer; - - // Since a group is assigned, don't add the layer to the category's layer + this._byGroup[group][stamp] = layer; // Since a group is assigned, don't add the layer to the category's layer // group; instead, use the group's layer group. // This one liner creates the _groupContainers[group] entry if it doesn't // already exist. + container = this.getLayerGroup(group, true); - } + } // Update category index + - // Update category index this._byCategory[category] = this._byCategory[category] || {}; - this._byCategory[category][stamp] = layer; + this._byCategory[category][stamp] = layer; // Update stamp index - // Update stamp index var layerInfo = this._byStamp[stamp] = { layer: layer, group: group, @@ -912,9 +955,8 @@ var LayerManager = function () { category: category, container: container, hidden: false - }; + }; // Update crosstalk group index - // Update crosstalk group index if (ctGroup) { if (layer.setStyle) { // Need to save this info so we know what to set opacity to later @@ -923,6 +965,7 @@ var LayerManager = function () { } var ctg = this._byCrosstalkGroup[ctGroup]; + if (!ctg) { ctg = this._byCrosstalkGroup[ctGroup] = {}; var crosstalk = global.crosstalk; @@ -930,64 +973,80 @@ var LayerManager = function () { var handleFilter = function handleFilter(e) { if (!e.value) { var groupKeys = Object.keys(ctg); + for (var i = 0; i < groupKeys.length; i++) { var key = groupKeys[i]; var _layerInfo = _this._byStamp[ctg[key]]; + _this._setVisibility(_layerInfo, true); } } else { var selectedKeys = {}; + for (var _i = 0; _i < e.value.length; _i++) { selectedKeys[e.value[_i]] = true; } + var _groupKeys = Object.keys(ctg); + for (var _i2 = 0; _i2 < _groupKeys.length; _i2++) { var _key = _groupKeys[_i2]; var _layerInfo2 = _this._byStamp[ctg[_key]]; + _this._setVisibility(_layerInfo2, selectedKeys[_groupKeys[_i2]]); } } }; + var filterHandle = new crosstalk.FilterHandle(ctGroup); filterHandle.on("change", handleFilter); var handleSelection = function handleSelection(e) { if (!e.value || !e.value.length) { var groupKeys = Object.keys(ctg); + for (var i = 0; i < groupKeys.length; i++) { var key = groupKeys[i]; var _layerInfo3 = _this._byStamp[ctg[key]]; + _this._setOpacity(_layerInfo3, 1.0); } } else { var selectedKeys = {}; + for (var _i3 = 0; _i3 < e.value.length; _i3++) { selectedKeys[e.value[_i3]] = true; } + var _groupKeys2 = Object.keys(ctg); + for (var _i4 = 0; _i4 < _groupKeys2.length; _i4++) { var _key2 = _groupKeys2[_i4]; var _layerInfo4 = _this._byStamp[ctg[_key2]]; + _this._setOpacity(_layerInfo4, selectedKeys[_groupKeys2[_i4]] ? 1.0 : 0.2); } } }; + var selHandle = new crosstalk.SelectionHandle(ctGroup); selHandle.on("change", handleSelection); - setTimeout(function () { - handleFilter({ value: filterHandle.filteredKeys }); - handleSelection({ value: selHandle.value }); + handleFilter({ + value: filterHandle.filteredKeys + }); + handleSelection({ + value: selHandle.value + }); }, 100); } if (!ctg[ctKey]) ctg[ctKey] = []; ctg[ctKey].push(stamp); - } + } // Add to container - // Add to container - if (!layerInfo.hidden) container.addLayer(layer); + if (!layerInfo.hidden) container.addLayer(layer); return oldLayer; } }, { @@ -996,18 +1055,17 @@ var LayerManager = function () { var _this2 = this; /* eslint-disable no-console */ - // For each Crosstalk group... Object.keys(this._byCrosstalkGroup).forEach(function (ctGroupName) { var ctg = _this2._byCrosstalkGroup[ctGroupName]; - var selection = []; - // ...iterate over each Crosstalk key (each of which may have multiple + var selection = []; // ...iterate over each Crosstalk key (each of which may have multiple // layers)... + Object.keys(ctg).forEach(function (ctKey) { // ...and for each layer... ctg[ctKey].forEach(function (stamp) { - var layerInfo = _this2._byStamp[stamp]; - // ...if it's something with a point... + var layerInfo = _this2._byStamp[stamp]; // ...if it's something with a point... + if (layerInfo.layer.getLatLng) { // ... and it's inside the selection bounds... // TODO: Use pixel containment, not lat/lng containment @@ -1064,8 +1122,9 @@ var LayerManager = function () { var _this3 = this; // Find layer info - _jquery2.default.each((0, _util.asArray)(layerIds), function (i, layerId) { + _jquery2["default"].each((0, _util.asArray)(layerIds), function (i, layerId) { var layer = _this3._byLayerId[_this3._layerIdKey(category, layerId)]; + if (layer) { _this3._removeLayer(layer); } @@ -1078,17 +1137,20 @@ var LayerManager = function () { // Find all layers in _byCategory[category] var catTable = this._byCategory[category]; + if (!catTable) { return false; - } - - // Remove all layers. Make copy of keys to avoid mutating the collection + } // Remove all layers. Make copy of keys to avoid mutating the collection // behind the iterator you're accessing. + + var stamps = []; - _jquery2.default.each(catTable, function (k, v) { + + _jquery2["default"].each(catTable, function (k, v) { stamps.push(k); }); - _jquery2.default.each(stamps, function (i, stamp) { + + _jquery2["default"].each(stamps, function (i, stamp) { _this4._removeLayer(stamp); }); } @@ -1096,12 +1158,14 @@ var LayerManager = function () { key: "getLayerGroup", value: function getLayerGroup(group, ensureExists) { var g = this._groupContainers[group]; + if (ensureExists && !g) { this._byGroup[group] = this._byGroup[group] || {}; - g = this._groupContainers[group] = _leaflet2.default.featureGroup(); + g = this._groupContainers[group] = _leaflet2["default"].featureGroup(); g.groupname = group; g.addTo(this._map); } + return g; } }, { @@ -1115,20 +1179,24 @@ var LayerManager = function () { var _this5 = this; var result = []; - _jquery2.default.each(this._groupContainers, function (k, v) { + + _jquery2["default"].each(this._groupContainers, function (k, v) { if (_this5._map.hasLayer(v)) { result.push(k); } }); + return result; } }, { key: "getAllGroupNames", value: function getAllGroupNames() { var result = []; - _jquery2.default.each(this._groupContainers, function (k, v) { + + _jquery2["default"].each(this._groupContainers, function (k, v) { result.push(k); }); + return result; } }, { @@ -1138,17 +1206,20 @@ var LayerManager = function () { // Find all layers in _byGroup[group] var groupTable = this._byGroup[group]; + if (!groupTable) { return false; - } - - // Remove all layers. Make copy of keys to avoid mutating the collection + } // Remove all layers. Make copy of keys to avoid mutating the collection // behind the iterator you're accessing. + + var stamps = []; - _jquery2.default.each(groupTable, function (k, v) { + + _jquery2["default"].each(groupTable, function (k, v) { stamps.push(k); }); - _jquery2.default.each(stamps, function (i, stamp) { + + _jquery2["default"].each(stamps, function (i, stamp) { _this6._removeLayer(stamp); }); } @@ -1157,46 +1228,58 @@ var LayerManager = function () { value: function clear() { function clearLayerGroup(key, layerGroup) { layerGroup.clearLayers(); - } - // Clear all indices and layerGroups + } // Clear all indices and layerGroups + + this._byGroup = {}; this._byCategory = {}; this._byLayerId = {}; this._byStamp = {}; this._byCrosstalkGroup = {}; - _jquery2.default.each(this._categoryContainers, clearLayerGroup); + + _jquery2["default"].each(this._categoryContainers, clearLayerGroup); + this._categoryContainers = {}; - _jquery2.default.each(this._groupContainers, clearLayerGroup); + + _jquery2["default"].each(this._groupContainers, clearLayerGroup); + this._groupContainers = {}; } }, { key: "_removeLayer", value: function _removeLayer(layer) { - var stamp = void 0; + var stamp; + if (typeof layer === "string") { stamp = layer; } else { - stamp = _leaflet2.default.Util.stamp(layer); + stamp = _leaflet2["default"].Util.stamp(layer); } var layerInfo = this._byStamp[stamp]; + if (!layerInfo) { return false; } layerInfo.container.removeLayer(stamp); + if (typeof layerInfo.group === "string") { delete this._byGroup[layerInfo.group][stamp]; } + if (typeof layerInfo.layerId === "string") { delete this._byLayerId[this._layerIdKey(layerInfo.category, layerInfo.layerId)]; } + delete this._byCategory[layerInfo.category][stamp]; delete this._byStamp[stamp]; + if (layerInfo.ctGroup) { var ctGroup = this._byCrosstalkGroup[layerInfo.ctGroup]; var layersForKey = ctGroup[layerInfo.ctKey]; var idx = layersForKey ? layersForKey.indexOf(stamp) : -1; + if (idx >= 0) { if (layersForKey.length === 1) { delete ctGroup[layerInfo.ctKey]; @@ -1216,12 +1299,12 @@ var LayerManager = function () { return LayerManager; }(); -exports.default = LayerManager; +exports["default"] = LayerManager; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{"./global/jquery":9,"./global/leaflet":10,"./util":17}],15:[function(require,module,exports){ -(function (global){ +(function (global){(function (){ "use strict"; Object.defineProperty(exports, "__esModule", { @@ -1260,30 +1343,38 @@ var _mipmapper = require("./mipmapper"); var _mipmapper2 = _interopRequireDefault(_mipmapper); -function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { default: obj }; } +function _interopRequireDefault(obj) { return obj && obj.__esModule ? obj : { "default": obj }; } var methods = {}; -exports.default = methods; - +exports["default"] = methods; function mouseHandler(mapId, layerId, group, eventName, extraInfo) { return function (e) { - if (!_htmlwidgets2.default.shinyMode) return; - + if (!_htmlwidgets2["default"].shinyMode) return; var latLng = e.target.getLatLng ? e.target.getLatLng() : e.latlng; + if (latLng) { // retrieve only lat, lon values to remove prototype // and extra parameters added by 3rd party modules // these objects are for json serialization, not javascript - var latLngVal = _leaflet2.default.latLng(latLng); // make sure it has consistent shape - latLng = { lat: latLngVal.lat, lng: latLngVal.lng }; + var latLngVal = _leaflet2["default"].latLng(latLng); // make sure it has consistent shape + + + latLng = { + lat: latLngVal.lat, + lng: latLngVal.lng + }; } - var eventInfo = _jquery2.default.extend({ + + var eventInfo = _jquery2["default"].extend({ id: layerId, ".nonce": Math.random() // force reactivity - }, group !== null ? { group: group } : null, latLng, extraInfo); - _shiny2.default.onInputChange(mapId + "_" + eventName, eventInfo); + }, group !== null ? { + group: group + } : null, latLng, extraInfo); + + _shiny2["default"].onInputChange(mapId + "_" + eventName, eventInfo); }; } @@ -1292,7 +1383,7 @@ methods.mouseHandler = mouseHandler; methods.clearGroup = function (group) { var _this = this; - _jquery2.default.each((0, _util.asArray)(group), function (i, v) { + _jquery2["default"].each((0, _util.asArray)(group), function (i, v) { _this.layerManager.clearGroup(v); }); }; @@ -1320,12 +1411,13 @@ methods.setMaxBounds = function (lat1, lng1, lat2, lng2) { methods.addPopups = function (lat, lng, popup, layerId, group, options) { var _this2 = this; - var df = new _dataframe2.default().col("lat", lat).col("lng", lng).col("popup", popup).col("layerId", layerId).col("group", group).cbind(options); + var df = new _dataframe2["default"]().col("lat", lat).col("lng", lng).col("popup", popup).col("layerId", layerId).col("group", group).cbind(options); var _loop = function _loop(i) { - if (_jquery2.default.isNumeric(df.get(i, "lat")) && _jquery2.default.isNumeric(df.get(i, "lng"))) { + if (_jquery2["default"].isNumeric(df.get(i, "lat")) && _jquery2["default"].isNumeric(df.get(i, "lng"))) { (function () { - var popup = _leaflet2.default.popup(df.get(i)).setLatLng([df.get(i, "lat"), df.get(i, "lng")]).setContent(df.get(i, "popup")); + var popup = _leaflet2["default"].popup(df.get(i)).setLatLng([df.get(i, "lat"), df.get(i, "lng")]).setContent(df.get(i, "popup")); + var thisId = df.get(i, "layerId"); var thisGroup = df.get(i, "group"); this.layerManager.addLayer(popup, "popup", thisId, thisGroup); @@ -1347,7 +1439,7 @@ methods.clearPopups = function () { }; methods.addTiles = function (urlTemplate, layerId, group, options) { - this.layerManager.addLayer(_leaflet2.default.tileLayer(urlTemplate, options), "tile", layerId, group); + this.layerManager.addLayer(_leaflet2["default"].tileLayer(urlTemplate, options), "tile", layerId, group); }; methods.removeTiles = function (layerId) { @@ -1362,25 +1454,26 @@ methods.addWMSTiles = function (baseUrl, layerId, group, options) { if (options && options.crs) { options.crs = (0, _crs_utils.getCRS)(options.crs); } - this.layerManager.addLayer(_leaflet2.default.tileLayer.wms(baseUrl, options), "tile", layerId, group); -}; -// Given: + this.layerManager.addLayer(_leaflet2["default"].tileLayer.wms(baseUrl, options), "tile", layerId, group); +}; // Given: // {data: ["a", "b", "c"], index: [0, 1, 0, 2]} // returns: // ["a", "b", "a", "c"] + + function unpackStrings(iconset) { if (!iconset) { return iconset; } + if (typeof iconset.index === "undefined") { return iconset; } iconset.data = (0, _util.asArray)(iconset.data); iconset.index = (0, _util.asArray)(iconset.index); - - return _jquery2.default.map(iconset.index, function (e, i) { + return _jquery2["default"].map(iconset.index, function (e, i) { return iconset.data[e]; }); } @@ -1391,30 +1484,39 @@ function addMarkers(map, df, group, clusterOptions, clusterId, markerFunc) { var clusterGroup = this.layerManager.getLayer("cluster", clusterId), cluster = clusterOptions !== null; + if (cluster && !clusterGroup) { - clusterGroup = _leaflet2.default.markerClusterGroup.layerSupport(clusterOptions); + clusterGroup = _leaflet2["default"].markerClusterGroup.layerSupport(clusterOptions); + if (clusterOptions.freezeAtZoom) { var freezeAtZoom = clusterOptions.freezeAtZoom; delete clusterOptions.freezeAtZoom; clusterGroup.freezeAtZoom(freezeAtZoom); } - clusterGroup.clusterLayerStore = new _clusterLayerStore2.default(clusterGroup); + + clusterGroup.clusterLayerStore = new _clusterLayerStore2["default"](clusterGroup); } - var extraInfo = cluster ? { clusterId: clusterId } : {}; + + var extraInfo = cluster ? { + clusterId: clusterId + } : {}; var _loop2 = function _loop2(i) { - if (_jquery2.default.isNumeric(df.get(i, "lat")) && _jquery2.default.isNumeric(df.get(i, "lng"))) { + if (_jquery2["default"].isNumeric(df.get(i, "lat")) && _jquery2["default"].isNumeric(df.get(i, "lng"))) { (function () { var marker = markerFunc(df, i); var thisId = df.get(i, "layerId"); var thisGroup = cluster ? null : df.get(i, "group"); + if (cluster) { clusterGroup.clusterLayerStore.add(marker, thisId); } else { this.layerManager.addLayer(marker, "marker", thisId, thisGroup, df.get(i, "ctGroup", true), df.get(i, "ctKey", true)); } + var popup = df.get(i, "popup"); var popupOptions = df.get(i, "popupOptions"); + if (popup !== null) { if (popupOptions !== null) { marker.bindPopup(popup, popupOptions); @@ -1422,8 +1524,10 @@ function addMarkers(map, df, group, clusterOptions, clusterId, markerFunc) { marker.bindPopup(popup); } } + var label = df.get(i, "label"); var labelOptions = df.get(i, "labelOptions"); + if (label !== null) { if (labelOptions !== null) { if (labelOptions.permanent) { @@ -1435,6 +1539,7 @@ function addMarkers(map, df, group, clusterOptions, clusterId, markerFunc) { marker.bindTooltip(label); } } + marker.on("click", mouseHandler(this.id, thisId, thisGroup, "marker_click", extraInfo), this); marker.on("mouseover", mouseHandler(this.id, thisId, thisGroup, "marker_mouseover", extraInfo), this); marker.on("mouseout", mouseHandler(this.id, thisId, thisGroup, "marker_mouseout", extraInfo), this); @@ -1456,97 +1561,95 @@ function addMarkers(map, df, group, clusterOptions, clusterId, markerFunc) { methods.addGenericMarkers = addMarkers; methods.addMarkers = function (lat, lng, icon, layerId, group, options, popup, popupOptions, clusterOptions, clusterId, label, labelOptions, crosstalkOptions) { - var icondf = void 0; - var getIcon = void 0; + var icondf; + var getIcon; if (icon) { // Unpack icons icon.iconUrl = unpackStrings(icon.iconUrl); icon.iconRetinaUrl = unpackStrings(icon.iconRetinaUrl); icon.shadowUrl = unpackStrings(icon.shadowUrl); - icon.shadowRetinaUrl = unpackStrings(icon.shadowRetinaUrl); - - // This cbinds the icon URLs and any other icon options; they're all + icon.shadowRetinaUrl = unpackStrings(icon.shadowRetinaUrl); // This cbinds the icon URLs and any other icon options; they're all // present on the icon object. - icondf = new _dataframe2.default().cbind(icon); - // Constructs an icon from a specified row of the icon dataframe. + icondf = new _dataframe2["default"]().cbind(icon); // Constructs an icon from a specified row of the icon dataframe. + getIcon = function getIcon(i) { var opts = icondf.get(i); - if (!opts.iconUrl) { - return new _leaflet2.default.Icon.Default(); - } - // Composite options (like points or sizes) are passed from R with each + if (!opts.iconUrl) { + return new _leaflet2["default"].Icon.Default(); + } // Composite options (like points or sizes) are passed from R with each // individual component as its own option. We need to combine them now // into their composite form. + + if (opts.iconWidth) { opts.iconSize = [opts.iconWidth, opts.iconHeight]; } + if (opts.shadowWidth) { opts.shadowSize = [opts.shadowWidth, opts.shadowHeight]; } + if (opts.iconAnchorX) { opts.iconAnchor = [opts.iconAnchorX, opts.iconAnchorY]; } + if (opts.shadowAnchorX) { opts.shadowAnchor = [opts.shadowAnchorX, opts.shadowAnchorY]; } + if (opts.popupAnchorX) { opts.popupAnchor = [opts.popupAnchorX, opts.popupAnchorY]; } - return new _leaflet2.default.Icon(opts); + return new _leaflet2["default"].Icon(opts); }; } - if (!(_jquery2.default.isEmptyObject(lat) || _jquery2.default.isEmptyObject(lng)) || _jquery2.default.isNumeric(lat) && _jquery2.default.isNumeric(lng)) { - - var df = new _dataframe2.default().col("lat", lat).col("lng", lng).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(options).cbind(crosstalkOptions || {}); - + if (!(_jquery2["default"].isEmptyObject(lat) || _jquery2["default"].isEmptyObject(lng)) || _jquery2["default"].isNumeric(lat) && _jquery2["default"].isNumeric(lng)) { + var df = new _dataframe2["default"]().col("lat", lat).col("lng", lng).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(options).cbind(crosstalkOptions || {}); if (icon) icondf.effectiveLength = df.nrow(); - addMarkers(this, df, group, clusterOptions, clusterId, function (df, i) { var options = df.get(i); if (icon) options.icon = getIcon(i); - return _leaflet2.default.marker([df.get(i, "lat"), df.get(i, "lng")], options); + return _leaflet2["default"].marker([df.get(i, "lat"), df.get(i, "lng")], options); }); } }; methods.addAwesomeMarkers = function (lat, lng, icon, layerId, group, options, popup, popupOptions, clusterOptions, clusterId, label, labelOptions, crosstalkOptions) { - var icondf = void 0; - var getIcon = void 0; - if (icon) { + var icondf; + var getIcon; + if (icon) { // This cbinds the icon URLs and any other icon options; they're all // present on the icon object. - icondf = new _dataframe2.default().cbind(icon); + icondf = new _dataframe2["default"]().cbind(icon); // Constructs an icon from a specified row of the icon dataframe. - // Constructs an icon from a specified row of the icon dataframe. getIcon = function getIcon(i) { var opts = icondf.get(i); + if (!opts) { - return new _leaflet2.default.AwesomeMarkers.icon(); + return new _leaflet2["default"].AwesomeMarkers.icon(); } if (opts.squareMarker) { opts.className = "awesome-marker awesome-marker-square"; } - return new _leaflet2.default.AwesomeMarkers.icon(opts); + + return new _leaflet2["default"].AwesomeMarkers.icon(opts); }; } - if (!(_jquery2.default.isEmptyObject(lat) || _jquery2.default.isEmptyObject(lng)) || _jquery2.default.isNumeric(lat) && _jquery2.default.isNumeric(lng)) { - - var df = new _dataframe2.default().col("lat", lat).col("lng", lng).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(options).cbind(crosstalkOptions || {}); - + if (!(_jquery2["default"].isEmptyObject(lat) || _jquery2["default"].isEmptyObject(lng)) || _jquery2["default"].isNumeric(lat) && _jquery2["default"].isNumeric(lng)) { + var df = new _dataframe2["default"]().col("lat", lat).col("lng", lng).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(options).cbind(crosstalkOptions || {}); if (icon) icondf.effectiveLength = df.nrow(); - addMarkers(this, df, group, clusterOptions, clusterId, function (df, i) { var options = df.get(i); if (icon) options.icon = getIcon(i); - return _leaflet2.default.marker([df.get(i, "lat"), df.get(i, "lng")], options); + return _leaflet2["default"].marker([df.get(i, "lat"), df.get(i, "lng")], options); }); } }; @@ -1555,13 +1658,16 @@ function addLayers(map, category, df, layerFunc) { var _loop3 = function _loop3(i) { (function () { var layer = layerFunc(df, i); - if (!_jquery2.default.isEmptyObject(layer)) { + + if (!_jquery2["default"].isEmptyObject(layer)) { var thisId = df.get(i, "layerId"); var thisGroup = df.get(i, "group"); this.layerManager.addLayer(layer, category, thisId, thisGroup, df.get(i, "ctGroup", true), df.get(i, "ctKey", true)); + if (layer.bindPopup) { var popup = df.get(i, "popup"); var popupOptions = df.get(i, "popupOptions"); + if (popup !== null) { if (popupOptions !== null) { layer.bindPopup(popup, popupOptions); @@ -1570,9 +1676,11 @@ function addLayers(map, category, df, layerFunc) { } } } + if (layer.bindTooltip) { var label = df.get(i, "label"); var labelOptions = df.get(i, "labelOptions"); + if (label !== null) { if (labelOptions !== null) { layer.bindTooltip(label, labelOptions); @@ -1581,15 +1689,16 @@ function addLayers(map, category, df, layerFunc) { } } } + layer.on("click", mouseHandler(this.id, thisId, thisGroup, category + "_click"), this); layer.on("mouseover", mouseHandler(this.id, thisId, thisGroup, category + "_mouseover"), this); layer.on("mouseout", mouseHandler(this.id, thisId, thisGroup, category + "_mouseout"), this); var highlightStyle = df.get(i, "highlightOptions"); - if (!_jquery2.default.isEmptyObject(highlightStyle)) { - + if (!_jquery2["default"].isEmptyObject(highlightStyle)) { var defaultStyle = {}; - _jquery2.default.each(highlightStyle, function (k, v) { + + _jquery2["default"].each(highlightStyle, function (k, v) { if (k != "bringToFront" && k != "sendToBack") { if (df.get(i, k)) { defaultStyle[k] = df.get(i, k); @@ -1599,12 +1708,14 @@ function addLayers(map, category, df, layerFunc) { layer.on("mouseover", function (e) { this.setStyle(highlightStyle); + if (highlightStyle.bringToFront) { this.bringToFront(); } }); layer.on("mouseout", function (e) { this.setStyle(defaultStyle); + if (highlightStyle.sendToBack) { this.bringToBack(); } @@ -1622,12 +1733,11 @@ function addLayers(map, category, df, layerFunc) { methods.addGenericLayers = addLayers; methods.addCircles = function (lat, lng, radius, layerId, group, options, popup, popupOptions, label, labelOptions, highlightOptions, crosstalkOptions) { - if (!(_jquery2.default.isEmptyObject(lat) || _jquery2.default.isEmptyObject(lng)) || _jquery2.default.isNumeric(lat) && _jquery2.default.isNumeric(lng)) { - var df = new _dataframe2.default().col("lat", lat).col("lng", lng).col("radius", radius).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options).cbind(crosstalkOptions || {}); - + if (!(_jquery2["default"].isEmptyObject(lat) || _jquery2["default"].isEmptyObject(lng)) || _jquery2["default"].isNumeric(lat) && _jquery2["default"].isNumeric(lng)) { + var df = new _dataframe2["default"]().col("lat", lat).col("lng", lng).col("radius", radius).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options).cbind(crosstalkOptions || {}); addLayers(this, "shape", df, function (df, i) { - if (_jquery2.default.isNumeric(df.get(i, "lat")) && _jquery2.default.isNumeric(df.get(i, "lng")) && _jquery2.default.isNumeric(df.get(i, "radius"))) { - return _leaflet2.default.circle([df.get(i, "lat"), df.get(i, "lng")], df.get(i, "radius"), df.get(i)); + if (_jquery2["default"].isNumeric(df.get(i, "lat")) && _jquery2["default"].isNumeric(df.get(i, "lng")) && _jquery2["default"].isNumeric(df.get(i, "radius"))) { + return _leaflet2["default"].circle([df.get(i, "lat"), df.get(i, "lng")], df.get(i, "radius"), df.get(i)); } else { return null; } @@ -1636,32 +1746,32 @@ methods.addCircles = function (lat, lng, radius, layerId, group, options, popup, }; methods.addCircleMarkers = function (lat, lng, radius, layerId, group, options, clusterOptions, clusterId, popup, popupOptions, label, labelOptions, crosstalkOptions) { - if (!(_jquery2.default.isEmptyObject(lat) || _jquery2.default.isEmptyObject(lng)) || _jquery2.default.isNumeric(lat) && _jquery2.default.isNumeric(lng)) { - var df = new _dataframe2.default().col("lat", lat).col("lng", lng).col("radius", radius).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(crosstalkOptions || {}).cbind(options); - + if (!(_jquery2["default"].isEmptyObject(lat) || _jquery2["default"].isEmptyObject(lng)) || _jquery2["default"].isNumeric(lat) && _jquery2["default"].isNumeric(lng)) { + var df = new _dataframe2["default"]().col("lat", lat).col("lng", lng).col("radius", radius).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).cbind(crosstalkOptions || {}).cbind(options); addMarkers(this, df, group, clusterOptions, clusterId, function (df, i) { - return _leaflet2.default.circleMarker([df.get(i, "lat"), df.get(i, "lng")], df.get(i)); + return _leaflet2["default"].circleMarker([df.get(i, "lat"), df.get(i, "lng")], df.get(i)); }); } }; - /* * @param lat Array of arrays of latitude coordinates for polylines * @param lng Array of arrays of longitude coordinates for polylines */ + + methods.addPolylines = function (polygons, layerId, group, options, popup, popupOptions, label, labelOptions, highlightOptions) { if (polygons.length > 0) { - var df = new _dataframe2.default().col("shapes", polygons).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); - + var df = new _dataframe2["default"]().col("shapes", polygons).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); addLayers(this, "shape", df, function (df, i) { var shapes = df.get(i, "shapes"); shapes = shapes.map(function (shape) { - return _htmlwidgets2.default.dataframeToD3(shape[0]); + return _htmlwidgets2["default"].dataframeToD3(shape[0]); }); + if (shapes.length > 1) { - return _leaflet2.default.polyline(shapes, df.get(i)); + return _leaflet2["default"].polyline(shapes, df.get(i)); } else { - return _leaflet2.default.polyline(shapes[0], df.get(i)); + return _leaflet2["default"].polyline(shapes[0], df.get(i)); } }); } @@ -1698,25 +1808,24 @@ methods.clearShapes = function () { }; methods.addRectangles = function (lat1, lng1, lat2, lng2, layerId, group, options, popup, popupOptions, label, labelOptions, highlightOptions) { - var df = new _dataframe2.default().col("lat1", lat1).col("lng1", lng1).col("lat2", lat2).col("lng2", lng2).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); - + var df = new _dataframe2["default"]().col("lat1", lat1).col("lng1", lng1).col("lat2", lat2).col("lng2", lng2).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); addLayers(this, "shape", df, function (df, i) { - if (_jquery2.default.isNumeric(df.get(i, "lat1")) && _jquery2.default.isNumeric(df.get(i, "lng1")) && _jquery2.default.isNumeric(df.get(i, "lat2")) && _jquery2.default.isNumeric(df.get(i, "lng2"))) { - return _leaflet2.default.rectangle([[df.get(i, "lat1"), df.get(i, "lng1")], [df.get(i, "lat2"), df.get(i, "lng2")]], df.get(i)); + if (_jquery2["default"].isNumeric(df.get(i, "lat1")) && _jquery2["default"].isNumeric(df.get(i, "lng1")) && _jquery2["default"].isNumeric(df.get(i, "lat2")) && _jquery2["default"].isNumeric(df.get(i, "lng2"))) { + return _leaflet2["default"].rectangle([[df.get(i, "lat1"), df.get(i, "lng1")], [df.get(i, "lat2"), df.get(i, "lng2")]], df.get(i)); } else { return null; } }); }; - /* * @param lat Array of arrays of latitude coordinates for polygons * @param lng Array of arrays of longitude coordinates for polygons */ + + methods.addPolygons = function (polygons, layerId, group, options, popup, popupOptions, label, labelOptions, highlightOptions) { if (polygons.length > 0) { - var df = new _dataframe2.default().col("shapes", polygons).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); - + var df = new _dataframe2["default"]().col("shapes", polygons).col("layerId", layerId).col("group", group).col("popup", popup).col("popupOptions", popupOptions).col("label", label).col("labelOptions", labelOptions).col("highlightOptions", highlightOptions).cbind(options); addLayers(this, "shape", df, function (df, i) { // This code used to use L.multiPolygon, but that caused // double-click on a multipolygon to fail to zoom in on the @@ -1724,11 +1833,11 @@ methods.addPolygons = function (polygons, layerId, group, options, popup, popupO // polygon seems to still work; complicated multipolygons // are still rendered correctly. var shapes = df.get(i, "shapes").map(function (polygon) { - return polygon.map(_htmlwidgets2.default.dataframeToD3); + return polygon.map(_htmlwidgets2["default"].dataframeToD3); }).reduce(function (acc, val) { return acc.concat(val); }, []); - return _leaflet2.default.polygon(shapes, df.get(i)); + return _leaflet2["default"].polygon(shapes, df.get(i)); }); } }; @@ -1737,16 +1846,17 @@ methods.addGeoJSON = function (data, layerId, group, style) { // This time, self is actually needed because the callbacks below need // to access both the inner and outer senses of "this" var self = this; + if (typeof data === "string") { data = JSON.parse(data); } - var globalStyle = _jquery2.default.extend({}, style, data.style || {}); + var globalStyle = _jquery2["default"].extend({}, style, data.style || {}); - var gjlayer = _leaflet2.default.geoJson(data, { + var gjlayer = _leaflet2["default"].geoJson(data, { style: function style(feature) { if (feature.style || feature.properties.style) { - return _jquery2.default.extend({}, globalStyle, feature.style, feature.properties.style); + return _jquery2["default"].extend({}, globalStyle, feature.style, feature.properties.style); } else { return globalStyle; } @@ -1763,6 +1873,7 @@ methods.addGeoJSON = function (data, layerId, group, style) { layer.on("mouseout", mouseHandler(self.id, layerId, group, "geojson_mouseout", extraInfo), this); } }); + this.layerManager.addLayer(gjlayer, "geojson", layerId, group); }; @@ -1778,16 +1889,17 @@ methods.addTopoJSON = function (data, layerId, group, style) { // This time, self is actually needed because the callbacks below need // to access both the inner and outer senses of "this" var self = this; + if (typeof data === "string") { data = JSON.parse(data); } - var globalStyle = _jquery2.default.extend({}, style, data.style || {}); + var globalStyle = _jquery2["default"].extend({}, style, data.style || {}); - var gjlayer = _leaflet2.default.geoJson(null, { + var gjlayer = _leaflet2["default"].geoJson(null, { style: function style(feature) { if (feature.style || feature.properties.style) { - return _jquery2.default.extend({}, globalStyle, feature.style, feature.properties.style); + return _jquery2["default"].extend({}, globalStyle, feature.style, feature.properties.style); } else { return globalStyle; } @@ -1804,6 +1916,7 @@ methods.addTopoJSON = function (data, layerId, group, style) { layer.on("mouseout", mouseHandler(self.id, layerId, group, "topojson_mouseout", extraInfo), this); } }); + global.omnivore.topojson.parse(data, null, gjlayer); this.layerManager.addLayer(gjlayer, "topojson", layerId, group); }; @@ -1818,38 +1931,45 @@ methods.clearTopoJSON = function () { methods.addControl = function (html, position, layerId, classes) { function onAdd(map) { - var div = _leaflet2.default.DomUtil.create("div", classes); + var div = _leaflet2["default"].DomUtil.create("div", classes); + if (typeof layerId !== "undefined" && layerId !== null) { div.setAttribute("id", layerId); } - this._div = div; - // It's possible for window.Shiny to be true but Shiny.initializeInputs to + this._div = div; // It's possible for window.Shiny to be true but Shiny.initializeInputs to // not be, when a static leaflet widget is included as part of the shiny // UI directly (not through leafletOutput or uiOutput). In this case we // don't do the normal Shiny stuff as that will all happen when Shiny // itself loads and binds the entire doc. - if (window.Shiny && _shiny2.default.initializeInputs) { - _shiny2.default.renderHtml(html, this._div); - _shiny2.default.initializeInputs(this._div); - _shiny2.default.bindAll(this._div); + if (window.Shiny && _shiny2["default"].initializeInputs) { + _shiny2["default"].renderHtml(html, this._div); + + _shiny2["default"].initializeInputs(this._div); + + _shiny2["default"].bindAll(this._div); } else { this._div.innerHTML = html; } return this._div; } + function onRemove(map) { - if (window.Shiny && _shiny2.default.unbindAll) { - _shiny2.default.unbindAll(this._div); + if (window.Shiny && _shiny2["default"].unbindAll) { + _shiny2["default"].unbindAll(this._div); } } - var Control = _leaflet2.default.Control.extend({ - options: { position: position }, + + var Control = _leaflet2["default"].Control.extend({ + options: { + position: position + }, onAdd: onAdd, onRemove: onRemove }); + this.controls.add(new Control(), layerId, html); }; @@ -1870,37 +1990,42 @@ methods.clearControls = function () { }; methods.addLegend = function (options) { - var legend = _leaflet2.default.control({ position: options.position }); - var gradSpan = void 0; + var legend = _leaflet2["default"].control({ + position: options.position + }); + + var gradSpan; legend.onAdd = function (map) { - var div = _leaflet2.default.DomUtil.create("div", options.className), + var div = _leaflet2["default"].DomUtil.create("div", options.className), colors = options.colors, labels = options.labels, legendHTML = ""; + if (options.type === "numeric") { // # Formatting constants. var singleBinHeight = 20; // The distance between tick marks, in px + var vMargin = 8; // If 1st tick mark starts at top of gradient, how // many extra px are needed for the top half of the // 1st label? (ditto for last tick mark/label) + var tickWidth = 4; // How wide should tick marks be, in px? + var labelPadding = 6; // How much distance to reserve for tick mark? // (Must be >= tickWidth) - // # Derived formatting parameters. - // What's the height of a single bin, in percentage (of gradient height)? // It might not just be 1/(n-1), if the gradient extends past the tick // marks (which can be the case for pretty cut points). - var singleBinPct = (options.extra.p_n - options.extra.p_1) / (labels.length - 1); - // Each bin is `singleBinHeight` high. How tall is the gradient? - var totalHeight = 1 / singleBinPct * singleBinHeight + 1; - // How far should the first tick be shifted down, relative to the top + + var singleBinPct = (options.extra.p_n - options.extra.p_1) / (labels.length - 1); // Each bin is `singleBinHeight` high. How tall is the gradient? + + var totalHeight = 1 / singleBinPct * singleBinHeight + 1; // How far should the first tick be shifted down, relative to the top // of the gradient? - var tickOffset = singleBinHeight / singleBinPct * options.extra.p_1; - gradSpan = (0, _jquery2.default)("").css({ + var tickOffset = singleBinHeight / singleBinPct * options.extra.p_1; + gradSpan = (0, _jquery2["default"])("").css({ "background": "linear-gradient(" + colors + ")", "opacity": options.opacity, "height": totalHeight + "px", @@ -1908,68 +2033,65 @@ methods.addLegend = function (options) { "display": "block", "margin-top": vMargin + "px" }); - var leftDiv = (0, _jquery2.default)("
    ").css("float", "left"), - rightDiv = (0, _jquery2.default)("
    ").css("float", "left"); + var leftDiv = (0, _jquery2["default"])("
    ").css("float", "left"), + rightDiv = (0, _jquery2["default"])("
    ").css("float", "left"); leftDiv.append(gradSpan); - (0, _jquery2.default)(div).append(leftDiv).append(rightDiv).append((0, _jquery2.default)("
    ")); - - // Have to attach the div to the body at this early point, so that the + (0, _jquery2["default"])(div).append(leftDiv).append(rightDiv).append((0, _jquery2["default"])("
    ")); // Have to attach the div to the body at this early point, so that the // svg text getComputedTextLength() actually works, below. - document.body.appendChild(div); + document.body.appendChild(div); var ns = "http://www.w3.org/2000/svg"; var svg = document.createElementNS(ns, "svg"); rightDiv.append(svg); var g = document.createElementNS(ns, "g"); - (0, _jquery2.default)(g).attr("transform", "translate(0, " + vMargin + ")"); - svg.appendChild(g); + (0, _jquery2["default"])(g).attr("transform", "translate(0, " + vMargin + ")"); + svg.appendChild(g); // max label width needed to set width of svg, and right-justify text - // max label width needed to set width of svg, and right-justify text - var maxLblWidth = 0; + var maxLblWidth = 0; // Create tick marks and labels - // Create tick marks and labels - _jquery2.default.each(labels, function (i, label) { + _jquery2["default"].each(labels, function (i, label) { var y = tickOffset + i * singleBinHeight + 0.5; - var thisLabel = document.createElementNS(ns, "text"); - (0, _jquery2.default)(thisLabel).text(labels[i]).attr("y", y).attr("dx", labelPadding).attr("dy", "0.5ex"); + (0, _jquery2["default"])(thisLabel).text(labels[i]).attr("y", y).attr("dx", labelPadding).attr("dy", "0.5ex"); g.appendChild(thisLabel); maxLblWidth = Math.max(maxLblWidth, thisLabel.getComputedTextLength()); - var thisTick = document.createElementNS(ns, "line"); - (0, _jquery2.default)(thisTick).attr("x1", 0).attr("x2", tickWidth).attr("y1", y).attr("y2", y).attr("stroke-width", 1); + (0, _jquery2["default"])(thisTick).attr("x1", 0).attr("x2", tickWidth).attr("y1", y).attr("y2", y).attr("stroke-width", 1); g.appendChild(thisTick); - }); + }); // Now that we know the max label width, we can right-justify - // Now that we know the max label width, we can right-justify - (0, _jquery2.default)(svg).find("text").attr("dx", labelPadding + maxLblWidth).attr("text-anchor", "end"); - // Final size for - (0, _jquery2.default)(svg).css({ + + (0, _jquery2["default"])(svg).find("text").attr("dx", labelPadding + maxLblWidth).attr("text-anchor", "end"); // Final size for + + (0, _jquery2["default"])(svg).css({ width: maxLblWidth + labelPadding + "px", height: totalHeight + vMargin * 2 + "px" }); - if (options.na_color && _jquery2.default.inArray(options.na_label, labels) < 0) { - (0, _jquery2.default)(div).append("
    " + options.na_label + "
    "); + if (options.na_color && _jquery2["default"].inArray(options.na_label, labels) < 0) { + (0, _jquery2["default"])(div).append("
    " + options.na_label + "
    "); } } else { - if (options.na_color && _jquery2.default.inArray(options.na_label, labels) < 0) { + if (options.na_color && _jquery2["default"].inArray(options.na_label, labels) < 0) { colors.push(options.na_color); labels.push(options.na_label); } + for (var i = 0; i < colors.length; i++) { legendHTML += " " + labels[i] + "
    "; } + div.innerHTML = legendHTML; } - if (options.title) (0, _jquery2.default)(div).prepend("
    " + options.title + "
    "); + + if (options.title) (0, _jquery2["default"])(div).prepend("
    " + options.title + "
    "); return div; }; if (options.group) { // Auto generate a layerID if not provided if (!options.layerId) { - options.layerId = _leaflet2.default.Util.stamp(legend); + options.layerId = _leaflet2["default"].Util.stamp(legend); } var map = this; @@ -2003,15 +2125,15 @@ methods.addLayersControl = function (baseGroups, overlayGroups, options) { // Only allow one layers control at a time methods.removeLayersControl.call(this); - var firstLayer = true; var base = {}; - _jquery2.default.each((0, _util.asArray)(baseGroups), function (i, g) { + + _jquery2["default"].each((0, _util.asArray)(baseGroups), function (i, g) { var layer = _this4.layerManager.getLayerGroup(g, true); + if (layer) { - base[g] = layer; + base[g] = layer; // Check if >1 base layers are visible; if so, hide all but the first one - // Check if >1 base layers are visible; if so, hide all but the first one if (_this4.hasLayer(layer)) { if (firstLayer) { firstLayer = false; @@ -2021,15 +2143,18 @@ methods.addLayersControl = function (baseGroups, overlayGroups, options) { } } }); + var overlay = {}; - _jquery2.default.each((0, _util.asArray)(overlayGroups), function (i, g) { + + _jquery2["default"].each((0, _util.asArray)(overlayGroups), function (i, g) { var layer = _this4.layerManager.getLayerGroup(g, true); + if (layer) { overlay[g] = layer; } }); - this.currentLayersControl = _leaflet2.default.control.layers(base, overlay, options); + this.currentLayersControl = _leaflet2["default"].control.layers(base, overlay, options); this.addControl(this.currentLayersControl); }; @@ -2041,11 +2166,11 @@ methods.removeLayersControl = function () { }; methods.addScaleBar = function (options) { - // Only allow one scale bar at a time methods.removeScaleBar.call(this); - var scaleBar = _leaflet2.default.control.scale(options).addTo(this); + var scaleBar = _leaflet2["default"].control.scale(options).addTo(this); + this.currentScaleBar = scaleBar; }; @@ -2059,8 +2184,9 @@ methods.removeScaleBar = function () { methods.hideGroup = function (group) { var _this5 = this; - _jquery2.default.each((0, _util.asArray)(group), function (i, g) { + _jquery2["default"].each((0, _util.asArray)(group), function (i, g) { var layer = _this5.layerManager.getLayerGroup(g, true); + if (layer) { _this5.removeLayer(layer); } @@ -2070,8 +2196,9 @@ methods.hideGroup = function (group) { methods.showGroup = function (group) { var _this6 = this; - _jquery2.default.each((0, _util.asArray)(group), function (i, g) { + _jquery2["default"].each((0, _util.asArray)(group), function (i, g) { var layer = _this6.layerManager.getLayerGroup(g, true); + if (layer) { _this6.addLayer(layer); } @@ -2082,26 +2209,33 @@ function setupShowHideGroupsOnZoom(map) { if (map.leafletr._hasInitializedShowHideGroups) { return; } + map.leafletr._hasInitializedShowHideGroups = true; function setVisibility(layer, visible, group) { if (visible !== map.hasLayer(layer)) { if (visible) { map.addLayer(layer); - map.fire("groupadd", { "name": group, "layer": layer }); + map.fire("groupadd", { + "name": group, + "layer": layer + }); } else { map.removeLayer(layer); - map.fire("groupremove", { "name": group, "layer": layer }); + map.fire("groupremove", { + "name": group, + "layer": layer + }); } } } function showHideGroupsOnZoom() { if (!map.layerManager) return; - var zoom = map.getZoom(); map.layerManager.getAllGroupNames().forEach(function (group) { var layer = map.layerManager.getLayerGroup(group, false); + if (layer && typeof layer.zoomLevels !== "undefined") { setVisibility(layer, layer.zoomLevels === true || layer.zoomLevels.indexOf(zoom) >= 0, group); } @@ -2115,9 +2249,10 @@ function setupShowHideGroupsOnZoom(map) { methods.setGroupOptions = function (group, options) { var _this7 = this; - _jquery2.default.each((0, _util.asArray)(group), function (i, g) { - var layer = _this7.layerManager.getLayerGroup(g, true); - // This slightly tortured check is because 0 is a valid value for zoomLevels + _jquery2["default"].each((0, _util.asArray)(group), function (i, g) { + var layer = _this7.layerManager.getLayerGroup(g, true); // This slightly tortured check is because 0 is a valid value for zoomLevels + + if (typeof options.zoomLevels !== "undefined" && options.zoomLevels !== null) { layer.zoomLevels = (0, _util.asArray)(options.zoomLevels); } @@ -2130,13 +2265,10 @@ methods.setGroupOptions = function (group, options) { methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, group) { // uri is a data URI containing an image. We want to paint this image as a // layer at (top-left) bounds[0] to (bottom-right) bounds[1]. - // We can't simply use ImageOverlay, as it uses bilinear scaling which looks // awful as you zoom in (and sometimes shifts positions or disappears). // Instead, we'll use a TileLayer.Canvas to draw pieces of the image. - // First, some helper functions. - // degree2tile converts latitude, longitude, and zoom to x and y tile // numbers. The tile numbers returned can be non-integral, as there's no // reason to expect that the lat/lng inputs are exactly on the border of two @@ -2153,28 +2285,33 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g var n = Math.pow(2, zoom); var x = (lng + 180) / 360 * n; var y = (1 - Math.log(Math.tan(latRad) + 1 / Math.cos(latRad)) / Math.PI) / 2 * n; - return { x: x, y: y }; - } - - // Given a range [from,to) and either one or two numbers, returns true if + return { + x: x, + y: y + }; + } // Given a range [from,to) and either one or two numbers, returns true if // there is any overlap between [x,x1) and the range--or if x1 is omitted, // then returns true if x is within [from,to). - function overlap(from, to, x, /* optional */x1) { + + + function overlap(from, to, x, + /* optional */ + x1) { if (arguments.length == 3) x1 = x; return x < to && x1 >= from; } function getCanvasSmoothingProperty(ctx) { var candidates = ["imageSmoothingEnabled", "mozImageSmoothingEnabled", "webkitImageSmoothingEnabled", "msImageSmoothingEnabled"]; + for (var i = 0; i < candidates.length; i++) { if (typeof ctx[candidates[i]] !== "undefined") { return candidates[i]; } } - return null; - } - // Our general strategy is to: + return null; + } // Our general strategy is to: // 1. Load the data URI in an Image() object, so we can get its pixel // dimensions and the underlying image data. (We could have done this // by not encoding as PNG at all but just send an array of RGBA values @@ -2182,26 +2319,27 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g // 2. Create a hidden canvas that we use just to extract the image data // from the Image (using Context2D.getImageData()). // 3. Create a TileLayer.Canvas and add it to the map. - // We want to synchronously create and attach the TileLayer.Canvas (so an // immediate call to clearRasters() will be respected, for example), but // Image loads its data asynchronously. Fortunately we can resolve this // by putting TileLayer.Canvas into async mode, which will let us create // and attach the layer but have it wait until the image is loaded before // it actually draws anything. - // These are the variables that we will populate once the image is loaded. + + var imgData = null; // 1d row-major array, four [0-255] integers per pixel + var imgDataMipMapper = null; var w = null; // image width in pixels - var h = null; // image height in pixels + var h = null; // image height in pixels // We'll use this array to store callbacks that need to be invoked once // imgData, w, and h have been resolved. - var imgDataCallbacks = []; - // Consumers of imgData, w, and h can call this to be notified when data + var imgDataCallbacks = []; // Consumers of imgData, w, and h can call this to be notified when data // is available. + function getImageData(callback) { if (imgData != null) { // Must not invoke the callback immediately; it's too confusing and @@ -2216,83 +2354,79 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g } var img = new Image(); + img.onload = function () { // Save size w = img.width; - h = img.height; + h = img.height; // Create a dummy canvas to extract the image data - // Create a dummy canvas to extract the image data var imgDataCanvas = document.createElement("canvas"); imgDataCanvas.width = w; imgDataCanvas.height = h; imgDataCanvas.style.display = "none"; document.body.appendChild(imgDataCanvas); - var imgDataCtx = imgDataCanvas.getContext("2d"); - imgDataCtx.drawImage(img, 0, 0); + imgDataCtx.drawImage(img, 0, 0); // Save the image data. - // Save the image data. imgData = imgDataCtx.getImageData(0, 0, w, h).data; - imgDataMipMapper = new _mipmapper2.default(img); + imgDataMipMapper = new _mipmapper2["default"](img); // Done with the canvas, remove it from the page so it can be gc'd. - // Done with the canvas, remove it from the page so it can be gc'd. - document.body.removeChild(imgDataCanvas); + document.body.removeChild(imgDataCanvas); // Alert any getImageData callers who are waiting. - // Alert any getImageData callers who are waiting. for (var i = 0; i < imgDataCallbacks.length; i++) { imgDataCallbacks[i](imgData, w, h, imgDataMipMapper); } + imgDataCallbacks = []; }; + img.src = uri; - var canvasTiles = _leaflet2.default.gridLayer({ + var canvasTiles = _leaflet2["default"].gridLayer({ opacity: opacity, attribution: attribution, detectRetina: true, async: true - }); - - // NOTE: The done() function MUST NOT be invoked until after the current + }); // NOTE: The done() function MUST NOT be invoked until after the current // tick; done() looks in Leaflet's tile cache for the current tile, and // since it's still being constructed, it won't be found. + + canvasTiles.createTile = function (tilePoint, done) { var zoom = tilePoint.z; - var canvas = _leaflet2.default.DomUtil.create("canvas"); - var error = void 0; - // setup tile width and height according to the options + var canvas = _leaflet2["default"].DomUtil.create("canvas"); + + var error; // setup tile width and height according to the options + var size = this.getTileSize(); canvas.width = size.x; canvas.height = size.y; - getImageData(function (imgData, w, h, mipmapper) { try { // The Context2D we'll being drawing onto. It's always 256x256. - var ctx = canvas.getContext("2d"); - - // Convert our image data's top-left and bottom-right locations into + var ctx = canvas.getContext("2d"); // Convert our image data's top-left and bottom-right locations into // x/y tile coordinates. This is essentially doing a spherical mercator // projection, then multiplying by 2^zoom. + var topLeft = degree2tile(bounds[0][0], bounds[0][1], zoom); - var bottomRight = degree2tile(bounds[1][0], bounds[1][1], zoom); - // The size of the image in x/y tile coordinates. - var extent = { x: bottomRight.x - topLeft.x, y: bottomRight.y - topLeft.y }; + var bottomRight = degree2tile(bounds[1][0], bounds[1][1], zoom); // The size of the image in x/y tile coordinates. - // Short circuit if tile is totally disjoint from image. - if (!overlap(tilePoint.x, tilePoint.x + 1, topLeft.x, bottomRight.x)) return; - if (!overlap(tilePoint.y, tilePoint.y + 1, topLeft.y, bottomRight.y)) return; + var extent = { + x: bottomRight.x - topLeft.x, + y: bottomRight.y - topLeft.y + }; // Short circuit if tile is totally disjoint from image. - // The linear resolution of the tile we're drawing is always 256px per tile unit. + if (!overlap(tilePoint.x, tilePoint.x + 1, topLeft.x, bottomRight.x)) return; + if (!overlap(tilePoint.y, tilePoint.y + 1, topLeft.y, bottomRight.y)) return; // The linear resolution of the tile we're drawing is always 256px per tile unit. // If the linear resolution (in either direction) of the image is less than 256px // per tile unit, then use nearest neighbor; otherwise, use the canvas's built-in // scaling. + var imgRes = { x: w / extent.x, y: h / extent.y - }; - - // We can do the actual drawing in one of three ways: + }; // We can do the actual drawing in one of three ways: // - Call drawImage(). This is easy and fast, and results in smooth // interpolation (bilinear?). This is what we want when we are // reducing the image from its native size. @@ -2304,36 +2438,32 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g // back to when enlarging, and imageSmoothingEnabled isn't supported. // In theory it's slower, but still pretty fast on my machine, and the // results look the same AFAICT. - // Is imageSmoothingEnabled supported? If so, we can let canvas do // nearest-neighbor interpolation for us. + var smoothingProperty = getCanvasSmoothingProperty(ctx); if (smoothingProperty || imgRes.x >= 256 && imgRes.y >= 256) { // Use built-in scaling - // Turn off anti-aliasing if necessary if (smoothingProperty) { ctx[smoothingProperty] = imgRes.x >= 256 && imgRes.y >= 256; - } - - // Don't necessarily draw with the full-size image; if we're + } // Don't necessarily draw with the full-size image; if we're // downscaling, use the mipmapper to get a pre-downscaled image // (see comments on Mipmapper class for why this matters). + + mipmapper.getBySize(extent.x * 256, extent.y * 256, function (mip) { // It's possible that the image will go off the edge of the canvas-- // that's OK, the canvas should clip appropriately. - ctx.drawImage(mip, - // Convert abs tile coords to rel tile coords, then *256 to convert + ctx.drawImage(mip, // Convert abs tile coords to rel tile coords, then *256 to convert // to rel pixel coords - (topLeft.x - tilePoint.x) * 256, (topLeft.y - tilePoint.y) * 256, - // Always draw the whole thing and let canvas clip; so we can just + (topLeft.x - tilePoint.x) * 256, (topLeft.y - tilePoint.y) * 256, // Always draw the whole thing and let canvas clip; so we can just // convert from size in tile coords straight to pixels extent.x * 256, extent.y * 256); }); } else { // Use manual nearest-neighbor interpolation - // Calculate the source image pixel coordinates that correspond with // the top-left and bottom-right of this tile. (If the source image // only partially overlaps the tile, we use max/min to limit the @@ -2345,17 +2475,15 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g var sourceEnd = { x: Math.min(w, Math.ceil((tilePoint.x + 1 - topLeft.x) * imgRes.x)), y: Math.min(h, Math.ceil((tilePoint.y + 1 - topLeft.y) * imgRes.y)) - }; - - // The size, in dest pixels, that each source pixel should occupy. + }; // The size, in dest pixels, that each source pixel should occupy. // This might be greater or less than 1 (e.g. if x and y resolution // are very different). + var pixelSize = { x: 256 / imgRes.x, y: 256 / imgRes.y - }; + }; // For each pixel in the source image that overlaps the tile... - // For each pixel in the source image that overlaps the tile... for (var row = sourceStart.y; row < sourceEnd.y; row++) { for (var col = sourceStart.x; col < sourceEnd.x; col++) { // ...extract the pixel data... @@ -2364,18 +2492,15 @@ methods.addRasterImage = function (uri, bounds, opacity, attribution, layerId, g var g = imgData[i + 1]; var b = imgData[i + 2]; var a = imgData[i + 3]; - ctx.fillStyle = "rgba(" + [r, g, b, a / 255].join(",") + ")"; - - // ...calculate the corresponding pixel coord in the dest image + ctx.fillStyle = "rgba(" + [r, g, b, a / 255].join(",") + ")"; // ...calculate the corresponding pixel coord in the dest image // where it should be drawn... + var pixelPos = { x: (col / imgRes.x + topLeft.x - tilePoint.x) * 256, y: (row / imgRes.y + topLeft.y - tilePoint.y) * 256 - }; + }; // ...and draw a rectangle there. - // ...and draw a rectangle there. - ctx.fillRect(Math.round(pixelPos.x), Math.round(pixelPos.y), - // Looks crazy, but this is necessary to prevent rounding from + ctx.fillRect(Math.round(pixelPos.x), Math.round(pixelPos.y), // Looks crazy, but this is necessary to prevent rounding from // causing overlap between this rect and its neighbors. The // minuend is the location of the next pixel, while the // subtrahend is the position of the current pixel (to turn an @@ -2409,7 +2534,7 @@ methods.addMeasure = function (options) { // if a measureControl already exists, then remove it and // replace with a new one methods.removeMeasure.call(this); - this.measureControl = _leaflet2.default.control.measure(options); + this.measureControl = _leaflet2["default"].control.measure(options); this.addControl(this.measureControl); }; @@ -2424,15 +2549,14 @@ methods.addSelect = function (ctGroup) { var _this8 = this; methods.removeSelect.call(this); - - this._selectButton = _leaflet2.default.easyButton({ + this._selectButton = _leaflet2["default"].easyButton({ states: [{ stateName: "select-inactive", icon: "ion-qr-scanner", title: "Make a selection", onClick: function onClick(btn, map) { btn.state("select-active"); - _this8._locationFilter = new _leaflet2.default.LocationFilter2(); + _this8._locationFilter = new _leaflet2["default"].LocationFilter2(); if (ctGroup) { var selectionHandle = new global.crosstalk.SelectionHandle(ctGroup); @@ -2440,15 +2564,22 @@ methods.addSelect = function (ctGroup) { if (e.sender !== selectionHandle) { if (_this8._locationFilter) { _this8._locationFilter.disable(); + btn.state("select-inactive"); } } }); + var handler = function handler(e) { - _this8.layerManager.brush(_this8._locationFilter.getBounds(), { sender: selectionHandle }); + _this8.layerManager.brush(_this8._locationFilter.getBounds(), { + sender: selectionHandle + }); }; + _this8._locationFilter.on("enabled", handler); + _this8._locationFilter.on("change", handler); + _this8._locationFilter.on("disabled", function () { selectionHandle.close(); _this8._locationFilter = null; @@ -2463,8 +2594,10 @@ methods.addSelect = function (ctGroup) { title: "Dismiss selection", onClick: function onClick(btn, map) { btn.state("select-inactive"); - _this8._locationFilter.disable(); - // If explicitly dismissed, clear the crosstalk selections + + _this8._locationFilter.disable(); // If explicitly dismissed, clear the crosstalk selections + + _this8.layerManager.unbrush(); } }] @@ -2490,7 +2623,7 @@ methods.createMapPane = function (name, zIndex) { }; -}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) +}).call(this)}).call(this,typeof global !== "undefined" ? global : typeof self !== "undefined" ? self : typeof window !== "undefined" ? window : {}) },{"./cluster-layer-store":1,"./crs_utils":3,"./dataframe":4,"./global/htmlwidgets":8,"./global/jquery":9,"./global/leaflet":10,"./global/shiny":12,"./mipmapper":16,"./util":17}],16:[function(require,module,exports){ "use strict"; @@ -2498,23 +2631,23 @@ Object.defineProperty(exports, "__esModule", { value: true }); -var _createClass = function () { function defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } return function (Constructor, protoProps, staticProps) { if (protoProps) defineProperties(Constructor.prototype, protoProps); if (staticProps) defineProperties(Constructor, staticProps); return Constructor; }; }(); - function _classCallCheck(instance, Constructor) { if (!(instance instanceof Constructor)) { throw new TypeError("Cannot call a class as a function"); } } +function _defineProperties(target, props) { for (var i = 0; i < props.length; i++) { var descriptor = props[i]; descriptor.enumerable = descriptor.enumerable || false; descriptor.configurable = true; if ("value" in descriptor) descriptor.writable = true; Object.defineProperty(target, descriptor.key, descriptor); } } + +function _createClass(Constructor, protoProps, staticProps) { if (protoProps) _defineProperties(Constructor.prototype, protoProps); if (staticProps) _defineProperties(Constructor, staticProps); return Constructor; } + // This class simulates a mipmap, which shrinks images by powers of two. This // stepwise reduction results in "pixel-perfect downscaling" (where every // pixel of the original image has some contribution to the downscaled image) // as opposed to a single-step downscaling which will discard a lot of data // (and with sparse images at small scales can give very surprising results). -var Mipmapper = function () { +var Mipmapper = /*#__PURE__*/function () { function Mipmapper(img) { _classCallCheck(this, Mipmapper); this._layers = [img]; - } - - // The various functions on this class take a callback function BUT MAY OR MAY + } // The various functions on this class take a callback function BUT MAY OR MAY // NOT actually behave asynchronously. @@ -2525,6 +2658,7 @@ var Mipmapper = function () { var i = 0; var lastImg = this._layers[0]; + var testNext = function testNext() { _this.getByIndex(i, function (img) { // If current image is invalid (i.e. too small to be rendered) or @@ -2540,6 +2674,7 @@ var Mipmapper = function () { } }); }; + testNext(); } }, { @@ -2558,14 +2693,16 @@ var Mipmapper = function () { callback(null); return; } + if (prevImg.width < 2 || prevImg.height < 2) { // Can't reduce this image any further callback(null); return; - } - // If reduce ever becomes truly asynchronous, we should stuff a promise or + } // If reduce ever becomes truly asynchronous, we should stuff a promise or // something into this._layers[i] before calling this.reduce(), to prevent // redundant reduce operations from happening. + + _this2.reduce(prevImg, function (reducedImg) { _this2._layers[i] = reducedImg; callback(reducedImg); @@ -2581,6 +2718,7 @@ var Mipmapper = function () { imgDataCanvas.height = Math.ceil(img.height / 2); imgDataCanvas.style.display = "none"; document.body.appendChild(imgDataCanvas); + try { var imgDataCtx = imgDataCanvas.getContext("2d"); imgDataCtx.drawImage(img, 0, 0, img.width / 2, img.height / 2); @@ -2594,7 +2732,7 @@ var Mipmapper = function () { return Mipmapper; }(); -exports.default = Mipmapper; +exports["default"] = Mipmapper; },{}],17:[function(require,module,exports){ @@ -2606,6 +2744,7 @@ Object.defineProperty(exports, "__esModule", { exports.log = log; exports.recycle = recycle; exports.asArray = asArray; + function log(message) { /* eslint-disable no-console */ if (console && console.log) console.log(message); @@ -2619,18 +2758,22 @@ function recycle(values, length, inPlace) { if (inPlace) { throw new Error("Can't do in-place recycling of a non-Array value"); } + values = [values]; } - if (typeof length === "undefined") length = values.length; + if (typeof length === "undefined") length = values.length; var dest = inPlace ? values : []; var origLength = values.length; + while (dest.length < length) { dest.push(values[dest.length % origLength]); } + if (dest.length > length) { dest.splice(length, dest.length - length); } + return dest; } diff --git a/rmarkdown-libs/leaflet/leaflet.js b/rmarkdown-libs/leaflet/leaflet.js index 764fbb8..4eca8af 100644 --- a/rmarkdown-libs/leaflet/leaflet.js +++ b/rmarkdown-libs/leaflet/leaflet.js @@ -1,5 +1,5 @@ /* @preserve - * Leaflet 1.3.1+Detached: ba6f97fff8647e724e4dfe66d2ed7da11f908989.ba6f97f, a JS library for interactive maps. http://leafletjs.com + * Leaflet 1.3.1+Detached: ba6f97fff8647e724e4dfe66d2ed7da11f908989.ba6f97f, a JS library for interactive maps. https://leafletjs.com * (c) 2010-2017 Vladimir Agafonkin, (c) 2010-2011 CloudMade */ -!function(t,i){"object"==typeof exports&&"undefined"!=typeof module?i(exports):"function"==typeof define&&define.amd?define(["exports"],i):i(t.L={})}(this,function(t){"use strict";function i(t){var i,e,n,o;for(e=1,n=arguments.length;e=0}function I(t,i,e,n){return"touchstart"===i?O(t,e,n):"touchmove"===i?W(t,e,n):"touchend"===i&&H(t,e,n),this}function B(t,i,e){var n=t["_leaflet_"+i+e];return"touchstart"===i?t.removeEventListener(Qi,n,!1):"touchmove"===i?t.removeEventListener(te,n,!1):"touchend"===i&&(t.removeEventListener(ie,n,!1),t.removeEventListener(ee,n,!1)),this}function O(t,i,n){var o=e(function(t){if("mouse"!==t.pointerType&&t.MSPOINTER_TYPE_MOUSE&&t.pointerType!==t.MSPOINTER_TYPE_MOUSE){if(!(ne.indexOf(t.target.tagName)<0))return;$(t)}j(t,i)});t["_leaflet_touchstart"+n]=o,t.addEventListener(Qi,o,!1),se||(document.documentElement.addEventListener(Qi,R,!0),document.documentElement.addEventListener(te,D,!0),document.documentElement.addEventListener(ie,N,!0),document.documentElement.addEventListener(ee,N,!0),se=!0)}function R(t){oe[t.pointerId]=t,re++}function D(t){oe[t.pointerId]&&(oe[t.pointerId]=t)}function N(t){delete oe[t.pointerId],re--}function j(t,i){t.touches=[];for(var e in oe)t.touches.push(oe[e]);t.changedTouches=[t],i(t)}function W(t,i,e){var n=function(t){(t.pointerType!==t.MSPOINTER_TYPE_MOUSE&&"mouse"!==t.pointerType||0!==t.buttons)&&j(t,i)};t["_leaflet_touchmove"+e]=n,t.addEventListener(te,n,!1)}function H(t,i,e){var n=function(t){j(t,i)};t["_leaflet_touchend"+e]=n,t.addEventListener(ie,n,!1),t.addEventListener(ee,n,!1)}function F(t,i,e){function n(t){var i;if(Ui){if(!Pi||"mouse"===t.pointerType)return;i=re}else i=t.touches.length;if(!(i>1)){var e=Date.now(),n=e-(s||e);r=t.touches?t.touches[0]:t,a=n>0&&n<=h,s=e}}function o(t){if(a&&!r.cancelBubble){if(Ui){if(!Pi||"mouse"===t.pointerType)return;var e,n,o={};for(n in r)e=r[n],o[n]=e&&e.bind?e.bind(r):e;r=o}r.type="dblclick",i(r),s=null}}var s,r,a=!1,h=250;return t[ue+ae+e]=n,t[ue+he+e]=o,t[ue+"dblclick"+e]=i,t.addEventListener(ae,n,!1),t.addEventListener(he,o,!1),t.addEventListener("dblclick",i,!1),this}function U(t,i){var e=t[ue+ae+i],n=t[ue+he+i],o=t[ue+"dblclick"+i];return t.removeEventListener(ae,e,!1),t.removeEventListener(he,n,!1),Pi||t.removeEventListener("dblclick",o,!1),this}function V(t,i,e,n){if("object"==typeof i)for(var o in i)G(t,o,i[o],e);else for(var s=0,r=(i=u(i)).length;s100&&n<500||t.target._simulatedClick&&!t._simulated?Q(t):(pi=e,i(t))}function rt(t){return"string"==typeof t?document.getElementById(t):t}function at(t,i){var e=t.style[i]||t.currentStyle&&t.currentStyle[i];if((!e||"auto"===e)&&document.defaultView){var n=document.defaultView.getComputedStyle(t,null);e=n?n[i]:null}return"auto"===e?null:e}function ht(t,i,e){var n=document.createElement(t);return n.className=i||"",e&&e.appendChild(n),n}function ut(t){var i=t.parentNode;i&&i.removeChild(t)}function lt(t){for(;t.firstChild;)t.removeChild(t.firstChild)}function ct(t){var i=t.parentNode;i.lastChild!==t&&i.appendChild(t)}function _t(t){var i=t.parentNode;i.firstChild!==t&&i.insertBefore(t,i.firstChild)}function dt(t,i){if(void 0!==t.classList)return t.classList.contains(i);var e=gt(t);return e.length>0&&new RegExp("(^|\\s)"+i+"(\\s|$)").test(e)}function pt(t,i){if(void 0!==t.classList)for(var e=u(i),n=0,o=e.length;nh&&(s=r,h=a);h>e&&(i[s]=1,Et(t,i,e,n,s),Et(t,i,e,s,o))}function kt(t,i){for(var e=[t[0]],n=1,o=0,s=t.length;ni&&(e.push(t[n]),o=n);return oi.max.x&&(e|=2),t.yi.max.y&&(e|=8),e}function Ot(t,i){var e=i.x-t.x,n=i.y-t.y;return e*e+n*n}function Rt(t,i,e,n){var o,s=i.x,r=i.y,a=e.x-s,h=e.y-r,u=a*a+h*h;return u>0&&((o=((t.x-s)*a+(t.y-r)*h)/u)>1?(s=e.x,r=e.y):o>0&&(s+=a*o,r+=h*o)),a=t.x-s,h=t.y-r,n?a*a+h*h:new x(s,r)}function Dt(t){return!ei(t[0])||"object"!=typeof t[0][0]&&void 0!==t[0][0]}function Nt(t){return console.warn("Deprecated use of _flat, please use L.LineUtil.isFlat instead."),Dt(t)}function jt(t,i,e){var n,o,s,r,a,h,u,l,c,_=[1,4,2,8];for(o=0,u=t.length;o0?Math.floor(t):Math.ceil(t)};x.prototype={clone:function(){return new x(this.x,this.y)},add:function(t){return this.clone()._add(w(t))},_add:function(t){return this.x+=t.x,this.y+=t.y,this},subtract:function(t){return this.clone()._subtract(w(t))},_subtract:function(t){return this.x-=t.x,this.y-=t.y,this},divideBy:function(t){return this.clone()._divideBy(t)},_divideBy:function(t){return this.x/=t,this.y/=t,this},multiplyBy:function(t){return this.clone()._multiplyBy(t)},_multiplyBy:function(t){return this.x*=t,this.y*=t,this},scaleBy:function(t){return new x(this.x*t.x,this.y*t.y)},unscaleBy:function(t){return new x(this.x/t.x,this.y/t.y)},round:function(){return this.clone()._round()},_round:function(){return this.x=Math.round(this.x),this.y=Math.round(this.y),this},floor:function(){return this.clone()._floor()},_floor:function(){return this.x=Math.floor(this.x),this.y=Math.floor(this.y),this},ceil:function(){return this.clone()._ceil()},_ceil:function(){return this.x=Math.ceil(this.x),this.y=Math.ceil(this.y),this},trunc:function(){return this.clone()._trunc()},_trunc:function(){return this.x=li(this.x),this.y=li(this.y),this},distanceTo:function(t){var i=(t=w(t)).x-this.x,e=t.y-this.y;return Math.sqrt(i*i+e*e)},equals:function(t){return(t=w(t)).x===this.x&&t.y===this.y},contains:function(t){return t=w(t),Math.abs(t.x)<=Math.abs(this.x)&&Math.abs(t.y)<=Math.abs(this.y)},toString:function(){return"Point("+a(this.x)+", "+a(this.y)+")"}},P.prototype={extend:function(t){return t=w(t),this.min||this.max?(this.min.x=Math.min(t.x,this.min.x),this.max.x=Math.max(t.x,this.max.x),this.min.y=Math.min(t.y,this.min.y),this.max.y=Math.max(t.y,this.max.y)):(this.min=t.clone(),this.max=t.clone()),this},getCenter:function(t){return new x((this.min.x+this.max.x)/2,(this.min.y+this.max.y)/2,t)},getBottomLeft:function(){return new x(this.min.x,this.max.y)},getTopRight:function(){return new x(this.max.x,this.min.y)},getTopLeft:function(){return this.min},getBottomRight:function(){return this.max},getSize:function(){return this.max.subtract(this.min)},contains:function(t){var i,e;return(t="number"==typeof t[0]||t instanceof x?w(t):b(t))instanceof P?(i=t.min,e=t.max):i=e=t,i.x>=this.min.x&&e.x<=this.max.x&&i.y>=this.min.y&&e.y<=this.max.y},intersects:function(t){t=b(t);var i=this.min,e=this.max,n=t.min,o=t.max,s=o.x>=i.x&&n.x<=e.x,r=o.y>=i.y&&n.y<=e.y;return s&&r},overlaps:function(t){t=b(t);var i=this.min,e=this.max,n=t.min,o=t.max,s=o.x>i.x&&n.xi.y&&n.y=n.lat&&e.lat<=o.lat&&i.lng>=n.lng&&e.lng<=o.lng},intersects:function(t){t=z(t);var i=this._southWest,e=this._northEast,n=t.getSouthWest(),o=t.getNorthEast(),s=o.lat>=i.lat&&n.lat<=e.lat,r=o.lng>=i.lng&&n.lng<=e.lng;return s&&r},overlaps:function(t){t=z(t);var i=this._southWest,e=this._northEast,n=t.getSouthWest(),o=t.getNorthEast(),s=o.lat>i.lat&&n.lati.lng&&n.lng1,Yi=!!document.createElement("canvas").getContext,Xi=!(!document.createElementNS||!E("svg").createSVGRect),Ji=!Xi&&function(){try{var t=document.createElement("div");t.innerHTML='';var i=t.firstChild;return i.style.behavior="url(#default#VML)",i&&"object"==typeof i.adj}catch(t){return!1}}(),$i=(Object.freeze||Object)({ie:wi,ielt9:Li,edge:Pi,webkit:bi,android:Ti,android23:zi,androidStock:Ci,opera:Zi,chrome:Si,gecko:Ei,safari:ki,phantom:Ai,opera12:Ii,win:Bi,ie3d:Oi,webkit3d:Ri,gecko3d:Di,any3d:Ni,mobile:ji,mobileWebkit:Wi,mobileWebkit3d:Hi,msPointer:Fi,pointer:Ui,touch:Vi,mobileOpera:qi,mobileGecko:Gi,retina:Ki,canvas:Yi,svg:Xi,vml:Ji}),Qi=Fi?"MSPointerDown":"pointerdown",te=Fi?"MSPointerMove":"pointermove",ie=Fi?"MSPointerUp":"pointerup",ee=Fi?"MSPointerCancel":"pointercancel",ne=["INPUT","SELECT","OPTION"],oe={},se=!1,re=0,ae=Fi?"MSPointerDown":Ui?"pointerdown":"touchstart",he=Fi?"MSPointerUp":Ui?"pointerup":"touchend",ue="_leaflet_",le="_leaflet_events",ce=Bi&&Si?2*window.devicePixelRatio:Ei?window.devicePixelRatio:1,_e={},de=(Object.freeze||Object)({on:V,off:q,stopPropagation:Y,disableScrollPropagation:X,disableClickPropagation:J,preventDefault:$,stop:Q,getMousePosition:tt,getWheelDelta:it,fakeStop:et,skipped:nt,isExternalTarget:ot,addListener:V,removeListener:q}),pe=xt(["transform","WebkitTransform","OTransform","MozTransform","msTransform"]),me=xt(["webkitTransition","transition","OTransition","MozTransition","msTransition"]),fe="webkitTransition"===me||"OTransition"===me?me+"End":"transitionend";if("onselectstart"in document)mi=function(){V(window,"selectstart",$)},fi=function(){q(window,"selectstart",$)};else{var ge=xt(["userSelect","WebkitUserSelect","OUserSelect","MozUserSelect","msUserSelect"]);mi=function(){if(ge){var t=document.documentElement.style;gi=t[ge],t[ge]="none"}},fi=function(){ge&&(document.documentElement.style[ge]=gi,gi=void 0)}}var ve,ye,xe=(Object.freeze||Object)({TRANSFORM:pe,TRANSITION:me,TRANSITION_END:fe,get:rt,getStyle:at,create:ht,remove:ut,empty:lt,toFront:ct,toBack:_t,hasClass:dt,addClass:pt,removeClass:mt,setClass:ft,getClass:gt,setOpacity:vt,testProp:xt,setTransform:wt,setPosition:Lt,getPosition:Pt,disableTextSelection:mi,enableTextSelection:fi,disableImageDrag:bt,enableImageDrag:Tt,preventOutline:zt,restoreOutline:Mt}),we=ui.extend({run:function(t,i,e,n){this.stop(),this._el=t,this._inProgress=!0,this._duration=e||.25,this._easeOutPower=1/Math.max(n||.5,.2),this._startPos=Pt(t),this._offset=i.subtract(this._startPos),this._startTime=+new Date,this.fire("start"),this._animate()},stop:function(){this._inProgress&&(this._step(!0),this._complete())},_animate:function(){this._animId=f(this._animate,this),this._step()},_step:function(t){var i=+new Date-this._startTime,e=1e3*this._duration;ithis.options.maxZoom)?this.setZoom(t):this},panInsideBounds:function(t,i){this._enforcingBounds=!0;var e=this.getCenter(),n=this._limitCenter(e,this._zoom,z(t));return e.equals(n)||this.panTo(n,i),this._enforcingBounds=!1,this},invalidateSize:function(t){if(!this._loaded)return this;t=i({animate:!1,pan:!0},!0===t?{animate:!0}:t);var n=this.getSize();this._sizeChanged=!0,this._lastCenter=null;var o=this.getSize(),s=n.divideBy(2).round(),r=o.divideBy(2).round(),a=s.subtract(r);return a.x||a.y?(t.animate&&t.pan?this.panBy(a):(t.pan&&this._rawPanBy(a),this.fire("move"),t.debounceMoveend?(clearTimeout(this._sizeTimer),this._sizeTimer=setTimeout(e(this.fire,this,"moveend"),200)):this.fire("moveend")),this.fire("resize",{oldSize:n,newSize:o})):this},stop:function(){return this.setZoom(this._limitZoom(this._zoom)),this.options.zoomSnap||this.fire("viewreset"),this._stop()},locate:function(t){if(t=this._locateOptions=i({timeout:1e4,watch:!1},t),!("geolocation"in navigator))return this._handleGeolocationError({code:0,message:"Geolocation not supported."}),this;var n=e(this._handleGeolocationResponse,this),o=e(this._handleGeolocationError,this);return t.watch?this._locationWatchId=navigator.geolocation.watchPosition(n,o,t):navigator.geolocation.getCurrentPosition(n,o,t),this},stopLocate:function(){return navigator.geolocation&&navigator.geolocation.clearWatch&&navigator.geolocation.clearWatch(this._locationWatchId),this._locateOptions&&(this._locateOptions.setView=!1),this},_handleGeolocationError:function(t){var i=t.code,e=t.message||(1===i?"permission denied":2===i?"position unavailable":"timeout");this._locateOptions.setView&&!this._loaded&&this.fitWorld(),this.fire("locationerror",{code:i,message:"Geolocation error: "+e+"."})},_handleGeolocationResponse:function(t){var i=new M(t.coords.latitude,t.coords.longitude),e=i.toBounds(t.coords.accuracy),n=this._locateOptions;if(n.setView){var o=this.getBoundsZoom(e);this.setView(i,n.maxZoom?Math.min(o,n.maxZoom):o)}var s={latlng:i,bounds:e,timestamp:t.timestamp};for(var r in t.coords)"number"==typeof t.coords[r]&&(s[r]=t.coords[r]);this.fire("locationfound",s)},addHandler:function(t,i){if(!i)return this;var e=this[t]=new i(this);return this._handlers.push(e),this.options[t]&&e.enable(),this},remove:function(){if(this._initEvents(!0),this._containerId!==this._container._leaflet_id)throw new Error("Map container is being reused by another instance");try{delete this._container._leaflet_id,delete this._containerId}catch(t){this._container._leaflet_id=void 0,this._containerId=void 0}void 0!==this._locationWatchId&&this.stopLocate(),this._stop(),ut(this._mapPane),this._clearControlPos&&this._clearControlPos(),this._clearHandlers(),this._loaded&&this.fire("unload");var t;for(t in this._layers)this._layers[t].remove();for(t in this._panes)ut(this._panes[t]);return this._layers=[],this._panes=[],delete this._mapPane,delete this._renderer,this},createPane:function(t,i){var e=ht("div","leaflet-pane"+(t?" leaflet-"+t.replace("Pane","")+"-pane":""),i||this._mapPane);return t&&(this._panes[t]=e),e},getCenter:function(){return this._checkIfLoaded(),this._lastCenter&&!this._moved()?this._lastCenter:this.layerPointToLatLng(this._getCenterLayerPoint())},getZoom:function(){return this._zoom},getBounds:function(){var t=this.getPixelBounds();return new T(this.unproject(t.getBottomLeft()),this.unproject(t.getTopRight()))},getMinZoom:function(){return void 0===this.options.minZoom?this._layersMinZoom||0:this.options.minZoom},getMaxZoom:function(){return void 0===this.options.maxZoom?void 0===this._layersMaxZoom?1/0:this._layersMaxZoom:this.options.maxZoom},getBoundsZoom:function(t,i,e){t=z(t),e=w(e||[0,0]);var n=this.getZoom()||0,o=this.getMinZoom(),s=this.getMaxZoom(),r=t.getNorthWest(),a=t.getSouthEast(),h=this.getSize().subtract(e),u=b(this.project(a,n),this.project(r,n)).getSize(),l=Ni?this.options.zoomSnap:1,c=h.x/u.x,_=h.y/u.y,d=i?Math.max(c,_):Math.min(c,_);return n=this.getScaleZoom(d,n),l&&(n=Math.round(n/(l/100))*(l/100),n=i?Math.ceil(n/l)*l:Math.floor(n/l)*l),Math.max(o,Math.min(s,n))},getSize:function(){return this._size&&!this._sizeChanged||(this._size=new x(this._container.clientWidth||0,this._container.clientHeight||0),this._sizeChanged=!1),this._size.clone()},getPixelBounds:function(t,i){var e=this._getTopLeftPoint(t,i);return new P(e,e.add(this.getSize()))},getPixelOrigin:function(){return this._checkIfLoaded(),this._pixelOrigin},getPixelWorldBounds:function(t){return this.options.crs.getProjectedBounds(void 0===t?this.getZoom():t)},getPane:function(t){return"string"==typeof t?this._panes[t]:t},getPanes:function(){return this._panes},getContainer:function(){return this._container},getZoomScale:function(t,i){var e=this.options.crs;return i=void 0===i?this._zoom:i,e.scale(t)/e.scale(i)},getScaleZoom:function(t,i){var e=this.options.crs;i=void 0===i?this._zoom:i;var n=e.zoom(t*e.scale(i));return isNaN(n)?1/0:n},project:function(t,i){return i=void 0===i?this._zoom:i,this.options.crs.latLngToPoint(C(t),i)},unproject:function(t,i){return i=void 0===i?this._zoom:i,this.options.crs.pointToLatLng(w(t),i)},layerPointToLatLng:function(t){var i=w(t).add(this.getPixelOrigin());return this.unproject(i)},latLngToLayerPoint:function(t){return this.project(C(t))._round()._subtract(this.getPixelOrigin())},wrapLatLng:function(t){return this.options.crs.wrapLatLng(C(t))},wrapLatLngBounds:function(t){return this.options.crs.wrapLatLngBounds(z(t))},distance:function(t,i){return this.options.crs.distance(C(t),C(i))},containerPointToLayerPoint:function(t){return w(t).subtract(this._getMapPanePos())},layerPointToContainerPoint:function(t){return w(t).add(this._getMapPanePos())},containerPointToLatLng:function(t){var i=this.containerPointToLayerPoint(w(t));return this.layerPointToLatLng(i)},latLngToContainerPoint:function(t){return this.layerPointToContainerPoint(this.latLngToLayerPoint(C(t)))},mouseEventToContainerPoint:function(t){return tt(t,this._container)},mouseEventToLayerPoint:function(t){return this.containerPointToLayerPoint(this.mouseEventToContainerPoint(t))},mouseEventToLatLng:function(t){return this.layerPointToLatLng(this.mouseEventToLayerPoint(t))},_initContainer:function(t){var i=this._container=rt(t);if(!i)throw new Error("Map container not found.");if(i._leaflet_id)throw new Error("Map container is already initialized.");V(i,"scroll",this._onScroll,this),this._containerId=n(i)},_initLayout:function(){var t=this._container;this._fadeAnimated=this.options.fadeAnimation&&Ni,pt(t,"leaflet-container"+(Vi?" leaflet-touch":"")+(Ki?" leaflet-retina":"")+(Li?" leaflet-oldie":"")+(ki?" leaflet-safari":"")+(this._fadeAnimated?" leaflet-fade-anim":""));var i=at(t,"position");"absolute"!==i&&"relative"!==i&&"fixed"!==i&&(t.style.position="relative"),this._initPanes(),this._initControlPos&&this._initControlPos()},_initPanes:function(){var t=this._panes={};this._paneRenderers={},this._mapPane=this.createPane("mapPane",this._container),Lt(this._mapPane,new x(0,0)),this.createPane("tilePane"),this.createPane("shadowPane"),this.createPane("overlayPane"),this.createPane("markerPane"),this.createPane("tooltipPane"),this.createPane("popupPane"),this.options.markerZoomAnimation||(pt(t.markerPane,"leaflet-zoom-hide"),pt(t.shadowPane,"leaflet-zoom-hide"))},_resetView:function(t,i){Lt(this._mapPane,new x(0,0));var e=!this._loaded;this._loaded=!0,i=this._limitZoom(i),this.fire("viewprereset");var n=this._zoom!==i;this._moveStart(n,!1)._move(t,i)._moveEnd(n),this.fire("viewreset"),e&&this.fire("load")},_moveStart:function(t,i){return t&&this.fire("zoomstart"),i||this.fire("movestart"),this},_move:function(t,i,e){void 0===i&&(i=this._zoom);var n=this._zoom!==i;return this._zoom=i,this._lastCenter=t,this._pixelOrigin=this._getNewPixelOrigin(t),(n||e&&e.pinch)&&this.fire("zoom",e),this.fire("move",e)},_moveEnd:function(t){return t&&this.fire("zoomend"),this.fire("moveend")},_stop:function(){return g(this._flyToFrame),this._panAnim&&this._panAnim.stop(),this},_rawPanBy:function(t){Lt(this._mapPane,this._getMapPanePos().subtract(t))},_getZoomSpan:function(){return this.getMaxZoom()-this.getMinZoom()},_panInsideMaxBounds:function(){this._enforcingBounds||this.panInsideBounds(this.options.maxBounds)},_checkIfLoaded:function(){if(!this._loaded)throw new Error("Set map center and zoom first.")},_initEvents:function(t){this._targets={},this._targets[n(this._container)]=this;var i=t?q:V;i(this._container,"click dblclick mousedown mouseup mouseover mouseout mousemove contextmenu keypress",this._handleDOMEvent,this),this.options.trackResize&&i(window,"resize",this._onResize,this),Ni&&this.options.transform3DLimit&&(t?this.off:this.on).call(this,"moveend",this._onMoveEnd)},_onResize:function(){g(this._resizeRequest),this._resizeRequest=f(function(){this.invalidateSize({debounceMoveend:!0})},this)},_onScroll:function(){this._container.scrollTop=0,this._container.scrollLeft=0},_onMoveEnd:function(){var t=this._getMapPanePos();Math.max(Math.abs(t.x),Math.abs(t.y))>=this.options.transform3DLimit&&this._resetView(this.getCenter(),this.getZoom())},_findEventTargets:function(t,i){for(var e,o=[],s="mouseout"===i||"mouseover"===i,r=t.target||t.srcElement,a=!1;r;){if((e=this._targets[n(r)])&&("click"===i||"preclick"===i)&&!t._simulated&&this._draggableMoved(e)){a=!0;break}if(e&&e.listens(i,!0)){if(s&&!ot(r,t))break;if(o.push(e),s)break}if(r===this._container)break;r=r.parentNode}return o.length||a||s||!ot(r,t)||(o=[this]),o},_handleDOMEvent:function(t){if(this._loaded&&!nt(t)){var i=t.type;"mousedown"!==i&&"keypress"!==i||zt(t.target||t.srcElement),this._fireDOMEvent(t,i)}},_mouseEvents:["click","dblclick","mouseover","mouseout","contextmenu"],_fireDOMEvent:function(t,e,n){if("click"===t.type){var o=i({},t);o.type="preclick",this._fireDOMEvent(o,o.type,n)}if(!t._stopped&&(n=(n||[]).concat(this._findEventTargets(t,e))).length){var s=n[0];"contextmenu"===e&&s.listens(e,!0)&&$(t);var r={originalEvent:t};if("keypress"!==t.type){var a=s.getLatLng&&(!s._radius||s._radius<=10);r.containerPoint=a?this.latLngToContainerPoint(s.getLatLng()):this.mouseEventToContainerPoint(t),r.layerPoint=this.containerPointToLayerPoint(r.containerPoint),r.latlng=a?s.getLatLng():this.layerPointToLatLng(r.layerPoint)}for(var h=0;h0?Math.round(t-i)/2:Math.max(0,Math.ceil(t))-Math.max(0,Math.floor(i))},_limitZoom:function(t){var i=this.getMinZoom(),e=this.getMaxZoom(),n=Ni?this.options.zoomSnap:1;return n&&(t=Math.round(t/n)*n),Math.max(i,Math.min(e,t))},_onPanTransitionStep:function(){this.fire("move")},_onPanTransitionEnd:function(){mt(this._mapPane,"leaflet-pan-anim"),this.fire("moveend")},_tryAnimatedPan:function(t,i){var e=this._getCenterOffset(t)._trunc();return!(!0!==(i&&i.animate)&&!this.getSize().contains(e))&&(this.panBy(e,i),!0)},_createAnimProxy:function(){var t=this._proxy=ht("div","leaflet-proxy leaflet-zoom-animated");this._panes.mapPane.appendChild(t),this.on("zoomanim",function(t){var i=pe,e=this._proxy.style[i];wt(this._proxy,this.project(t.center,t.zoom),this.getZoomScale(t.zoom,1)),e===this._proxy.style[i]&&this._animatingZoom&&this._onZoomTransitionEnd()},this),this.on("load moveend",function(){var t=this.getCenter(),i=this.getZoom();wt(this._proxy,this.project(t,i),this.getZoomScale(i,1))},this),this._on("unload",this._destroyAnimProxy,this)},_destroyAnimProxy:function(){ut(this._proxy),delete this._proxy},_catchTransitionEnd:function(t){this._animatingZoom&&t.propertyName.indexOf("transform")>=0&&this._onZoomTransitionEnd()},_nothingToAnimate:function(){return!this._container.getElementsByClassName("leaflet-zoom-animated").length},_tryAnimatedZoom:function(t,i,e){if(this._animatingZoom)return!0;if(e=e||{},!this._zoomAnimated||!1===e.animate||this._nothingToAnimate()||Math.abs(i-this._zoom)>this.options.zoomAnimationThreshold)return!1;var n=this.getZoomScale(i),o=this._getCenterOffset(t)._divideBy(1-1/n);return!(!0!==e.animate&&!this.getSize().contains(o))&&(f(function(){this._moveStart(!0,!1)._animateZoom(t,i,!0)},this),!0)},_animateZoom:function(t,i,n,o){this._mapPane&&(n&&(this._animatingZoom=!0,this._animateToCenter=t,this._animateToZoom=i,pt(this._mapPane,"leaflet-zoom-anim")),this.fire("zoomanim",{center:t,zoom:i,noUpdate:o}),setTimeout(e(this._onZoomTransitionEnd,this),250))},_onZoomTransitionEnd:function(){this._animatingZoom&&(this._mapPane&&mt(this._mapPane,"leaflet-zoom-anim"),this._animatingZoom=!1,this._move(this._animateToCenter,this._animateToZoom),f(function(){this._moveEnd(!0)},this))}}),Pe=v.extend({options:{position:"topright"},initialize:function(t){l(this,t)},getPosition:function(){return this.options.position},setPosition:function(t){var i=this._map;return i&&i.removeControl(this),this.options.position=t,i&&i.addControl(this),this},getContainer:function(){return this._container},addTo:function(t){this.remove(),this._map=t;var i=this._container=this.onAdd(t),e=this.getPosition(),n=t._controlCorners[e];return pt(i,"leaflet-control"),-1!==e.indexOf("bottom")?n.insertBefore(i,n.firstChild):n.appendChild(i),this},remove:function(){return this._map?(ut(this._container),this.onRemove&&this.onRemove(this._map),this._map=null,this):this},_refocusOnMap:function(t){this._map&&t&&t.screenX>0&&t.screenY>0&&this._map.getContainer().focus()}}),be=function(t){return new Pe(t)};Le.include({addControl:function(t){return t.addTo(this),this},removeControl:function(t){return t.remove(),this},_initControlPos:function(){function t(t,o){var s=e+t+" "+e+o;i[t+o]=ht("div",s,n)}var i=this._controlCorners={},e="leaflet-",n=this._controlContainer=ht("div",e+"control-container",this._container);t("top","left"),t("top","right"),t("bottom","left"),t("bottom","right")},_clearControlPos:function(){for(var t in this._controlCorners)ut(this._controlCorners[t]);ut(this._controlContainer),delete this._controlCorners,delete this._controlContainer}});var Te=Pe.extend({options:{collapsed:!0,position:"topright",autoZIndex:!0,hideSingleBase:!1,sortLayers:!1,sortFunction:function(t,i,e,n){return e1,this._baseLayersList.style.display=t?"":"none"),this._separator.style.display=i&&t?"":"none",this},_onLayerChange:function(t){this._handlingClick||this._update();var i=this._getLayer(n(t.target)),e=i.overlay?"add"===t.type?"overlayadd":"overlayremove":"add"===t.type?"baselayerchange":null;e&&this._map.fire(e,i)},_createRadioElement:function(t,i){var e='",n=document.createElement("div");return n.innerHTML=e,n.firstChild},_addItem:function(t){var i,e=document.createElement("label"),o=this._map.hasLayer(t.layer);t.overlay?((i=document.createElement("input")).type="checkbox",i.className="leaflet-control-layers-selector",i.defaultChecked=o):i=this._createRadioElement("leaflet-base-layers",o),this._layerControlInputs.push(i),i.layerId=n(t.layer),V(i,"click",this._onInputClick,this);var s=document.createElement("span");s.innerHTML=" "+t.name;var r=document.createElement("div");return e.appendChild(r),r.appendChild(i),r.appendChild(s),(t.overlay?this._overlaysList:this._baseLayersList).appendChild(e),this._checkDisabledLayers(),e},_onInputClick:function(){var t,i,e=this._layerControlInputs,n=[],o=[];this._handlingClick=!0;for(var s=e.length-1;s>=0;s--)t=e[s],i=this._getLayer(t.layerId).layer,t.checked?n.push(i):t.checked||o.push(i);for(s=0;s=0;o--)t=e[o],i=this._getLayer(t.layerId).layer,t.disabled=void 0!==i.options.minZoom&&ni.options.maxZoom},_expandIfNotCollapsed:function(){return this._map&&!this.options.collapsed&&this.expand(),this},_expand:function(){return this.expand()},_collapse:function(){return this.collapse()}}),ze=Pe.extend({options:{position:"topleft",zoomInText:"+",zoomInTitle:"Zoom in",zoomOutText:"−",zoomOutTitle:"Zoom out"},onAdd:function(t){var i="leaflet-control-zoom",e=ht("div",i+" leaflet-bar"),n=this.options;return this._zoomInButton=this._createButton(n.zoomInText,n.zoomInTitle,i+"-in",e,this._zoomIn),this._zoomOutButton=this._createButton(n.zoomOutText,n.zoomOutTitle,i+"-out",e,this._zoomOut),this._updateDisabled(),t.on("zoomend zoomlevelschange",this._updateDisabled,this),e},onRemove:function(t){t.off("zoomend zoomlevelschange",this._updateDisabled,this)},disable:function(){return this._disabled=!0,this._updateDisabled(),this},enable:function(){return this._disabled=!1,this._updateDisabled(),this},_zoomIn:function(t){!this._disabled&&this._map._zoomthis._map.getMinZoom()&&this._map.zoomOut(this._map.options.zoomDelta*(t.shiftKey?3:1))},_createButton:function(t,i,e,n,o){var s=ht("a",e,n);return s.innerHTML=t,s.href="#",s.title=i,s.setAttribute("role","button"),s.setAttribute("aria-label",i),J(s),V(s,"click",Q),V(s,"click",o,this),V(s,"click",this._refocusOnMap,this),s},_updateDisabled:function(){var t=this._map,i="leaflet-disabled";mt(this._zoomInButton,i),mt(this._zoomOutButton,i),(this._disabled||t._zoom===t.getMinZoom())&&pt(this._zoomOutButton,i),(this._disabled||t._zoom===t.getMaxZoom())&&pt(this._zoomInButton,i)}});Le.mergeOptions({zoomControl:!0}),Le.addInitHook(function(){this.options.zoomControl&&(this.zoomControl=new ze,this.addControl(this.zoomControl))});var Me=Pe.extend({options:{position:"bottomleft",maxWidth:100,metric:!0,imperial:!0},onAdd:function(t){var i=ht("div","leaflet-control-scale"),e=this.options;return this._addScales(e,"leaflet-control-scale-line",i),t.on(e.updateWhenIdle?"moveend":"move",this._update,this),t.whenReady(this._update,this),i},onRemove:function(t){t.off(this.options.updateWhenIdle?"moveend":"move",this._update,this)},_addScales:function(t,i,e){t.metric&&(this._mScale=ht("div",i,e)),t.imperial&&(this._iScale=ht("div",i,e))},_update:function(){var t=this._map,i=t.getSize().y/2,e=t.distance(t.containerPointToLatLng([0,i]),t.containerPointToLatLng([this.options.maxWidth,i]));this._updateScales(e)},_updateScales:function(t){this.options.metric&&t&&this._updateMetric(t),this.options.imperial&&t&&this._updateImperial(t)},_updateMetric:function(t){var i=this._getRoundNum(t),e=i<1e3?i+" m":i/1e3+" km";this._updateScale(this._mScale,e,i/t)},_updateImperial:function(t){var i,e,n,o=3.2808399*t;o>5280?(i=o/5280,e=this._getRoundNum(i),this._updateScale(this._iScale,e+" mi",e/i)):(n=this._getRoundNum(o),this._updateScale(this._iScale,n+" ft",n/o))},_updateScale:function(t,i,e){t.style.width=Math.round(this.options.maxWidth*e)+"px",t.innerHTML=i},_getRoundNum:function(t){var i=Math.pow(10,(Math.floor(t)+"").length-1),e=t/i;return e=e>=10?10:e>=5?5:e>=3?3:e>=2?2:1,i*e}}),Ce=Pe.extend({options:{position:"bottomright",prefix:'    Leaflet'},initialize:function(t){l(this,t),this._attributions={}},onAdd:function(t){t.attributionControl=this,this._container=ht("div","leaflet-control-attribution"),J(this._container);for(var i in t._layers)t._layers[i].getAttribution&&this.addAttribution(t._layers[i].getAttribution());return this._update(),this._container},setPrefix:function(t){return this.options.prefix=t,this._update(),this},addAttribution:function(t){return t?(this._attributions[t]||(this._attributions[t]=0),this._attributions[t]++,this._update(),this):this},removeAttribution:function(t){return t?(this._attributions[t]&&(this._attributions[t]--,this._update()),this):this},_update:function(){if(this._map){var t=[];for(var i in this._attributions)this._attributions[i]&&t.push(i);var e=[];this.options.prefix&&e.push(this.options.prefix),t.length&&e.push(t.join(", ")),this._container.innerHTML=e.join(" | ")}}});Le.mergeOptions({attributionControl:!0}),Le.addInitHook(function(){this.options.attributionControl&&(new Ce).addTo(this)});Pe.Layers=Te,Pe.Zoom=ze,Pe.Scale=Me,Pe.Attribution=Ce,be.layers=function(t,i,e){return new Te(t,i,e)},be.zoom=function(t){return new ze(t)},be.scale=function(t){return new Me(t)},be.attribution=function(t){return new Ce(t)};var Ze=v.extend({initialize:function(t){this._map=t},enable:function(){return this._enabled?this:(this._enabled=!0,this.addHooks(),this)},disable:function(){return this._enabled?(this._enabled=!1,this.removeHooks(),this):this},enabled:function(){return!!this._enabled}});Ze.addTo=function(t,i){return t.addHandler(i,this),this};var Se,Ee={Events:hi},ke=Vi?"touchstart mousedown":"mousedown",Ae={mousedown:"mouseup",touchstart:"touchend",pointerdown:"touchend",MSPointerDown:"touchend"},Ie={mousedown:"mousemove",touchstart:"touchmove",pointerdown:"touchmove",MSPointerDown:"touchmove"},Be=ui.extend({options:{clickTolerance:3},initialize:function(t,i,e,n){l(this,n),this._element=t,this._dragStartTarget=i||t,this._preventOutline=e},enable:function(){this._enabled||(V(this._dragStartTarget,ke,this._onDown,this),this._enabled=!0)},disable:function(){this._enabled&&(Be._dragging===this&&this.finishDrag(),q(this._dragStartTarget,ke,this._onDown,this),this._enabled=!1,this._moved=!1)},_onDown:function(t){if(!t._simulated&&this._enabled&&(this._moved=!1,!dt(this._element,"leaflet-zoom-anim")&&!(Be._dragging||t.shiftKey||1!==t.which&&1!==t.button&&!t.touches||(Be._dragging=this,this._preventOutline&&zt(this._element),bt(),mi(),this._moving)))){this.fire("down");var i=t.touches?t.touches[0]:t;this._startPoint=new x(i.clientX,i.clientY),V(document,Ie[t.type],this._onMove,this),V(document,Ae[t.type],this._onUp,this)}},_onMove:function(t){if(!t._simulated&&this._enabled)if(t.touches&&t.touches.length>1)this._moved=!0;else{var i=t.touches&&1===t.touches.length?t.touches[0]:t,e=new x(i.clientX,i.clientY).subtract(this._startPoint);(e.x||e.y)&&(Math.abs(e.x)+Math.abs(e.y)1e-7;h++)i=s*Math.sin(a),i=Math.pow((1-i)/(1+i),s/2),a+=u=Math.PI/2-2*Math.atan(r*i)-a;return new M(a*e,t.x*e/n)}},je=(Object.freeze||Object)({LonLat:De,Mercator:Ne,SphericalMercator:di}),We=i({},_i,{code:"EPSG:3395",projection:Ne,transformation:function(){var t=.5/(Math.PI*Ne.R);return S(t,.5,-t,.5)}()}),He=i({},_i,{code:"EPSG:4326",projection:De,transformation:S(1/180,1,-1/180,.5)}),Fe=i({},ci,{projection:De,transformation:S(1,0,-1,0),scale:function(t){return Math.pow(2,t)},zoom:function(t){return Math.log(t)/Math.LN2},distance:function(t,i){var e=i.lng-t.lng,n=i.lat-t.lat;return Math.sqrt(e*e+n*n)},infinite:!0});ci.Earth=_i,ci.EPSG3395=We,ci.EPSG3857=vi,ci.EPSG900913=yi,ci.EPSG4326=He,ci.Simple=Fe;var Ue=ui.extend({options:{pane:"overlayPane",attribution:null,bubblingMouseEvents:!0},addTo:function(t){return t.addLayer(this),this},remove:function(){return this.removeFrom(this._map||this._mapToAdd)},removeFrom:function(t){return t&&t.removeLayer(this),this},getPane:function(t){return this._map.getPane(t?this.options[t]||t:this.options.pane)},addInteractiveTarget:function(t){return this._map._targets[n(t)]=this,this},removeInteractiveTarget:function(t){return delete this._map._targets[n(t)],this},getAttribution:function(){return this.options.attribution},_layerAdd:function(t){var i=t.target;if(i.hasLayer(this)){if(this._map=i,this._zoomAnimated=i._zoomAnimated,this.getEvents){var e=this.getEvents();i.on(e,this),this.once("remove",function(){i.off(e,this)},this)}this.onAdd(i),this.getAttribution&&i.attributionControl&&i.attributionControl.addAttribution(this.getAttribution()),this.fire("add"),i.fire("layeradd",{layer:this})}}});Le.include({addLayer:function(t){if(!t._layerAdd)throw new Error("The provided object is not a Layer.");var i=n(t);return this._layers[i]?this:(this._layers[i]=t,t._mapToAdd=this,t.beforeAdd&&t.beforeAdd(this),this.whenReady(t._layerAdd,t),this)},removeLayer:function(t){var i=n(t);return this._layers[i]?(this._loaded&&t.onRemove(this),t.getAttribution&&this.attributionControl&&this.attributionControl.removeAttribution(t.getAttribution()),delete this._layers[i],this._loaded&&(this.fire("layerremove",{layer:t}),t.fire("remove")),t._map=t._mapToAdd=null,this):this},hasLayer:function(t){return!!t&&n(t)in this._layers},eachLayer:function(t,i){for(var e in this._layers)t.call(i,this._layers[e]);return this},_addLayers:function(t){for(var i=0,e=(t=t?ei(t)?t:[t]:[]).length;ithis._layersMaxZoom&&this.setZoom(this._layersMaxZoom),void 0===this.options.minZoom&&this._layersMinZoom&&this.getZoom()i)return r=(n-i)/e,this._map.layerPointToLatLng([s.x-r*(s.x-o.x),s.y-r*(s.y-o.y)])},getBounds:function(){return this._bounds},addLatLng:function(t,i){return i=i||this._defaultShape(),t=C(t),i.push(t),this._bounds.extend(t),this.redraw()},_setLatLngs:function(t){this._bounds=new T,this._latlngs=this._convertLatLngs(t)},_defaultShape:function(){return Dt(this._latlngs)?this._latlngs:this._latlngs[0]},_convertLatLngs:function(t){for(var i=[],e=Dt(t),n=0,o=t.length;n=2&&i[0]instanceof M&&i[0].equals(i[e-1])&&i.pop(),i},_setLatLngs:function(t){tn.prototype._setLatLngs.call(this,t),Dt(this._latlngs)&&(this._latlngs=[this._latlngs])},_defaultShape:function(){return Dt(this._latlngs[0])?this._latlngs[0]:this._latlngs[0][0]},_clipPoints:function(){var t=this._renderer._bounds,i=this.options.weight,e=new x(i,i);if(t=new P(t.min.subtract(e),t.max.add(e)),this._parts=[],this._pxBounds&&this._pxBounds.intersects(t))if(this.options.noClip)this._parts=this._rings;else for(var n,o=0,s=this._rings.length;ot.y!=n.y>t.y&&t.x<(n.x-e.x)*(t.y-e.y)/(n.y-e.y)+e.x&&(u=!u);return u||tn.prototype._containsPoint.call(this,t,!0)}}),nn=qe.extend({initialize:function(t,i){l(this,i),this._layers={},t&&this.addData(t)},addData:function(t){var i,e,n,o=ei(t)?t:t.features;if(o){for(i=0,e=o.length;i0?o:[i.src]}else{ei(this._url)||(this._url=[this._url]),i.autoplay=!!this.options.autoplay,i.loop=!!this.options.loop;for(var a=0;ao?(i.height=o+"px",pt(t,"leaflet-popup-scrolled")):mt(t,"leaflet-popup-scrolled"),this._containerWidth=this._container.offsetWidth},_animateZoom:function(t){var i=this._map._latLngToNewLayerPoint(this._latlng,t.zoom,t.center),e=this._getAnchor();Lt(this._container,i.add(e))},_adjustPan:function(){if(!(!this.options.autoPan||this._map._panAnim&&this._map._panAnim._inProgress)){var t=this._map,i=parseInt(at(this._container,"marginBottom"),10)||0,e=this._container.offsetHeight+i,n=this._containerWidth,o=new x(this._containerLeft,-e-this._containerBottom);o._add(Pt(this._container));var s=t.layerPointToContainerPoint(o),r=w(this.options.autoPanPadding),a=w(this.options.autoPanPaddingTopLeft||r),h=w(this.options.autoPanPaddingBottomRight||r),u=t.getSize(),l=0,c=0;s.x+n+h.x>u.x&&(l=s.x+n-u.x+h.x),s.x-l-a.x<0&&(l=s.x-a.x),s.y+e+h.y>u.y&&(c=s.y+e-u.y+h.y),s.y-c-a.y<0&&(c=s.y-a.y),(l||c)&&t.fire("autopanstart").panBy([l,c])}},_onCloseButtonClick:function(t){this._close(),Q(t)},_getAnchor:function(){return w(this._source&&this._source._getPopupAnchor?this._source._getPopupAnchor():[0,0])}});Le.mergeOptions({closePopupOnClick:!0}),Le.include({openPopup:function(t,i,e){return t instanceof un||(t=new un(e).setContent(t)),i&&t.setLatLng(i),this.hasLayer(t)?this:(this._popup&&this._popup.options.autoClose&&this.closePopup(),this._popup=t,this.addLayer(t))},closePopup:function(t){return t&&t!==this._popup||(t=this._popup,this._popup=null),t&&this.removeLayer(t),this}}),Ue.include({bindPopup:function(t,i){return t instanceof un?(l(t,i),this._popup=t,t._source=this):(this._popup&&!i||(this._popup=new un(i,this)),this._popup.setContent(t)),this._popupHandlersAdded||(this.on({click:this._openPopup,keypress:this._onKeyPress,remove:this.closePopup,move:this._movePopup}),this._popupHandlersAdded=!0),this},unbindPopup:function(){return this._popup&&(this.off({click:this._openPopup,keypress:this._onKeyPress,remove:this.closePopup,move:this._movePopup}),this._popupHandlersAdded=!1,this._popup=null),this},openPopup:function(t,i){if(t instanceof Ue||(i=t,t=this),t instanceof qe)for(var e in this._layers){t=this._layers[e];break}return i||(i=t.getCenter?t.getCenter():t.getLatLng()),this._popup&&this._map&&(this._popup._source=t,this._popup.update(),this._map.openPopup(this._popup,i)),this},closePopup:function(){return this._popup&&this._popup._close(),this},togglePopup:function(t){return this._popup&&(this._popup._map?this.closePopup():this.openPopup(t)),this},isPopupOpen:function(){return!!this._popup&&this._popup.isOpen()},setPopupContent:function(t){return this._popup&&this._popup.setContent(t),this},getPopup:function(){return this._popup},_openPopup:function(t){var i=t.layer||t.target;this._popup&&this._map&&(Q(t),i instanceof Je?this.openPopup(t.layer||t.target,t.latlng):this._map.hasLayer(this._popup)&&this._popup._source===i?this.closePopup():this.openPopup(i,t.latlng))},_movePopup:function(t){this._popup.setLatLng(t.latlng)},_onKeyPress:function(t){13===t.originalEvent.keyCode&&this._openPopup(t)}});var ln=hn.extend({options:{pane:"tooltipPane",offset:[0,0],direction:"auto",permanent:!1,sticky:!1,interactive:!1,opacity:.9},onAdd:function(t){hn.prototype.onAdd.call(this,t),this.setOpacity(this.options.opacity),t.fire("tooltipopen",{tooltip:this}),this._source&&this._source.fire("tooltipopen",{tooltip:this},!0)},onRemove:function(t){hn.prototype.onRemove.call(this,t),t.fire("tooltipclose",{tooltip:this}),this._source&&this._source.fire("tooltipclose",{tooltip:this},!0)},getEvents:function(){var t=hn.prototype.getEvents.call(this);return Vi&&!this.options.permanent&&(t.preclick=this._close),t},_close:function(){this._map&&this._map.closeTooltip(this)},_initLayout:function(){var t="leaflet-tooltip "+(this.options.className||"")+" leaflet-zoom-"+(this._zoomAnimated?"animated":"hide");this._contentNode=this._container=ht("div",t)},_updateLayout:function(){},_adjustPan:function(){},_setPosition:function(t){var i=this._map,e=this._container,n=i.latLngToContainerPoint(i.getCenter()),o=i.layerPointToContainerPoint(t),s=this.options.direction,r=e.offsetWidth,a=e.offsetHeight,h=w(this.options.offset),u=this._getAnchor();"top"===s?t=t.add(w(-r/2+h.x,-a+h.y+u.y,!0)):"bottom"===s?t=t.subtract(w(r/2-h.x,-h.y,!0)):"center"===s?t=t.subtract(w(r/2+h.x,a/2-u.y+h.y,!0)):"right"===s||"auto"===s&&o.xthis.options.maxZoom||en&&this._retainParent(o,s,r,n))},_retainChildren:function(t,i,e,n){for(var o=2*t;o<2*t+2;o++)for(var s=2*i;s<2*i+2;s++){var r=new x(o,s);r.z=e+1;var a=this._tileCoordsToKey(r),h=this._tiles[a];h&&h.active?h.retain=!0:(h&&h.loaded&&(h.retain=!0),e+1this.options.maxZoom||void 0!==this.options.minZoom&&o1)this._setView(t,e);else{for(var c=o.min.y;c<=o.max.y;c++)for(var _=o.min.x;_<=o.max.x;_++){var d=new x(_,c);if(d.z=this._tileZoom,this._isValidTile(d)){var p=this._tiles[this._tileCoordsToKey(d)];p?p.current=!0:r.push(d)}}if(r.sort(function(t,i){return t.distanceTo(s)-i.distanceTo(s)}),0!==r.length){this._loading||(this._loading=!0,this.fire("loading"));var m=document.createDocumentFragment();for(_=0;_e.max.x)||!i.wrapLat&&(t.ye.max.y))return!1}if(!this.options.bounds)return!0;var n=this._tileCoordsToBounds(t);return z(this.options.bounds).overlaps(n)},_keyToBounds:function(t){return this._tileCoordsToBounds(this._keyToTileCoords(t))},_tileCoordsToNwSe:function(t){var i=this._map,e=this.getTileSize(),n=t.scaleBy(e),o=n.add(e);return[i.unproject(n,t.z),i.unproject(o,t.z)]},_tileCoordsToBounds:function(t){var i=this._tileCoordsToNwSe(t),e=new T(i[0],i[1]);return this.options.noWrap||(e=this._map.wrapLatLngBounds(e)),e},_tileCoordsToKey:function(t){return t.x+":"+t.y+":"+t.z},_keyToTileCoords:function(t){var i=t.split(":"),e=new x(+i[0],+i[1]);return e.z=+i[2],e},_removeTile:function(t){var i=this._tiles[t];i&&(Ci||i.el.setAttribute("src",ni),ut(i.el),delete this._tiles[t],this.fire("tileunload",{tile:i.el,coords:this._keyToTileCoords(t)}))},_initTile:function(t){pt(t,"leaflet-tile");var i=this.getTileSize();t.style.width=i.x+"px",t.style.height=i.y+"px",t.onselectstart=r,t.onmousemove=r,Li&&this.options.opacity<1&&vt(t,this.options.opacity),Ti&&!zi&&(t.style.WebkitBackfaceVisibility="hidden")},_addTile:function(t,i){var n=this._getTilePos(t),o=this._tileCoordsToKey(t),s=this.createTile(this._wrapCoords(t),e(this._tileReady,this,t));this._initTile(s),this.createTile.length<2&&f(e(this._tileReady,this,t,null,s)),Lt(s,n),this._tiles[o]={el:s,coords:t,current:!0},i.appendChild(s),this.fire("tileloadstart",{tile:s,coords:t})},_tileReady:function(t,i,n){if(this._map){i&&this.fire("tileerror",{error:i,tile:n,coords:t});var o=this._tileCoordsToKey(t);(n=this._tiles[o])&&(n.loaded=+new Date,this._map._fadeAnimated?(vt(n.el,0),g(this._fadeFrame),this._fadeFrame=f(this._updateOpacity,this)):(n.active=!0,this._pruneTiles()),i||(pt(n.el,"leaflet-tile-loaded"),this.fire("tileload",{tile:n.el,coords:t})),this._noTilesToLoad()&&(this._loading=!1,this.fire("load"),Li||!this._map._fadeAnimated?f(this._pruneTiles,this):setTimeout(e(this._pruneTiles,this),250)))}},_getTilePos:function(t){return t.scaleBy(this.getTileSize()).subtract(this._level.origin)},_wrapCoords:function(t){var i=new x(this._wrapX?s(t.x,this._wrapX):t.x,this._wrapY?s(t.y,this._wrapY):t.y);return i.z=t.z,i},_pxBoundsToTileRange:function(t){var i=this.getTileSize();return new P(t.min.unscaleBy(i).floor(),t.max.unscaleBy(i).ceil().subtract([1,1]))},_noTilesToLoad:function(){for(var t in this._tiles)if(!this._tiles[t].loaded)return!1;return!0}}),dn=_n.extend({options:{minZoom:0,maxZoom:18,subdomains:"abc",errorTileUrl:"",zoomOffset:0,tms:!1,zoomReverse:!1,detectRetina:!1,crossOrigin:!1},initialize:function(t,i){this._url=t,(i=l(this,i)).detectRetina&&Ki&&i.maxZoom>0&&(i.tileSize=Math.floor(i.tileSize/2),i.zoomReverse?(i.zoomOffset--,i.minZoom++):(i.zoomOffset++,i.maxZoom--),i.minZoom=Math.max(0,i.minZoom)),"string"==typeof i.subdomains&&(i.subdomains=i.subdomains.split("")),Ti||this.on("tileunload",this._onTileRemove)},setUrl:function(t,i){return this._url=t,i||this.redraw(),this},createTile:function(t,i){var n=document.createElement("img");return V(n,"load",e(this._tileOnLoad,this,i,n)),V(n,"error",e(this._tileOnError,this,i,n)),this.options.crossOrigin&&(n.crossOrigin=""),n.alt="",n.setAttribute("role","presentation"),n.src=this.getTileUrl(t),n},getTileUrl:function(t){var e={r:Ki?"@2x":"",s:this._getSubdomain(t),x:t.x,y:t.y,z:this._getZoomForUrl()};if(this._map&&!this._map.options.crs.infinite){var n=this._globalTileRange.max.y-t.y;this.options.tms&&(e.y=n),e["-y"]=n}return _(this._url,i(e,this.options))},_tileOnLoad:function(t,i){Li?setTimeout(e(t,this,null,i),0):t(null,i)},_tileOnError:function(t,i,e){var n=this.options.errorTileUrl;n&&i.getAttribute("src")!==n&&(i.src=n),t(e,i)},_onTileRemove:function(t){t.tile.onload=null},_getZoomForUrl:function(){var t=this._tileZoom,i=this.options.maxZoom,e=this.options.zoomReverse,n=this.options.zoomOffset;return e&&(t=i-t),t+n},_getSubdomain:function(t){var i=Math.abs(t.x+t.y)%this.options.subdomains.length;return this.options.subdomains[i]},_abortLoading:function(){var t,i;for(t in this._tiles)this._tiles[t].coords.z!==this._tileZoom&&((i=this._tiles[t].el).onload=r,i.onerror=r,i.complete||(i.src=ni,ut(i),delete this._tiles[t]))}}),pn=dn.extend({defaultWmsParams:{service:"WMS",request:"GetMap",layers:"",styles:"",format:"image/jpeg",transparent:!1,version:"1.1.1"},options:{crs:null,uppercase:!1},initialize:function(t,e){this._url=t;var n=i({},this.defaultWmsParams);for(var o in e)o in this.options||(n[o]=e[o]);var s=(e=l(this,e)).detectRetina&&Ki?2:1,r=this.getTileSize();n.width=r.x*s,n.height=r.y*s,this.wmsParams=n},onAdd:function(t){this._crs=this.options.crs||t.options.crs,this._wmsVersion=parseFloat(this.wmsParams.version);var i=this._wmsVersion>=1.3?"crs":"srs";this.wmsParams[i]=this._crs.code,dn.prototype.onAdd.call(this,t)},getTileUrl:function(t){var i=this._tileCoordsToNwSe(t),e=this._crs,n=b(e.project(i[0]),e.project(i[1])),o=n.min,s=n.max,r=(this._wmsVersion>=1.3&&this._crs===He?[o.y,o.x,s.y,s.x]:[o.x,o.y,s.x,s.y]).join(","),a=L.TileLayer.prototype.getTileUrl.call(this,t);return a+c(this.wmsParams,a,this.options.uppercase)+(this.options.uppercase?"&BBOX=":"&bbox=")+r},setParams:function(t,e){return i(this.wmsParams,t),e||this.redraw(),this}});dn.WMS=pn,Yt.wms=function(t,i){return new pn(t,i)};var mn=Ue.extend({options:{padding:.1,tolerance:0},initialize:function(t){l(this,t),n(this),this._layers=this._layers||{}},onAdd:function(){this._container||(this._initContainer(),this._zoomAnimated&&pt(this._container,"leaflet-zoom-animated")),this.getPane().appendChild(this._container),this._update(),this.on("update",this._updatePaths,this)},onRemove:function(){this.off("update",this._updatePaths,this),this._destroyContainer()},getEvents:function(){var t={viewreset:this._reset,zoom:this._onZoom,moveend:this._update,zoomend:this._onZoomEnd};return this._zoomAnimated&&(t.zoomanim=this._onAnimZoom),t},_onAnimZoom:function(t){this._updateTransform(t.center,t.zoom)},_onZoom:function(){this._updateTransform(this._map.getCenter(),this._map.getZoom())},_updateTransform:function(t,i){var e=this._map.getZoomScale(i,this._zoom),n=Pt(this._container),o=this._map.getSize().multiplyBy(.5+this.options.padding),s=this._map.project(this._center,i),r=this._map.project(t,i).subtract(s),a=o.multiplyBy(-e).add(n).add(o).subtract(r);Ni?wt(this._container,a,e):Lt(this._container,a)},_reset:function(){this._update(),this._updateTransform(this._center,this._zoom);for(var t in this._layers)this._layers[t]._reset()},_onZoomEnd:function(){for(var t in this._layers)this._layers[t]._project()},_updatePaths:function(){for(var t in this._layers)this._layers[t]._update()},_update:function(){var t=this.options.padding,i=this._map.getSize(),e=this._map.containerPointToLayerPoint(i.multiplyBy(-t)).round();this._bounds=new P(e,e.add(i.multiplyBy(1+2*t)).round()),this._center=this._map.getCenter(),this._zoom=this._map.getZoom()}}),fn=mn.extend({getEvents:function(){var t=mn.prototype.getEvents.call(this);return t.viewprereset=this._onViewPreReset,t},_onViewPreReset:function(){this._postponeUpdatePaths=!0},onAdd:function(){mn.prototype.onAdd.call(this),this._draw()},_initContainer:function(){var t=this._container=document.createElement("canvas");V(t,"mousemove",o(this._onMouseMove,32,this),this),V(t,"click dblclick mousedown mouseup contextmenu",this._onClick,this),V(t,"mouseout",this._handleMouseOut,this),this._ctx=t.getContext("2d")},_destroyContainer:function(){delete this._ctx,ut(this._container),q(this._container),delete this._container},_updatePaths:function(){if(!this._postponeUpdatePaths){this._redrawBounds=null;for(var t in this._layers)this._layers[t]._update();this._redraw()}},_update:function(){if(!this._map._animatingZoom||!this._bounds){this._drawnLayers={},mn.prototype._update.call(this);var t=this._bounds,i=this._container,e=t.getSize(),n=Ki?2:1;Lt(i,t.min),i.width=n*e.x,i.height=n*e.y,i.style.width=e.x+"px",i.style.height=e.y+"px",Ki&&this._ctx.scale(2,2),this._ctx.translate(-t.min.x,-t.min.y),this.fire("update")}},_reset:function(){mn.prototype._reset.call(this),this._postponeUpdatePaths&&(this._postponeUpdatePaths=!1,this._updatePaths())},_initPath:function(t){this._updateDashArray(t),this._layers[n(t)]=t;var i=t._order={layer:t,prev:this._drawLast,next:null};this._drawLast&&(this._drawLast.next=i),this._drawLast=i,this._drawFirst=this._drawFirst||this._drawLast},_addPath:function(t){this._requestRedraw(t)},_removePath:function(t){var i=t._order,e=i.next,n=i.prev;e?e.prev=n:this._drawLast=n,n?n.next=e:this._drawFirst=e,delete t._order,delete this._layers[L.stamp(t)],this._requestRedraw(t)},_updatePath:function(t){this._extendRedrawBounds(t),t._project(),t._update(),this._requestRedraw(t)},_updateStyle:function(t){this._updateDashArray(t),this._requestRedraw(t)},_updateDashArray:function(t){if(t.options.dashArray){var i,e=t.options.dashArray.split(","),n=[];for(i=0;i')}}catch(t){return function(t){return document.createElement("<"+t+' xmlns="urn:schemas-microsoft.com:vml" class="lvml">')}}}(),vn={_initContainer:function(){this._container=ht("div","leaflet-vml-container")},_update:function(){this._map._animatingZoom||(mn.prototype._update.call(this),this.fire("update"))},_initPath:function(t){var i=t._container=gn("shape");pt(i,"leaflet-vml-shape "+(this.options.className||"")),i.coordsize="1 1",t._path=gn("path"),i.appendChild(t._path),this._updateStyle(t),this._layers[n(t)]=t},_addPath:function(t){var i=t._container;this._container.appendChild(i),t.options.interactive&&t.addInteractiveTarget(i)},_removePath:function(t){var i=t._container;ut(i),t.removeInteractiveTarget(i),delete this._layers[n(t)]},_updateStyle:function(t){var i=t._stroke,e=t._fill,n=t.options,o=t._container;o.stroked=!!n.stroke,o.filled=!!n.fill,n.stroke?(i||(i=t._stroke=gn("stroke")),o.appendChild(i),i.weight=n.weight+"px",i.color=n.color,i.opacity=n.opacity,n.dashArray?i.dashStyle=ei(n.dashArray)?n.dashArray.join(" "):n.dashArray.replace(/( *, *)/g," "):i.dashStyle="",i.endcap=n.lineCap.replace("butt","flat"),i.joinstyle=n.lineJoin):i&&(o.removeChild(i),t._stroke=null),n.fill?(e||(e=t._fill=gn("fill")),o.appendChild(e),e.color=n.fillColor||n.color,e.opacity=n.fillOpacity):e&&(o.removeChild(e),t._fill=null)},_updateCircle:function(t){var i=t._point.round(),e=Math.round(t._radius),n=Math.round(t._radiusY||e);this._setPath(t,t._empty()?"M0 0":"AL "+i.x+","+i.y+" "+e+","+n+" 0,23592600")},_setPath:function(t,i){t._path.v=i},_bringToFront:function(t){ct(t._container)},_bringToBack:function(t){_t(t._container)}},yn=Ji?gn:E,xn=mn.extend({getEvents:function(){var t=mn.prototype.getEvents.call(this);return t.zoomstart=this._onZoomStart,t},_initContainer:function(){this._container=yn("svg"),this._container.setAttribute("pointer-events","none"),this._rootGroup=yn("g"),this._container.appendChild(this._rootGroup)},_destroyContainer:function(){ut(this._container),q(this._container),delete this._container,delete this._rootGroup,delete this._svgSize},_onZoomStart:function(){this._update()},_update:function(){if(!this._map._animatingZoom||!this._bounds){mn.prototype._update.call(this);var t=this._bounds,i=t.getSize(),e=this._container;this._svgSize&&this._svgSize.equals(i)||(this._svgSize=i,e.setAttribute("width",i.x),e.setAttribute("height",i.y)),Lt(e,t.min),e.setAttribute("viewBox",[t.min.x,t.min.y,i.x,i.y].join(" ")),this.fire("update")}},_initPath:function(t){var i=t._path=yn("path");t.options.className&&pt(i,t.options.className),t.options.interactive&&pt(i,"leaflet-interactive"),this._updateStyle(t),this._layers[n(t)]=t},_addPath:function(t){this._rootGroup||this._initContainer(),this._rootGroup.appendChild(t._path),t.addInteractiveTarget(t._path)},_removePath:function(t){ut(t._path),t.removeInteractiveTarget(t._path),delete this._layers[n(t)]},_updatePath:function(t){t._project(),t._update()},_updateStyle:function(t){var i=t._path,e=t.options;i&&(e.stroke?(i.setAttribute("stroke",e.color),i.setAttribute("stroke-opacity",e.opacity),i.setAttribute("stroke-width",e.weight),i.setAttribute("stroke-linecap",e.lineCap),i.setAttribute("stroke-linejoin",e.lineJoin),e.dashArray?i.setAttribute("stroke-dasharray",e.dashArray):i.removeAttribute("stroke-dasharray"),e.dashOffset?i.setAttribute("stroke-dashoffset",e.dashOffset):i.removeAttribute("stroke-dashoffset")):i.setAttribute("stroke","none"),e.fill?(i.setAttribute("fill",e.fillColor||e.color),i.setAttribute("fill-opacity",e.fillOpacity),i.setAttribute("fill-rule",e.fillRule||"evenodd")):i.setAttribute("fill","none"))},_updatePoly:function(t,i){this._setPath(t,k(t._parts,i))},_updateCircle:function(t){var i=t._point,e=Math.max(Math.round(t._radius),1),n="a"+e+","+(Math.max(Math.round(t._radiusY),1)||e)+" 0 1,0 ",o=t._empty()?"M0 0":"M"+(i.x-e)+","+i.y+n+2*e+",0 "+n+2*-e+",0 ";this._setPath(t,o)},_setPath:function(t,i){t._path.setAttribute("d",i)},_bringToFront:function(t){ct(t._path)},_bringToBack:function(t){_t(t._path)}});Ji&&xn.include(vn),Le.include({getRenderer:function(t){var i=t.options.renderer||this._getPaneRenderer(t.options.pane)||this.options.renderer||this._renderer;return i||(i=this._renderer=this.options.preferCanvas&&Xt()||Jt()),this.hasLayer(i)||this.addLayer(i),i},_getPaneRenderer:function(t){if("overlayPane"===t||void 0===t)return!1;var i=this._paneRenderers[t];return void 0===i&&(i=xn&&Jt({pane:t})||fn&&Xt({pane:t}),this._paneRenderers[t]=i),i}});var wn=en.extend({initialize:function(t,i){en.prototype.initialize.call(this,this._boundsToLatLngs(t),i)},setBounds:function(t){return this.setLatLngs(this._boundsToLatLngs(t))},_boundsToLatLngs:function(t){return t=z(t),[t.getSouthWest(),t.getNorthWest(),t.getNorthEast(),t.getSouthEast()]}});xn.create=yn,xn.pointsToPath=k,nn.geometryToLayer=Wt,nn.coordsToLatLng=Ht,nn.coordsToLatLngs=Ft,nn.latLngToCoords=Ut,nn.latLngsToCoords=Vt,nn.getFeature=qt,nn.asFeature=Gt,Le.mergeOptions({boxZoom:!0});var Ln=Ze.extend({initialize:function(t){this._map=t,this._container=t._container,this._pane=t._panes.overlayPane,this._resetStateTimeout=0,t.on("unload",this._destroy,this)},addHooks:function(){V(this._container,"mousedown",this._onMouseDown,this)},removeHooks:function(){q(this._container,"mousedown",this._onMouseDown,this)},moved:function(){return this._moved},_destroy:function(){ut(this._pane),delete this._pane},_resetState:function(){this._resetStateTimeout=0,this._moved=!1},_clearDeferredResetState:function(){0!==this._resetStateTimeout&&(clearTimeout(this._resetStateTimeout),this._resetStateTimeout=0)},_onMouseDown:function(t){if(!t.shiftKey||1!==t.which&&1!==t.button)return!1;this._clearDeferredResetState(),this._resetState(),mi(),bt(),this._startPoint=this._map.mouseEventToContainerPoint(t),V(document,{contextmenu:Q,mousemove:this._onMouseMove,mouseup:this._onMouseUp,keydown:this._onKeyDown},this)},_onMouseMove:function(t){this._moved||(this._moved=!0,this._box=ht("div","leaflet-zoom-box",this._container),pt(this._container,"leaflet-crosshair"),this._map.fire("boxzoomstart")),this._point=this._map.mouseEventToContainerPoint(t);var i=new P(this._point,this._startPoint),e=i.getSize();Lt(this._box,i.min),this._box.style.width=e.x+"px",this._box.style.height=e.y+"px"},_finish:function(){this._moved&&(ut(this._box),mt(this._container,"leaflet-crosshair")),fi(),Tt(),q(document,{contextmenu:Q,mousemove:this._onMouseMove,mouseup:this._onMouseUp,keydown:this._onKeyDown},this)},_onMouseUp:function(t){if((1===t.which||1===t.button)&&(this._finish(),this._moved)){this._clearDeferredResetState(),this._resetStateTimeout=setTimeout(e(this._resetState,this),0);var i=new T(this._map.containerPointToLatLng(this._startPoint),this._map.containerPointToLatLng(this._point));this._map.fitBounds(i).fire("boxzoomend",{boxZoomBounds:i})}},_onKeyDown:function(t){27===t.keyCode&&this._finish()}});Le.addInitHook("addHandler","boxZoom",Ln),Le.mergeOptions({doubleClickZoom:!0});var Pn=Ze.extend({addHooks:function(){this._map.on("dblclick",this._onDoubleClick,this)},removeHooks:function(){this._map.off("dblclick",this._onDoubleClick,this)},_onDoubleClick:function(t){var i=this._map,e=i.getZoom(),n=i.options.zoomDelta,o=t.originalEvent.shiftKey?e-n:e+n;"center"===i.options.doubleClickZoom?i.setZoom(o):i.setZoomAround(t.containerPoint,o)}});Le.addInitHook("addHandler","doubleClickZoom",Pn),Le.mergeOptions({dragging:!0,inertia:!zi,inertiaDeceleration:3400,inertiaMaxSpeed:1/0,easeLinearity:.2,worldCopyJump:!1,maxBoundsViscosity:0});var bn=Ze.extend({addHooks:function(){if(!this._draggable){var t=this._map;this._draggable=new Be(t._mapPane,t._container),this._draggable.on({dragstart:this._onDragStart,drag:this._onDrag,dragend:this._onDragEnd},this),this._draggable.on("predrag",this._onPreDragLimit,this),t.options.worldCopyJump&&(this._draggable.on("predrag",this._onPreDragWrap,this),t.on("zoomend",this._onZoomEnd,this),t.whenReady(this._onZoomEnd,this))}pt(this._map._container,"leaflet-grab leaflet-touch-drag"),this._draggable.enable(),this._positions=[],this._times=[]},removeHooks:function(){mt(this._map._container,"leaflet-grab"),mt(this._map._container,"leaflet-touch-drag"),this._draggable.disable()},moved:function(){return this._draggable&&this._draggable._moved},moving:function(){return this._draggable&&this._draggable._moving},_onDragStart:function(){var t=this._map;if(t._stop(),this._map.options.maxBounds&&this._map.options.maxBoundsViscosity){var i=z(this._map.options.maxBounds);this._offsetLimit=b(this._map.latLngToContainerPoint(i.getNorthWest()).multiplyBy(-1),this._map.latLngToContainerPoint(i.getSouthEast()).multiplyBy(-1).add(this._map.getSize())),this._viscosity=Math.min(1,Math.max(0,this._map.options.maxBoundsViscosity))}else this._offsetLimit=null;t.fire("movestart").fire("dragstart"),t.options.inertia&&(this._positions=[],this._times=[])},_onDrag:function(t){if(this._map.options.inertia){var i=this._lastTime=+new Date,e=this._lastPos=this._draggable._absPos||this._draggable._newPos;this._positions.push(e),this._times.push(i),this._prunePositions(i)}this._map.fire("move",t).fire("drag",t)},_prunePositions:function(t){for(;this._positions.length>1&&t-this._times[0]>50;)this._positions.shift(),this._times.shift()},_onZoomEnd:function(){var t=this._map.getSize().divideBy(2),i=this._map.latLngToLayerPoint([0,0]);this._initialWorldOffset=i.subtract(t).x,this._worldWidth=this._map.getPixelWorldBounds().getSize().x},_viscousLimit:function(t,i){return t-(t-i)*this._viscosity},_onPreDragLimit:function(){if(this._viscosity&&this._offsetLimit){var t=this._draggable._newPos.subtract(this._draggable._startPos),i=this._offsetLimit;t.xi.max.x&&(t.x=this._viscousLimit(t.x,i.max.x)),t.y>i.max.y&&(t.y=this._viscousLimit(t.y,i.max.y)),this._draggable._newPos=this._draggable._startPos.add(t)}},_onPreDragWrap:function(){var t=this._worldWidth,i=Math.round(t/2),e=this._initialWorldOffset,n=this._draggable._newPos.x,o=(n-i+e)%t+i-e,s=(n+i+e)%t-i-e,r=Math.abs(o+e)0?s:-s))-i;this._delta=0,this._startTime=null,r&&("center"===t.options.scrollWheelZoom?t.setZoom(i+r):t.setZoomAround(this._lastMousePos,i+r))}});Le.addInitHook("addHandler","scrollWheelZoom",zn),Le.mergeOptions({tap:!0,tapTolerance:15});var Mn=Ze.extend({addHooks:function(){V(this._map._container,"touchstart",this._onDown,this)},removeHooks:function(){q(this._map._container,"touchstart",this._onDown,this)},_onDown:function(t){if(t.touches){if($(t),this._fireClick=!0,t.touches.length>1)return this._fireClick=!1,void clearTimeout(this._holdTimeout);var i=t.touches[0],n=i.target;this._startPos=this._newPos=new x(i.clientX,i.clientY),n.tagName&&"a"===n.tagName.toLowerCase()&&pt(n,"leaflet-active"),this._holdTimeout=setTimeout(e(function(){this._isTapValid()&&(this._fireClick=!1,this._onUp(),this._simulateEvent("contextmenu",i))},this),1e3),this._simulateEvent("mousedown",i),V(document,{touchmove:this._onMove,touchend:this._onUp},this)}},_onUp:function(t){if(clearTimeout(this._holdTimeout),q(document,{touchmove:this._onMove,touchend:this._onUp},this),this._fireClick&&t&&t.changedTouches){var i=t.changedTouches[0],e=i.target;e&&e.tagName&&"a"===e.tagName.toLowerCase()&&mt(e,"leaflet-active"),this._simulateEvent("mouseup",i),this._isTapValid()&&this._simulateEvent("click",i)}},_isTapValid:function(){return this._newPos.distanceTo(this._startPos)<=this._map.options.tapTolerance},_onMove:function(t){var i=t.touches[0];this._newPos=new x(i.clientX,i.clientY),this._simulateEvent("mousemove",i)},_simulateEvent:function(t,i){var e=document.createEvent("MouseEvents");e._simulated=!0,i.target._simulatedClick=!0,e.initMouseEvent(t,!0,!0,window,1,i.screenX,i.screenY,i.clientX,i.clientY,!1,!1,!1,!1,0,null),i.target.dispatchEvent(e)}});Vi&&!Ui&&Le.addInitHook("addHandler","tap",Mn),Le.mergeOptions({touchZoom:Vi&&!zi,bounceAtZoomLimits:!0});var Cn=Ze.extend({addHooks:function(){pt(this._map._container,"leaflet-touch-zoom"),V(this._map._container,"touchstart",this._onTouchStart,this)},removeHooks:function(){mt(this._map._container,"leaflet-touch-zoom"),q(this._map._container,"touchstart",this._onTouchStart,this)},_onTouchStart:function(t){var i=this._map;if(t.touches&&2===t.touches.length&&!i._animatingZoom&&!this._zooming){var e=i.mouseEventToContainerPoint(t.touches[0]),n=i.mouseEventToContainerPoint(t.touches[1]);this._centerPoint=i.getSize()._divideBy(2),this._startLatLng=i.containerPointToLatLng(this._centerPoint),"center"!==i.options.touchZoom&&(this._pinchStartLatLng=i.containerPointToLatLng(e.add(n)._divideBy(2))),this._startDist=e.distanceTo(n),this._startZoom=i.getZoom(),this._moved=!1,this._zooming=!0,i._stop(),V(document,"touchmove",this._onTouchMove,this),V(document,"touchend",this._onTouchEnd,this),$(t)}},_onTouchMove:function(t){if(t.touches&&2===t.touches.length&&this._zooming){var i=this._map,n=i.mouseEventToContainerPoint(t.touches[0]),o=i.mouseEventToContainerPoint(t.touches[1]),s=n.distanceTo(o)/this._startDist;if(this._zoom=i.getScaleZoom(s,this._startZoom),!i.options.bounceAtZoomLimits&&(this._zoomi.getMaxZoom()&&s>1)&&(this._zoom=i._limitZoom(this._zoom)),"center"===i.options.touchZoom){if(this._center=this._startLatLng,1===s)return}else{var r=n._add(o)._divideBy(2)._subtract(this._centerPoint);if(1===s&&0===r.x&&0===r.y)return;this._center=i.unproject(i.project(this._pinchStartLatLng,this._zoom).subtract(r),this._zoom)}this._moved||(i._moveStart(!0,!1),this._moved=!0),g(this._animRequest);var a=e(i._move,i,this._center,this._zoom,{pinch:!0,round:!1});this._animRequest=f(a,this,!0),$(t)}},_onTouchEnd:function(){this._moved&&this._zooming?(this._zooming=!1,g(this._animRequest),q(document,"touchmove",this._onTouchMove),q(document,"touchend",this._onTouchEnd),this._map.options.zoomAnimation?this._map._animateZoom(this._center,this._map._limitZoom(this._zoom),!0,this._map.options.zoomSnap):this._map._resetView(this._center,this._map._limitZoom(this._zoom))):this._zooming=!1}});Le.addInitHook("addHandler","touchZoom",Cn),Le.BoxZoom=Ln,Le.DoubleClickZoom=Pn,Le.Drag=bn,Le.Keyboard=Tn,Le.ScrollWheelZoom=zn,Le.Tap=Mn,Le.TouchZoom=Cn;var Zn=window.L;window.L=t,Object.freeze=$t,t.version="1.3.1+HEAD.ba6f97f",t.noConflict=function(){return window.L=Zn,this},t.Control=Pe,t.control=be,t.Browser=$i,t.Evented=ui,t.Mixin=Ee,t.Util=ai,t.Class=v,t.Handler=Ze,t.extend=i,t.bind=e,t.stamp=n,t.setOptions=l,t.DomEvent=de,t.DomUtil=xe,t.PosAnimation=we,t.Draggable=Be,t.LineUtil=Oe,t.PolyUtil=Re,t.Point=x,t.point=w,t.Bounds=P,t.bounds=b,t.Transformation=Z,t.transformation=S,t.Projection=je,t.LatLng=M,t.latLng=C,t.LatLngBounds=T,t.latLngBounds=z,t.CRS=ci,t.GeoJSON=nn,t.geoJSON=Kt,t.geoJson=sn,t.Layer=Ue,t.LayerGroup=Ve,t.layerGroup=function(t,i){return new Ve(t,i)},t.FeatureGroup=qe,t.featureGroup=function(t){return new qe(t)},t.ImageOverlay=rn,t.imageOverlay=function(t,i,e){return new rn(t,i,e)},t.VideoOverlay=an,t.videoOverlay=function(t,i,e){return new an(t,i,e)},t.DivOverlay=hn,t.Popup=un,t.popup=function(t,i){return new un(t,i)},t.Tooltip=ln,t.tooltip=function(t,i){return new ln(t,i)},t.Icon=Ge,t.icon=function(t){return new Ge(t)},t.DivIcon=cn,t.divIcon=function(t){return new cn(t)},t.Marker=Xe,t.marker=function(t,i){return new Xe(t,i)},t.TileLayer=dn,t.tileLayer=Yt,t.GridLayer=_n,t.gridLayer=function(t){return new _n(t)},t.SVG=xn,t.svg=Jt,t.Renderer=mn,t.Canvas=fn,t.canvas=Xt,t.Path=Je,t.CircleMarker=$e,t.circleMarker=function(t,i){return new $e(t,i)},t.Circle=Qe,t.circle=function(t,i,e){return new Qe(t,i,e)},t.Polyline=tn,t.polyline=function(t,i){return new tn(t,i)},t.Polygon=en,t.polygon=function(t,i){return new en(t,i)},t.Rectangle=wn,t.rectangle=function(t,i){return new wn(t,i)},t.Map=Le,t.map=function(t,i){return new Le(t,i)}}); \ No newline at end of file +!function(t,i){"object"==typeof exports&&"undefined"!=typeof module?i(exports):"function"==typeof define&&define.amd?define(["exports"],i):i(t.L={})}(this,function(t){"use strict";function i(t){var i,e,n,o;for(e=1,n=arguments.length;e=0}function I(t,i,e,n){return"touchstart"===i?O(t,e,n):"touchmove"===i?W(t,e,n):"touchend"===i&&H(t,e,n),this}function B(t,i,e){var n=t["_leaflet_"+i+e];return"touchstart"===i?t.removeEventListener(Qi,n,!1):"touchmove"===i?t.removeEventListener(te,n,!1):"touchend"===i&&(t.removeEventListener(ie,n,!1),t.removeEventListener(ee,n,!1)),this}function O(t,i,n){var o=e(function(t){if("mouse"!==t.pointerType&&t.MSPOINTER_TYPE_MOUSE&&t.pointerType!==t.MSPOINTER_TYPE_MOUSE){if(!(ne.indexOf(t.target.tagName)<0))return;$(t)}j(t,i)});t["_leaflet_touchstart"+n]=o,t.addEventListener(Qi,o,!1),se||(document.documentElement.addEventListener(Qi,R,!0),document.documentElement.addEventListener(te,D,!0),document.documentElement.addEventListener(ie,N,!0),document.documentElement.addEventListener(ee,N,!0),se=!0)}function R(t){oe[t.pointerId]=t,re++}function D(t){oe[t.pointerId]&&(oe[t.pointerId]=t)}function N(t){delete oe[t.pointerId],re--}function j(t,i){t.touches=[];for(var e in oe)t.touches.push(oe[e]);t.changedTouches=[t],i(t)}function W(t,i,e){var n=function(t){(t.pointerType!==t.MSPOINTER_TYPE_MOUSE&&"mouse"!==t.pointerType||0!==t.buttons)&&j(t,i)};t["_leaflet_touchmove"+e]=n,t.addEventListener(te,n,!1)}function H(t,i,e){var n=function(t){j(t,i)};t["_leaflet_touchend"+e]=n,t.addEventListener(ie,n,!1),t.addEventListener(ee,n,!1)}function F(t,i,e){function n(t){var i;if(Ui){if(!Pi||"mouse"===t.pointerType)return;i=re}else i=t.touches.length;if(!(i>1)){var e=Date.now(),n=e-(s||e);r=t.touches?t.touches[0]:t,a=n>0&&n<=h,s=e}}function o(t){if(a&&!r.cancelBubble){if(Ui){if(!Pi||"mouse"===t.pointerType)return;var e,n,o={};for(n in r)e=r[n],o[n]=e&&e.bind?e.bind(r):e;r=o}r.type="dblclick",i(r),s=null}}var s,r,a=!1,h=250;return t[ue+ae+e]=n,t[ue+he+e]=o,t[ue+"dblclick"+e]=i,t.addEventListener(ae,n,!1),t.addEventListener(he,o,!1),t.addEventListener("dblclick",i,!1),this}function U(t,i){var e=t[ue+ae+i],n=t[ue+he+i],o=t[ue+"dblclick"+i];return t.removeEventListener(ae,e,!1),t.removeEventListener(he,n,!1),Pi||t.removeEventListener("dblclick",o,!1),this}function V(t,i,e,n){if("object"==typeof i)for(var o in i)G(t,o,i[o],e);else for(var s=0,r=(i=u(i)).length;s100&&n<500||t.target._simulatedClick&&!t._simulated?Q(t):(pi=e,i(t))}function rt(t){return"string"==typeof t?document.getElementById(t):t}function at(t,i){var e=t.style[i]||t.currentStyle&&t.currentStyle[i];if((!e||"auto"===e)&&document.defaultView){var n=document.defaultView.getComputedStyle(t,null);e=n?n[i]:null}return"auto"===e?null:e}function ht(t,i,e){var n=document.createElement(t);return n.className=i||"",e&&e.appendChild(n),n}function ut(t){var i=t.parentNode;i&&i.removeChild(t)}function lt(t){for(;t.firstChild;)t.removeChild(t.firstChild)}function ct(t){var i=t.parentNode;i.lastChild!==t&&i.appendChild(t)}function _t(t){var i=t.parentNode;i.firstChild!==t&&i.insertBefore(t,i.firstChild)}function dt(t,i){if(void 0!==t.classList)return t.classList.contains(i);var e=gt(t);return e.length>0&&new RegExp("(^|\\s)"+i+"(\\s|$)").test(e)}function pt(t,i){if(void 0!==t.classList)for(var e=u(i),n=0,o=e.length;nh&&(s=r,h=a);h>e&&(i[s]=1,Et(t,i,e,n,s),Et(t,i,e,s,o))}function kt(t,i){for(var e=[t[0]],n=1,o=0,s=t.length;ni&&(e.push(t[n]),o=n);return oi.max.x&&(e|=2),t.yi.max.y&&(e|=8),e}function Ot(t,i){var e=i.x-t.x,n=i.y-t.y;return e*e+n*n}function Rt(t,i,e,n){var o,s=i.x,r=i.y,a=e.x-s,h=e.y-r,u=a*a+h*h;return u>0&&((o=((t.x-s)*a+(t.y-r)*h)/u)>1?(s=e.x,r=e.y):o>0&&(s+=a*o,r+=h*o)),a=t.x-s,h=t.y-r,n?a*a+h*h:new x(s,r)}function Dt(t){return!ei(t[0])||"object"!=typeof t[0][0]&&void 0!==t[0][0]}function Nt(t){return console.warn("Deprecated use of _flat, please use L.LineUtil.isFlat instead."),Dt(t)}function jt(t,i,e){var n,o,s,r,a,h,u,l,c,_=[1,4,2,8];for(o=0,u=t.length;o0?Math.floor(t):Math.ceil(t)};x.prototype={clone:function(){return new x(this.x,this.y)},add:function(t){return this.clone()._add(w(t))},_add:function(t){return this.x+=t.x,this.y+=t.y,this},subtract:function(t){return this.clone()._subtract(w(t))},_subtract:function(t){return this.x-=t.x,this.y-=t.y,this},divideBy:function(t){return this.clone()._divideBy(t)},_divideBy:function(t){return this.x/=t,this.y/=t,this},multiplyBy:function(t){return this.clone()._multiplyBy(t)},_multiplyBy:function(t){return this.x*=t,this.y*=t,this},scaleBy:function(t){return new x(this.x*t.x,this.y*t.y)},unscaleBy:function(t){return new x(this.x/t.x,this.y/t.y)},round:function(){return this.clone()._round()},_round:function(){return this.x=Math.round(this.x),this.y=Math.round(this.y),this},floor:function(){return this.clone()._floor()},_floor:function(){return this.x=Math.floor(this.x),this.y=Math.floor(this.y),this},ceil:function(){return this.clone()._ceil()},_ceil:function(){return this.x=Math.ceil(this.x),this.y=Math.ceil(this.y),this},trunc:function(){return this.clone()._trunc()},_trunc:function(){return this.x=li(this.x),this.y=li(this.y),this},distanceTo:function(t){var i=(t=w(t)).x-this.x,e=t.y-this.y;return Math.sqrt(i*i+e*e)},equals:function(t){return(t=w(t)).x===this.x&&t.y===this.y},contains:function(t){return t=w(t),Math.abs(t.x)<=Math.abs(this.x)&&Math.abs(t.y)<=Math.abs(this.y)},toString:function(){return"Point("+a(this.x)+", "+a(this.y)+")"}},P.prototype={extend:function(t){return t=w(t),this.min||this.max?(this.min.x=Math.min(t.x,this.min.x),this.max.x=Math.max(t.x,this.max.x),this.min.y=Math.min(t.y,this.min.y),this.max.y=Math.max(t.y,this.max.y)):(this.min=t.clone(),this.max=t.clone()),this},getCenter:function(t){return new x((this.min.x+this.max.x)/2,(this.min.y+this.max.y)/2,t)},getBottomLeft:function(){return new x(this.min.x,this.max.y)},getTopRight:function(){return new x(this.max.x,this.min.y)},getTopLeft:function(){return this.min},getBottomRight:function(){return this.max},getSize:function(){return this.max.subtract(this.min)},contains:function(t){var i,e;return(t="number"==typeof t[0]||t instanceof x?w(t):b(t))instanceof P?(i=t.min,e=t.max):i=e=t,i.x>=this.min.x&&e.x<=this.max.x&&i.y>=this.min.y&&e.y<=this.max.y},intersects:function(t){t=b(t);var i=this.min,e=this.max,n=t.min,o=t.max,s=o.x>=i.x&&n.x<=e.x,r=o.y>=i.y&&n.y<=e.y;return s&&r},overlaps:function(t){t=b(t);var i=this.min,e=this.max,n=t.min,o=t.max,s=o.x>i.x&&n.xi.y&&n.y=n.lat&&e.lat<=o.lat&&i.lng>=n.lng&&e.lng<=o.lng},intersects:function(t){t=z(t);var i=this._southWest,e=this._northEast,n=t.getSouthWest(),o=t.getNorthEast(),s=o.lat>=i.lat&&n.lat<=e.lat,r=o.lng>=i.lng&&n.lng<=e.lng;return s&&r},overlaps:function(t){t=z(t);var i=this._southWest,e=this._northEast,n=t.getSouthWest(),o=t.getNorthEast(),s=o.lat>i.lat&&n.lati.lng&&n.lng1,Yi=!!document.createElement("canvas").getContext,Xi=!(!document.createElementNS||!E("svg").createSVGRect),Ji=!Xi&&function(){try{var t=document.createElement("div");t.innerHTML='';var i=t.firstChild;return i.style.behavior="url(#default#VML)",i&&"object"==typeof i.adj}catch(t){return!1}}(),$i=(Object.freeze||Object)({ie:wi,ielt9:Li,edge:Pi,webkit:bi,android:Ti,android23:zi,androidStock:Ci,opera:Zi,chrome:Si,gecko:Ei,safari:ki,phantom:Ai,opera12:Ii,win:Bi,ie3d:Oi,webkit3d:Ri,gecko3d:Di,any3d:Ni,mobile:ji,mobileWebkit:Wi,mobileWebkit3d:Hi,msPointer:Fi,pointer:Ui,touch:Vi,mobileOpera:qi,mobileGecko:Gi,retina:Ki,canvas:Yi,svg:Xi,vml:Ji}),Qi=Fi?"MSPointerDown":"pointerdown",te=Fi?"MSPointerMove":"pointermove",ie=Fi?"MSPointerUp":"pointerup",ee=Fi?"MSPointerCancel":"pointercancel",ne=["INPUT","SELECT","OPTION"],oe={},se=!1,re=0,ae=Fi?"MSPointerDown":Ui?"pointerdown":"touchstart",he=Fi?"MSPointerUp":Ui?"pointerup":"touchend",ue="_leaflet_",le="_leaflet_events",ce=Bi&&Si?2*window.devicePixelRatio:Ei?window.devicePixelRatio:1,_e={},de=(Object.freeze||Object)({on:V,off:q,stopPropagation:Y,disableScrollPropagation:X,disableClickPropagation:J,preventDefault:$,stop:Q,getMousePosition:tt,getWheelDelta:it,fakeStop:et,skipped:nt,isExternalTarget:ot,addListener:V,removeListener:q}),pe=xt(["transform","WebkitTransform","OTransform","MozTransform","msTransform"]),me=xt(["webkitTransition","transition","OTransition","MozTransition","msTransition"]),fe="webkitTransition"===me||"OTransition"===me?me+"End":"transitionend";if("onselectstart"in document)mi=function(){V(window,"selectstart",$)},fi=function(){q(window,"selectstart",$)};else{var ge=xt(["userSelect","WebkitUserSelect","OUserSelect","MozUserSelect","msUserSelect"]);mi=function(){if(ge){var t=document.documentElement.style;gi=t[ge],t[ge]="none"}},fi=function(){ge&&(document.documentElement.style[ge]=gi,gi=void 0)}}var ve,ye,xe=(Object.freeze||Object)({TRANSFORM:pe,TRANSITION:me,TRANSITION_END:fe,get:rt,getStyle:at,create:ht,remove:ut,empty:lt,toFront:ct,toBack:_t,hasClass:dt,addClass:pt,removeClass:mt,setClass:ft,getClass:gt,setOpacity:vt,testProp:xt,setTransform:wt,setPosition:Lt,getPosition:Pt,disableTextSelection:mi,enableTextSelection:fi,disableImageDrag:bt,enableImageDrag:Tt,preventOutline:zt,restoreOutline:Mt}),we=ui.extend({run:function(t,i,e,n){this.stop(),this._el=t,this._inProgress=!0,this._duration=e||.25,this._easeOutPower=1/Math.max(n||.5,.2),this._startPos=Pt(t),this._offset=i.subtract(this._startPos),this._startTime=+new Date,this.fire("start"),this._animate()},stop:function(){this._inProgress&&(this._step(!0),this._complete())},_animate:function(){this._animId=f(this._animate,this),this._step()},_step:function(t){var i=+new Date-this._startTime,e=1e3*this._duration;ithis.options.maxZoom)?this.setZoom(t):this},panInsideBounds:function(t,i){this._enforcingBounds=!0;var e=this.getCenter(),n=this._limitCenter(e,this._zoom,z(t));return e.equals(n)||this.panTo(n,i),this._enforcingBounds=!1,this},invalidateSize:function(t){if(!this._loaded)return this;t=i({animate:!1,pan:!0},!0===t?{animate:!0}:t);var n=this.getSize();this._sizeChanged=!0,this._lastCenter=null;var o=this.getSize(),s=n.divideBy(2).round(),r=o.divideBy(2).round(),a=s.subtract(r);return a.x||a.y?(t.animate&&t.pan?this.panBy(a):(t.pan&&this._rawPanBy(a),this.fire("move"),t.debounceMoveend?(clearTimeout(this._sizeTimer),this._sizeTimer=setTimeout(e(this.fire,this,"moveend"),200)):this.fire("moveend")),this.fire("resize",{oldSize:n,newSize:o})):this},stop:function(){return this.setZoom(this._limitZoom(this._zoom)),this.options.zoomSnap||this.fire("viewreset"),this._stop()},locate:function(t){if(t=this._locateOptions=i({timeout:1e4,watch:!1},t),!("geolocation"in navigator))return this._handleGeolocationError({code:0,message:"Geolocation not supported."}),this;var n=e(this._handleGeolocationResponse,this),o=e(this._handleGeolocationError,this);return t.watch?this._locationWatchId=navigator.geolocation.watchPosition(n,o,t):navigator.geolocation.getCurrentPosition(n,o,t),this},stopLocate:function(){return navigator.geolocation&&navigator.geolocation.clearWatch&&navigator.geolocation.clearWatch(this._locationWatchId),this._locateOptions&&(this._locateOptions.setView=!1),this},_handleGeolocationError:function(t){var i=t.code,e=t.message||(1===i?"permission denied":2===i?"position unavailable":"timeout");this._locateOptions.setView&&!this._loaded&&this.fitWorld(),this.fire("locationerror",{code:i,message:"Geolocation error: "+e+"."})},_handleGeolocationResponse:function(t){var i=new M(t.coords.latitude,t.coords.longitude),e=i.toBounds(t.coords.accuracy),n=this._locateOptions;if(n.setView){var o=this.getBoundsZoom(e);this.setView(i,n.maxZoom?Math.min(o,n.maxZoom):o)}var s={latlng:i,bounds:e,timestamp:t.timestamp};for(var r in t.coords)"number"==typeof t.coords[r]&&(s[r]=t.coords[r]);this.fire("locationfound",s)},addHandler:function(t,i){if(!i)return this;var e=this[t]=new i(this);return this._handlers.push(e),this.options[t]&&e.enable(),this},remove:function(){if(this._initEvents(!0),this._containerId!==this._container._leaflet_id)throw new Error("Map container is being reused by another instance");try{delete this._container._leaflet_id,delete this._containerId}catch(t){this._container._leaflet_id=void 0,this._containerId=void 0}void 0!==this._locationWatchId&&this.stopLocate(),this._stop(),ut(this._mapPane),this._clearControlPos&&this._clearControlPos(),this._clearHandlers(),this._loaded&&this.fire("unload");var t;for(t in this._layers)this._layers[t].remove();for(t in this._panes)ut(this._panes[t]);return this._layers=[],this._panes=[],delete this._mapPane,delete this._renderer,this},createPane:function(t,i){var e=ht("div","leaflet-pane"+(t?" leaflet-"+t.replace("Pane","")+"-pane":""),i||this._mapPane);return t&&(this._panes[t]=e),e},getCenter:function(){return this._checkIfLoaded(),this._lastCenter&&!this._moved()?this._lastCenter:this.layerPointToLatLng(this._getCenterLayerPoint())},getZoom:function(){return this._zoom},getBounds:function(){var t=this.getPixelBounds();return new T(this.unproject(t.getBottomLeft()),this.unproject(t.getTopRight()))},getMinZoom:function(){return void 0===this.options.minZoom?this._layersMinZoom||0:this.options.minZoom},getMaxZoom:function(){return void 0===this.options.maxZoom?void 0===this._layersMaxZoom?1/0:this._layersMaxZoom:this.options.maxZoom},getBoundsZoom:function(t,i,e){t=z(t),e=w(e||[0,0]);var n=this.getZoom()||0,o=this.getMinZoom(),s=this.getMaxZoom(),r=t.getNorthWest(),a=t.getSouthEast(),h=this.getSize().subtract(e),u=b(this.project(a,n),this.project(r,n)).getSize(),l=Ni?this.options.zoomSnap:1,c=h.x/u.x,_=h.y/u.y,d=i?Math.max(c,_):Math.min(c,_);return n=this.getScaleZoom(d,n),l&&(n=Math.round(n/(l/100))*(l/100),n=i?Math.ceil(n/l)*l:Math.floor(n/l)*l),Math.max(o,Math.min(s,n))},getSize:function(){return this._size&&!this._sizeChanged||(this._size=new x(this._container.clientWidth||0,this._container.clientHeight||0),this._sizeChanged=!1),this._size.clone()},getPixelBounds:function(t,i){var e=this._getTopLeftPoint(t,i);return new P(e,e.add(this.getSize()))},getPixelOrigin:function(){return this._checkIfLoaded(),this._pixelOrigin},getPixelWorldBounds:function(t){return this.options.crs.getProjectedBounds(void 0===t?this.getZoom():t)},getPane:function(t){return"string"==typeof t?this._panes[t]:t},getPanes:function(){return this._panes},getContainer:function(){return this._container},getZoomScale:function(t,i){var e=this.options.crs;return i=void 0===i?this._zoom:i,e.scale(t)/e.scale(i)},getScaleZoom:function(t,i){var e=this.options.crs;i=void 0===i?this._zoom:i;var n=e.zoom(t*e.scale(i));return isNaN(n)?1/0:n},project:function(t,i){return i=void 0===i?this._zoom:i,this.options.crs.latLngToPoint(C(t),i)},unproject:function(t,i){return i=void 0===i?this._zoom:i,this.options.crs.pointToLatLng(w(t),i)},layerPointToLatLng:function(t){var i=w(t).add(this.getPixelOrigin());return this.unproject(i)},latLngToLayerPoint:function(t){return this.project(C(t))._round()._subtract(this.getPixelOrigin())},wrapLatLng:function(t){return this.options.crs.wrapLatLng(C(t))},wrapLatLngBounds:function(t){return this.options.crs.wrapLatLngBounds(z(t))},distance:function(t,i){return this.options.crs.distance(C(t),C(i))},containerPointToLayerPoint:function(t){return w(t).subtract(this._getMapPanePos())},layerPointToContainerPoint:function(t){return w(t).add(this._getMapPanePos())},containerPointToLatLng:function(t){var i=this.containerPointToLayerPoint(w(t));return this.layerPointToLatLng(i)},latLngToContainerPoint:function(t){return this.layerPointToContainerPoint(this.latLngToLayerPoint(C(t)))},mouseEventToContainerPoint:function(t){return tt(t,this._container)},mouseEventToLayerPoint:function(t){return this.containerPointToLayerPoint(this.mouseEventToContainerPoint(t))},mouseEventToLatLng:function(t){return this.layerPointToLatLng(this.mouseEventToLayerPoint(t))},_initContainer:function(t){var i=this._container=rt(t);if(!i)throw new Error("Map container not found.");if(i._leaflet_id)throw new Error("Map container is already initialized.");V(i,"scroll",this._onScroll,this),this._containerId=n(i)},_initLayout:function(){var t=this._container;this._fadeAnimated=this.options.fadeAnimation&&Ni,pt(t,"leaflet-container"+(Vi?" leaflet-touch":"")+(Ki?" leaflet-retina":"")+(Li?" leaflet-oldie":"")+(ki?" leaflet-safari":"")+(this._fadeAnimated?" leaflet-fade-anim":""));var i=at(t,"position");"absolute"!==i&&"relative"!==i&&"fixed"!==i&&(t.style.position="relative"),this._initPanes(),this._initControlPos&&this._initControlPos()},_initPanes:function(){var t=this._panes={};this._paneRenderers={},this._mapPane=this.createPane("mapPane",this._container),Lt(this._mapPane,new x(0,0)),this.createPane("tilePane"),this.createPane("shadowPane"),this.createPane("overlayPane"),this.createPane("markerPane"),this.createPane("tooltipPane"),this.createPane("popupPane"),this.options.markerZoomAnimation||(pt(t.markerPane,"leaflet-zoom-hide"),pt(t.shadowPane,"leaflet-zoom-hide"))},_resetView:function(t,i){Lt(this._mapPane,new x(0,0));var e=!this._loaded;this._loaded=!0,i=this._limitZoom(i),this.fire("viewprereset");var n=this._zoom!==i;this._moveStart(n,!1)._move(t,i)._moveEnd(n),this.fire("viewreset"),e&&this.fire("load")},_moveStart:function(t,i){return t&&this.fire("zoomstart"),i||this.fire("movestart"),this},_move:function(t,i,e){void 0===i&&(i=this._zoom);var n=this._zoom!==i;return this._zoom=i,this._lastCenter=t,this._pixelOrigin=this._getNewPixelOrigin(t),(n||e&&e.pinch)&&this.fire("zoom",e),this.fire("move",e)},_moveEnd:function(t){return t&&this.fire("zoomend"),this.fire("moveend")},_stop:function(){return g(this._flyToFrame),this._panAnim&&this._panAnim.stop(),this},_rawPanBy:function(t){Lt(this._mapPane,this._getMapPanePos().subtract(t))},_getZoomSpan:function(){return this.getMaxZoom()-this.getMinZoom()},_panInsideMaxBounds:function(){this._enforcingBounds||this.panInsideBounds(this.options.maxBounds)},_checkIfLoaded:function(){if(!this._loaded)throw new Error("Set map center and zoom first.")},_initEvents:function(t){this._targets={},this._targets[n(this._container)]=this;var i=t?q:V;i(this._container,"click dblclick mousedown mouseup mouseover mouseout mousemove contextmenu keypress",this._handleDOMEvent,this),this.options.trackResize&&i(window,"resize",this._onResize,this),Ni&&this.options.transform3DLimit&&(t?this.off:this.on).call(this,"moveend",this._onMoveEnd)},_onResize:function(){g(this._resizeRequest),this._resizeRequest=f(function(){this.invalidateSize({debounceMoveend:!0})},this)},_onScroll:function(){this._container.scrollTop=0,this._container.scrollLeft=0},_onMoveEnd:function(){var t=this._getMapPanePos();Math.max(Math.abs(t.x),Math.abs(t.y))>=this.options.transform3DLimit&&this._resetView(this.getCenter(),this.getZoom())},_findEventTargets:function(t,i){for(var e,o=[],s="mouseout"===i||"mouseover"===i,r=t.target||t.srcElement,a=!1;r;){if((e=this._targets[n(r)])&&("click"===i||"preclick"===i)&&!t._simulated&&this._draggableMoved(e)){a=!0;break}if(e&&e.listens(i,!0)){if(s&&!ot(r,t))break;if(o.push(e),s)break}if(r===this._container)break;r=r.parentNode}return o.length||a||s||!ot(r,t)||(o=[this]),o},_handleDOMEvent:function(t){if(this._loaded&&!nt(t)){var i=t.type;"mousedown"!==i&&"keypress"!==i||zt(t.target||t.srcElement),this._fireDOMEvent(t,i)}},_mouseEvents:["click","dblclick","mouseover","mouseout","contextmenu"],_fireDOMEvent:function(t,e,n){if("click"===t.type){var o=i({},t);o.type="preclick",this._fireDOMEvent(o,o.type,n)}if(!t._stopped&&(n=(n||[]).concat(this._findEventTargets(t,e))).length){var s=n[0];"contextmenu"===e&&s.listens(e,!0)&&$(t);var r={originalEvent:t};if("keypress"!==t.type){var a=s.getLatLng&&(!s._radius||s._radius<=10);r.containerPoint=a?this.latLngToContainerPoint(s.getLatLng()):this.mouseEventToContainerPoint(t),r.layerPoint=this.containerPointToLayerPoint(r.containerPoint),r.latlng=a?s.getLatLng():this.layerPointToLatLng(r.layerPoint)}for(var h=0;h0?Math.round(t-i)/2:Math.max(0,Math.ceil(t))-Math.max(0,Math.floor(i))},_limitZoom:function(t){var i=this.getMinZoom(),e=this.getMaxZoom(),n=Ni?this.options.zoomSnap:1;return n&&(t=Math.round(t/n)*n),Math.max(i,Math.min(e,t))},_onPanTransitionStep:function(){this.fire("move")},_onPanTransitionEnd:function(){mt(this._mapPane,"leaflet-pan-anim"),this.fire("moveend")},_tryAnimatedPan:function(t,i){var e=this._getCenterOffset(t)._trunc();return!(!0!==(i&&i.animate)&&!this.getSize().contains(e))&&(this.panBy(e,i),!0)},_createAnimProxy:function(){var t=this._proxy=ht("div","leaflet-proxy leaflet-zoom-animated");this._panes.mapPane.appendChild(t),this.on("zoomanim",function(t){var i=pe,e=this._proxy.style[i];wt(this._proxy,this.project(t.center,t.zoom),this.getZoomScale(t.zoom,1)),e===this._proxy.style[i]&&this._animatingZoom&&this._onZoomTransitionEnd()},this),this.on("load moveend",function(){var t=this.getCenter(),i=this.getZoom();wt(this._proxy,this.project(t,i),this.getZoomScale(i,1))},this),this._on("unload",this._destroyAnimProxy,this)},_destroyAnimProxy:function(){ut(this._proxy),delete this._proxy},_catchTransitionEnd:function(t){this._animatingZoom&&t.propertyName.indexOf("transform")>=0&&this._onZoomTransitionEnd()},_nothingToAnimate:function(){return!this._container.getElementsByClassName("leaflet-zoom-animated").length},_tryAnimatedZoom:function(t,i,e){if(this._animatingZoom)return!0;if(e=e||{},!this._zoomAnimated||!1===e.animate||this._nothingToAnimate()||Math.abs(i-this._zoom)>this.options.zoomAnimationThreshold)return!1;var n=this.getZoomScale(i),o=this._getCenterOffset(t)._divideBy(1-1/n);return!(!0!==e.animate&&!this.getSize().contains(o))&&(f(function(){this._moveStart(!0,!1)._animateZoom(t,i,!0)},this),!0)},_animateZoom:function(t,i,n,o){this._mapPane&&(n&&(this._animatingZoom=!0,this._animateToCenter=t,this._animateToZoom=i,pt(this._mapPane,"leaflet-zoom-anim")),this.fire("zoomanim",{center:t,zoom:i,noUpdate:o}),setTimeout(e(this._onZoomTransitionEnd,this),250))},_onZoomTransitionEnd:function(){this._animatingZoom&&(this._mapPane&&mt(this._mapPane,"leaflet-zoom-anim"),this._animatingZoom=!1,this._move(this._animateToCenter,this._animateToZoom),f(function(){this._moveEnd(!0)},this))}}),Pe=v.extend({options:{position:"topright"},initialize:function(t){l(this,t)},getPosition:function(){return this.options.position},setPosition:function(t){var i=this._map;return i&&i.removeControl(this),this.options.position=t,i&&i.addControl(this),this},getContainer:function(){return this._container},addTo:function(t){this.remove(),this._map=t;var i=this._container=this.onAdd(t),e=this.getPosition(),n=t._controlCorners[e];return pt(i,"leaflet-control"),-1!==e.indexOf("bottom")?n.insertBefore(i,n.firstChild):n.appendChild(i),this},remove:function(){return this._map?(ut(this._container),this.onRemove&&this.onRemove(this._map),this._map=null,this):this},_refocusOnMap:function(t){this._map&&t&&t.screenX>0&&t.screenY>0&&this._map.getContainer().focus()}}),be=function(t){return new Pe(t)};Le.include({addControl:function(t){return t.addTo(this),this},removeControl:function(t){return t.remove(),this},_initControlPos:function(){function t(t,o){var s=e+t+" "+e+o;i[t+o]=ht("div",s,n)}var i=this._controlCorners={},e="leaflet-",n=this._controlContainer=ht("div",e+"control-container",this._container);t("top","left"),t("top","right"),t("bottom","left"),t("bottom","right")},_clearControlPos:function(){for(var t in this._controlCorners)ut(this._controlCorners[t]);ut(this._controlContainer),delete this._controlCorners,delete this._controlContainer}});var Te=Pe.extend({options:{collapsed:!0,position:"topright",autoZIndex:!0,hideSingleBase:!1,sortLayers:!1,sortFunction:function(t,i,e,n){return e1,this._baseLayersList.style.display=t?"":"none"),this._separator.style.display=i&&t?"":"none",this},_onLayerChange:function(t){this._handlingClick||this._update();var i=this._getLayer(n(t.target)),e=i.overlay?"add"===t.type?"overlayadd":"overlayremove":"add"===t.type?"baselayerchange":null;e&&this._map.fire(e,i)},_createRadioElement:function(t,i){var e='",n=document.createElement("div");return n.innerHTML=e,n.firstChild},_addItem:function(t){var i,e=document.createElement("label"),o=this._map.hasLayer(t.layer);t.overlay?((i=document.createElement("input")).type="checkbox",i.className="leaflet-control-layers-selector",i.defaultChecked=o):i=this._createRadioElement("leaflet-base-layers",o),this._layerControlInputs.push(i),i.layerId=n(t.layer),V(i,"click",this._onInputClick,this);var s=document.createElement("span");s.innerHTML=" "+t.name;var r=document.createElement("div");return e.appendChild(r),r.appendChild(i),r.appendChild(s),(t.overlay?this._overlaysList:this._baseLayersList).appendChild(e),this._checkDisabledLayers(),e},_onInputClick:function(){var t,i,e=this._layerControlInputs,n=[],o=[];this._handlingClick=!0;for(var s=e.length-1;s>=0;s--)t=e[s],i=this._getLayer(t.layerId).layer,t.checked?n.push(i):t.checked||o.push(i);for(s=0;s=0;o--)t=e[o],i=this._getLayer(t.layerId).layer,t.disabled=void 0!==i.options.minZoom&&ni.options.maxZoom},_expandIfNotCollapsed:function(){return this._map&&!this.options.collapsed&&this.expand(),this},_expand:function(){return this.expand()},_collapse:function(){return this.collapse()}}),ze=Pe.extend({options:{position:"topleft",zoomInText:"+",zoomInTitle:"Zoom in",zoomOutText:"−",zoomOutTitle:"Zoom out"},onAdd:function(t){var i="leaflet-control-zoom",e=ht("div",i+" leaflet-bar"),n=this.options;return this._zoomInButton=this._createButton(n.zoomInText,n.zoomInTitle,i+"-in",e,this._zoomIn),this._zoomOutButton=this._createButton(n.zoomOutText,n.zoomOutTitle,i+"-out",e,this._zoomOut),this._updateDisabled(),t.on("zoomend zoomlevelschange",this._updateDisabled,this),e},onRemove:function(t){t.off("zoomend zoomlevelschange",this._updateDisabled,this)},disable:function(){return this._disabled=!0,this._updateDisabled(),this},enable:function(){return this._disabled=!1,this._updateDisabled(),this},_zoomIn:function(t){!this._disabled&&this._map._zoomthis._map.getMinZoom()&&this._map.zoomOut(this._map.options.zoomDelta*(t.shiftKey?3:1))},_createButton:function(t,i,e,n,o){var s=ht("a",e,n);return s.innerHTML=t,s.href="#",s.title=i,s.setAttribute("role","button"),s.setAttribute("aria-label",i),J(s),V(s,"click",Q),V(s,"click",o,this),V(s,"click",this._refocusOnMap,this),s},_updateDisabled:function(){var t=this._map,i="leaflet-disabled";mt(this._zoomInButton,i),mt(this._zoomOutButton,i),(this._disabled||t._zoom===t.getMinZoom())&&pt(this._zoomOutButton,i),(this._disabled||t._zoom===t.getMaxZoom())&&pt(this._zoomInButton,i)}});Le.mergeOptions({zoomControl:!0}),Le.addInitHook(function(){this.options.zoomControl&&(this.zoomControl=new ze,this.addControl(this.zoomControl))});var Me=Pe.extend({options:{position:"bottomleft",maxWidth:100,metric:!0,imperial:!0},onAdd:function(t){var i=ht("div","leaflet-control-scale"),e=this.options;return this._addScales(e,"leaflet-control-scale-line",i),t.on(e.updateWhenIdle?"moveend":"move",this._update,this),t.whenReady(this._update,this),i},onRemove:function(t){t.off(this.options.updateWhenIdle?"moveend":"move",this._update,this)},_addScales:function(t,i,e){t.metric&&(this._mScale=ht("div",i,e)),t.imperial&&(this._iScale=ht("div",i,e))},_update:function(){var t=this._map,i=t.getSize().y/2,e=t.distance(t.containerPointToLatLng([0,i]),t.containerPointToLatLng([this.options.maxWidth,i]));this._updateScales(e)},_updateScales:function(t){this.options.metric&&t&&this._updateMetric(t),this.options.imperial&&t&&this._updateImperial(t)},_updateMetric:function(t){var i=this._getRoundNum(t),e=i<1e3?i+" m":i/1e3+" km";this._updateScale(this._mScale,e,i/t)},_updateImperial:function(t){var i,e,n,o=3.2808399*t;o>5280?(i=o/5280,e=this._getRoundNum(i),this._updateScale(this._iScale,e+" mi",e/i)):(n=this._getRoundNum(o),this._updateScale(this._iScale,n+" ft",n/o))},_updateScale:function(t,i,e){t.style.width=Math.round(this.options.maxWidth*e)+"px",t.innerHTML=i},_getRoundNum:function(t){var i=Math.pow(10,(Math.floor(t)+"").length-1),e=t/i;return e=e>=10?10:e>=5?5:e>=3?3:e>=2?2:1,i*e}}),Ce=Pe.extend({options:{position:"bottomright",prefix:'Leaflet'},initialize:function(t){l(this,t),this._attributions={}},onAdd:function(t){t.attributionControl=this,this._container=ht("div","leaflet-control-attribution"),J(this._container);for(var i in t._layers)t._layers[i].getAttribution&&this.addAttribution(t._layers[i].getAttribution());return this._update(),this._container},setPrefix:function(t){return this.options.prefix=t,this._update(),this},addAttribution:function(t){return t?(this._attributions[t]||(this._attributions[t]=0),this._attributions[t]++,this._update(),this):this},removeAttribution:function(t){return t?(this._attributions[t]&&(this._attributions[t]--,this._update()),this):this},_update:function(){if(this._map){var t=[];for(var i in this._attributions)this._attributions[i]&&t.push(i);var e=[];this.options.prefix&&e.push(this.options.prefix),t.length&&e.push(t.join(", ")),this._container.innerHTML=e.join(" | ")}}});Le.mergeOptions({attributionControl:!0}),Le.addInitHook(function(){this.options.attributionControl&&(new Ce).addTo(this)});Pe.Layers=Te,Pe.Zoom=ze,Pe.Scale=Me,Pe.Attribution=Ce,be.layers=function(t,i,e){return new Te(t,i,e)},be.zoom=function(t){return new ze(t)},be.scale=function(t){return new Me(t)},be.attribution=function(t){return new Ce(t)};var Ze=v.extend({initialize:function(t){this._map=t},enable:function(){return this._enabled?this:(this._enabled=!0,this.addHooks(),this)},disable:function(){return this._enabled?(this._enabled=!1,this.removeHooks(),this):this},enabled:function(){return!!this._enabled}});Ze.addTo=function(t,i){return t.addHandler(i,this),this};var Se,Ee={Events:hi},ke=Vi?"touchstart mousedown":"mousedown",Ae={mousedown:"mouseup",touchstart:"touchend",pointerdown:"touchend",MSPointerDown:"touchend"},Ie={mousedown:"mousemove",touchstart:"touchmove",pointerdown:"touchmove",MSPointerDown:"touchmove"},Be=ui.extend({options:{clickTolerance:3},initialize:function(t,i,e,n){l(this,n),this._element=t,this._dragStartTarget=i||t,this._preventOutline=e},enable:function(){this._enabled||(V(this._dragStartTarget,ke,this._onDown,this),this._enabled=!0)},disable:function(){this._enabled&&(Be._dragging===this&&this.finishDrag(),q(this._dragStartTarget,ke,this._onDown,this),this._enabled=!1,this._moved=!1)},_onDown:function(t){if(!t._simulated&&this._enabled&&(this._moved=!1,!dt(this._element,"leaflet-zoom-anim")&&!(Be._dragging||t.shiftKey||1!==t.which&&1!==t.button&&!t.touches||(Be._dragging=this,this._preventOutline&&zt(this._element),bt(),mi(),this._moving)))){this.fire("down");var i=t.touches?t.touches[0]:t;this._startPoint=new x(i.clientX,i.clientY),V(document,Ie[t.type],this._onMove,this),V(document,Ae[t.type],this._onUp,this)}},_onMove:function(t){if(!t._simulated&&this._enabled)if(t.touches&&t.touches.length>1)this._moved=!0;else{var i=t.touches&&1===t.touches.length?t.touches[0]:t,e=new x(i.clientX,i.clientY).subtract(this._startPoint);(e.x||e.y)&&(Math.abs(e.x)+Math.abs(e.y)1e-7;h++)i=s*Math.sin(a),i=Math.pow((1-i)/(1+i),s/2),a+=u=Math.PI/2-2*Math.atan(r*i)-a;return new M(a*e,t.x*e/n)}},je=(Object.freeze||Object)({LonLat:De,Mercator:Ne,SphericalMercator:di}),We=i({},_i,{code:"EPSG:3395",projection:Ne,transformation:function(){var t=.5/(Math.PI*Ne.R);return S(t,.5,-t,.5)}()}),He=i({},_i,{code:"EPSG:4326",projection:De,transformation:S(1/180,1,-1/180,.5)}),Fe=i({},ci,{projection:De,transformation:S(1,0,-1,0),scale:function(t){return Math.pow(2,t)},zoom:function(t){return Math.log(t)/Math.LN2},distance:function(t,i){var e=i.lng-t.lng,n=i.lat-t.lat;return Math.sqrt(e*e+n*n)},infinite:!0});ci.Earth=_i,ci.EPSG3395=We,ci.EPSG3857=vi,ci.EPSG900913=yi,ci.EPSG4326=He,ci.Simple=Fe;var Ue=ui.extend({options:{pane:"overlayPane",attribution:null,bubblingMouseEvents:!0},addTo:function(t){return t.addLayer(this),this},remove:function(){return this.removeFrom(this._map||this._mapToAdd)},removeFrom:function(t){return t&&t.removeLayer(this),this},getPane:function(t){return this._map.getPane(t?this.options[t]||t:this.options.pane)},addInteractiveTarget:function(t){return this._map._targets[n(t)]=this,this},removeInteractiveTarget:function(t){return delete this._map._targets[n(t)],this},getAttribution:function(){return this.options.attribution},_layerAdd:function(t){var i=t.target;if(i.hasLayer(this)){if(this._map=i,this._zoomAnimated=i._zoomAnimated,this.getEvents){var e=this.getEvents();i.on(e,this),this.once("remove",function(){i.off(e,this)},this)}this.onAdd(i),this.getAttribution&&i.attributionControl&&i.attributionControl.addAttribution(this.getAttribution()),this.fire("add"),i.fire("layeradd",{layer:this})}}});Le.include({addLayer:function(t){if(!t._layerAdd)throw new Error("The provided object is not a Layer.");var i=n(t);return this._layers[i]?this:(this._layers[i]=t,t._mapToAdd=this,t.beforeAdd&&t.beforeAdd(this),this.whenReady(t._layerAdd,t),this)},removeLayer:function(t){var i=n(t);return this._layers[i]?(this._loaded&&t.onRemove(this),t.getAttribution&&this.attributionControl&&this.attributionControl.removeAttribution(t.getAttribution()),delete this._layers[i],this._loaded&&(this.fire("layerremove",{layer:t}),t.fire("remove")),t._map=t._mapToAdd=null,this):this},hasLayer:function(t){return!!t&&n(t)in this._layers},eachLayer:function(t,i){for(var e in this._layers)t.call(i,this._layers[e]);return this},_addLayers:function(t){for(var i=0,e=(t=t?ei(t)?t:[t]:[]).length;ithis._layersMaxZoom&&this.setZoom(this._layersMaxZoom),void 0===this.options.minZoom&&this._layersMinZoom&&this.getZoom()i)return r=(n-i)/e,this._map.layerPointToLatLng([s.x-r*(s.x-o.x),s.y-r*(s.y-o.y)])},getBounds:function(){return this._bounds},addLatLng:function(t,i){return i=i||this._defaultShape(),t=C(t),i.push(t),this._bounds.extend(t),this.redraw()},_setLatLngs:function(t){this._bounds=new T,this._latlngs=this._convertLatLngs(t)},_defaultShape:function(){return Dt(this._latlngs)?this._latlngs:this._latlngs[0]},_convertLatLngs:function(t){for(var i=[],e=Dt(t),n=0,o=t.length;n=2&&i[0]instanceof M&&i[0].equals(i[e-1])&&i.pop(),i},_setLatLngs:function(t){tn.prototype._setLatLngs.call(this,t),Dt(this._latlngs)&&(this._latlngs=[this._latlngs])},_defaultShape:function(){return Dt(this._latlngs[0])?this._latlngs[0]:this._latlngs[0][0]},_clipPoints:function(){var t=this._renderer._bounds,i=this.options.weight,e=new x(i,i);if(t=new P(t.min.subtract(e),t.max.add(e)),this._parts=[],this._pxBounds&&this._pxBounds.intersects(t))if(this.options.noClip)this._parts=this._rings;else for(var n,o=0,s=this._rings.length;ot.y!=n.y>t.y&&t.x<(n.x-e.x)*(t.y-e.y)/(n.y-e.y)+e.x&&(u=!u);return u||tn.prototype._containsPoint.call(this,t,!0)}}),nn=qe.extend({initialize:function(t,i){l(this,i),this._layers={},t&&this.addData(t)},addData:function(t){var i,e,n,o=ei(t)?t:t.features;if(o){for(i=0,e=o.length;i0?o:[i.src]}else{ei(this._url)||(this._url=[this._url]),i.autoplay=!!this.options.autoplay,i.loop=!!this.options.loop;for(var a=0;ao?(i.height=o+"px",pt(t,"leaflet-popup-scrolled")):mt(t,"leaflet-popup-scrolled"),this._containerWidth=this._container.offsetWidth},_animateZoom:function(t){var i=this._map._latLngToNewLayerPoint(this._latlng,t.zoom,t.center),e=this._getAnchor();Lt(this._container,i.add(e))},_adjustPan:function(){if(!(!this.options.autoPan||this._map._panAnim&&this._map._panAnim._inProgress)){var t=this._map,i=parseInt(at(this._container,"marginBottom"),10)||0,e=this._container.offsetHeight+i,n=this._containerWidth,o=new x(this._containerLeft,-e-this._containerBottom);o._add(Pt(this._container));var s=t.layerPointToContainerPoint(o),r=w(this.options.autoPanPadding),a=w(this.options.autoPanPaddingTopLeft||r),h=w(this.options.autoPanPaddingBottomRight||r),u=t.getSize(),l=0,c=0;s.x+n+h.x>u.x&&(l=s.x+n-u.x+h.x),s.x-l-a.x<0&&(l=s.x-a.x),s.y+e+h.y>u.y&&(c=s.y+e-u.y+h.y),s.y-c-a.y<0&&(c=s.y-a.y),(l||c)&&t.fire("autopanstart").panBy([l,c])}},_onCloseButtonClick:function(t){this._close(),Q(t)},_getAnchor:function(){return w(this._source&&this._source._getPopupAnchor?this._source._getPopupAnchor():[0,0])}});Le.mergeOptions({closePopupOnClick:!0}),Le.include({openPopup:function(t,i,e){return t instanceof un||(t=new un(e).setContent(t)),i&&t.setLatLng(i),this.hasLayer(t)?this:(this._popup&&this._popup.options.autoClose&&this.closePopup(),this._popup=t,this.addLayer(t))},closePopup:function(t){return t&&t!==this._popup||(t=this._popup,this._popup=null),t&&this.removeLayer(t),this}}),Ue.include({bindPopup:function(t,i){return t instanceof un?(l(t,i),this._popup=t,t._source=this):(this._popup&&!i||(this._popup=new un(i,this)),this._popup.setContent(t)),this._popupHandlersAdded||(this.on({click:this._openPopup,keypress:this._onKeyPress,remove:this.closePopup,move:this._movePopup}),this._popupHandlersAdded=!0),this},unbindPopup:function(){return this._popup&&(this.off({click:this._openPopup,keypress:this._onKeyPress,remove:this.closePopup,move:this._movePopup}),this._popupHandlersAdded=!1,this._popup=null),this},openPopup:function(t,i){if(t instanceof Ue||(i=t,t=this),t instanceof qe)for(var e in this._layers){t=this._layers[e];break}return i||(i=t.getCenter?t.getCenter():t.getLatLng()),this._popup&&this._map&&(this._popup._source=t,this._popup.update(),this._map.openPopup(this._popup,i)),this},closePopup:function(){return this._popup&&this._popup._close(),this},togglePopup:function(t){return this._popup&&(this._popup._map?this.closePopup():this.openPopup(t)),this},isPopupOpen:function(){return!!this._popup&&this._popup.isOpen()},setPopupContent:function(t){return this._popup&&this._popup.setContent(t),this},getPopup:function(){return this._popup},_openPopup:function(t){var i=t.layer||t.target;this._popup&&this._map&&(Q(t),i instanceof Je?this.openPopup(t.layer||t.target,t.latlng):this._map.hasLayer(this._popup)&&this._popup._source===i?this.closePopup():this.openPopup(i,t.latlng))},_movePopup:function(t){this._popup.setLatLng(t.latlng)},_onKeyPress:function(t){13===t.originalEvent.keyCode&&this._openPopup(t)}});var ln=hn.extend({options:{pane:"tooltipPane",offset:[0,0],direction:"auto",permanent:!1,sticky:!1,interactive:!1,opacity:.9},onAdd:function(t){hn.prototype.onAdd.call(this,t),this.setOpacity(this.options.opacity),t.fire("tooltipopen",{tooltip:this}),this._source&&this._source.fire("tooltipopen",{tooltip:this},!0)},onRemove:function(t){hn.prototype.onRemove.call(this,t),t.fire("tooltipclose",{tooltip:this}),this._source&&this._source.fire("tooltipclose",{tooltip:this},!0)},getEvents:function(){var t=hn.prototype.getEvents.call(this);return Vi&&!this.options.permanent&&(t.preclick=this._close),t},_close:function(){this._map&&this._map.closeTooltip(this)},_initLayout:function(){var t="leaflet-tooltip "+(this.options.className||"")+" leaflet-zoom-"+(this._zoomAnimated?"animated":"hide");this._contentNode=this._container=ht("div",t)},_updateLayout:function(){},_adjustPan:function(){},_setPosition:function(t){var i=this._map,e=this._container,n=i.latLngToContainerPoint(i.getCenter()),o=i.layerPointToContainerPoint(t),s=this.options.direction,r=e.offsetWidth,a=e.offsetHeight,h=w(this.options.offset),u=this._getAnchor();"top"===s?t=t.add(w(-r/2+h.x,-a+h.y+u.y,!0)):"bottom"===s?t=t.subtract(w(r/2-h.x,-h.y,!0)):"center"===s?t=t.subtract(w(r/2+h.x,a/2-u.y+h.y,!0)):"right"===s||"auto"===s&&o.xthis.options.maxZoom||en&&this._retainParent(o,s,r,n))},_retainChildren:function(t,i,e,n){for(var o=2*t;o<2*t+2;o++)for(var s=2*i;s<2*i+2;s++){var r=new x(o,s);r.z=e+1;var a=this._tileCoordsToKey(r),h=this._tiles[a];h&&h.active?h.retain=!0:(h&&h.loaded&&(h.retain=!0),e+1this.options.maxZoom||void 0!==this.options.minZoom&&o1)this._setView(t,e);else{for(var c=o.min.y;c<=o.max.y;c++)for(var _=o.min.x;_<=o.max.x;_++){var d=new x(_,c);if(d.z=this._tileZoom,this._isValidTile(d)){var p=this._tiles[this._tileCoordsToKey(d)];p?p.current=!0:r.push(d)}}if(r.sort(function(t,i){return t.distanceTo(s)-i.distanceTo(s)}),0!==r.length){this._loading||(this._loading=!0,this.fire("loading"));var m=document.createDocumentFragment();for(_=0;_e.max.x)||!i.wrapLat&&(t.ye.max.y))return!1}if(!this.options.bounds)return!0;var n=this._tileCoordsToBounds(t);return z(this.options.bounds).overlaps(n)},_keyToBounds:function(t){return this._tileCoordsToBounds(this._keyToTileCoords(t))},_tileCoordsToNwSe:function(t){var i=this._map,e=this.getTileSize(),n=t.scaleBy(e),o=n.add(e);return[i.unproject(n,t.z),i.unproject(o,t.z)]},_tileCoordsToBounds:function(t){var i=this._tileCoordsToNwSe(t),e=new T(i[0],i[1]);return this.options.noWrap||(e=this._map.wrapLatLngBounds(e)),e},_tileCoordsToKey:function(t){return t.x+":"+t.y+":"+t.z},_keyToTileCoords:function(t){var i=t.split(":"),e=new x(+i[0],+i[1]);return e.z=+i[2],e},_removeTile:function(t){var i=this._tiles[t];i&&(Ci||i.el.setAttribute("src",ni),ut(i.el),delete this._tiles[t],this.fire("tileunload",{tile:i.el,coords:this._keyToTileCoords(t)}))},_initTile:function(t){pt(t,"leaflet-tile");var i=this.getTileSize();t.style.width=i.x+"px",t.style.height=i.y+"px",t.onselectstart=r,t.onmousemove=r,Li&&this.options.opacity<1&&vt(t,this.options.opacity),Ti&&!zi&&(t.style.WebkitBackfaceVisibility="hidden")},_addTile:function(t,i){var n=this._getTilePos(t),o=this._tileCoordsToKey(t),s=this.createTile(this._wrapCoords(t),e(this._tileReady,this,t));this._initTile(s),this.createTile.length<2&&f(e(this._tileReady,this,t,null,s)),Lt(s,n),this._tiles[o]={el:s,coords:t,current:!0},i.appendChild(s),this.fire("tileloadstart",{tile:s,coords:t})},_tileReady:function(t,i,n){if(this._map){i&&this.fire("tileerror",{error:i,tile:n,coords:t});var o=this._tileCoordsToKey(t);(n=this._tiles[o])&&(n.loaded=+new Date,this._map._fadeAnimated?(vt(n.el,0),g(this._fadeFrame),this._fadeFrame=f(this._updateOpacity,this)):(n.active=!0,this._pruneTiles()),i||(pt(n.el,"leaflet-tile-loaded"),this.fire("tileload",{tile:n.el,coords:t})),this._noTilesToLoad()&&(this._loading=!1,this.fire("load"),Li||!this._map._fadeAnimated?f(this._pruneTiles,this):setTimeout(e(this._pruneTiles,this),250)))}},_getTilePos:function(t){return t.scaleBy(this.getTileSize()).subtract(this._level.origin)},_wrapCoords:function(t){var i=new x(this._wrapX?s(t.x,this._wrapX):t.x,this._wrapY?s(t.y,this._wrapY):t.y);return i.z=t.z,i},_pxBoundsToTileRange:function(t){var i=this.getTileSize();return new P(t.min.unscaleBy(i).floor(),t.max.unscaleBy(i).ceil().subtract([1,1]))},_noTilesToLoad:function(){for(var t in this._tiles)if(!this._tiles[t].loaded)return!1;return!0}}),dn=_n.extend({options:{minZoom:0,maxZoom:18,subdomains:"abc",errorTileUrl:"",zoomOffset:0,tms:!1,zoomReverse:!1,detectRetina:!1,crossOrigin:!1},initialize:function(t,i){this._url=t,(i=l(this,i)).detectRetina&&Ki&&i.maxZoom>0&&(i.tileSize=Math.floor(i.tileSize/2),i.zoomReverse?(i.zoomOffset--,i.minZoom++):(i.zoomOffset++,i.maxZoom--),i.minZoom=Math.max(0,i.minZoom)),"string"==typeof i.subdomains&&(i.subdomains=i.subdomains.split("")),Ti||this.on("tileunload",this._onTileRemove)},setUrl:function(t,i){return this._url=t,i||this.redraw(),this},createTile:function(t,i){var n=document.createElement("img");return V(n,"load",e(this._tileOnLoad,this,i,n)),V(n,"error",e(this._tileOnError,this,i,n)),this.options.crossOrigin&&(n.crossOrigin=""),n.alt="",n.setAttribute("role","presentation"),n.src=this.getTileUrl(t),n},getTileUrl:function(t){var e={r:Ki?"@2x":"",s:this._getSubdomain(t),x:t.x,y:t.y,z:this._getZoomForUrl()};if(this._map&&!this._map.options.crs.infinite){var n=this._globalTileRange.max.y-t.y;this.options.tms&&(e.y=n),e["-y"]=n}return _(this._url,i(e,this.options))},_tileOnLoad:function(t,i){Li?setTimeout(e(t,this,null,i),0):t(null,i)},_tileOnError:function(t,i,e){var n=this.options.errorTileUrl;n&&i.getAttribute("src")!==n&&(i.src=n),t(e,i)},_onTileRemove:function(t){t.tile.onload=null},_getZoomForUrl:function(){var t=this._tileZoom,i=this.options.maxZoom,e=this.options.zoomReverse,n=this.options.zoomOffset;return e&&(t=i-t),t+n},_getSubdomain:function(t){var i=Math.abs(t.x+t.y)%this.options.subdomains.length;return this.options.subdomains[i]},_abortLoading:function(){var t,i;for(t in this._tiles)this._tiles[t].coords.z!==this._tileZoom&&((i=this._tiles[t].el).onload=r,i.onerror=r,i.complete||(i.src=ni,ut(i),delete this._tiles[t]))}}),pn=dn.extend({defaultWmsParams:{service:"WMS",request:"GetMap",layers:"",styles:"",format:"image/jpeg",transparent:!1,version:"1.1.1"},options:{crs:null,uppercase:!1},initialize:function(t,e){this._url=t;var n=i({},this.defaultWmsParams);for(var o in e)o in this.options||(n[o]=e[o]);var s=(e=l(this,e)).detectRetina&&Ki?2:1,r=this.getTileSize();n.width=r.x*s,n.height=r.y*s,this.wmsParams=n},onAdd:function(t){this._crs=this.options.crs||t.options.crs,this._wmsVersion=parseFloat(this.wmsParams.version);var i=this._wmsVersion>=1.3?"crs":"srs";this.wmsParams[i]=this._crs.code,dn.prototype.onAdd.call(this,t)},getTileUrl:function(t){var i=this._tileCoordsToNwSe(t),e=this._crs,n=b(e.project(i[0]),e.project(i[1])),o=n.min,s=n.max,r=(this._wmsVersion>=1.3&&this._crs===He?[o.y,o.x,s.y,s.x]:[o.x,o.y,s.x,s.y]).join(","),a=L.TileLayer.prototype.getTileUrl.call(this,t);return a+c(this.wmsParams,a,this.options.uppercase)+(this.options.uppercase?"&BBOX=":"&bbox=")+r},setParams:function(t,e){return i(this.wmsParams,t),e||this.redraw(),this}});dn.WMS=pn,Yt.wms=function(t,i){return new pn(t,i)};var mn=Ue.extend({options:{padding:.1,tolerance:0},initialize:function(t){l(this,t),n(this),this._layers=this._layers||{}},onAdd:function(){this._container||(this._initContainer(),this._zoomAnimated&&pt(this._container,"leaflet-zoom-animated")),this.getPane().appendChild(this._container),this._update(),this.on("update",this._updatePaths,this)},onRemove:function(){this.off("update",this._updatePaths,this),this._destroyContainer()},getEvents:function(){var t={viewreset:this._reset,zoom:this._onZoom,moveend:this._update,zoomend:this._onZoomEnd};return this._zoomAnimated&&(t.zoomanim=this._onAnimZoom),t},_onAnimZoom:function(t){this._updateTransform(t.center,t.zoom)},_onZoom:function(){this._updateTransform(this._map.getCenter(),this._map.getZoom())},_updateTransform:function(t,i){var e=this._map.getZoomScale(i,this._zoom),n=Pt(this._container),o=this._map.getSize().multiplyBy(.5+this.options.padding),s=this._map.project(this._center,i),r=this._map.project(t,i).subtract(s),a=o.multiplyBy(-e).add(n).add(o).subtract(r);Ni?wt(this._container,a,e):Lt(this._container,a)},_reset:function(){this._update(),this._updateTransform(this._center,this._zoom);for(var t in this._layers)this._layers[t]._reset()},_onZoomEnd:function(){for(var t in this._layers)this._layers[t]._project()},_updatePaths:function(){for(var t in this._layers)this._layers[t]._update()},_update:function(){var t=this.options.padding,i=this._map.getSize(),e=this._map.containerPointToLayerPoint(i.multiplyBy(-t)).round();this._bounds=new P(e,e.add(i.multiplyBy(1+2*t)).round()),this._center=this._map.getCenter(),this._zoom=this._map.getZoom()}}),fn=mn.extend({getEvents:function(){var t=mn.prototype.getEvents.call(this);return t.viewprereset=this._onViewPreReset,t},_onViewPreReset:function(){this._postponeUpdatePaths=!0},onAdd:function(){mn.prototype.onAdd.call(this),this._draw()},_initContainer:function(){var t=this._container=document.createElement("canvas");V(t,"mousemove",o(this._onMouseMove,32,this),this),V(t,"click dblclick mousedown mouseup contextmenu",this._onClick,this),V(t,"mouseout",this._handleMouseOut,this),this._ctx=t.getContext("2d")},_destroyContainer:function(){delete this._ctx,ut(this._container),q(this._container),delete this._container},_updatePaths:function(){if(!this._postponeUpdatePaths){this._redrawBounds=null;for(var t in this._layers)this._layers[t]._update();this._redraw()}},_update:function(){if(!this._map._animatingZoom||!this._bounds){this._drawnLayers={},mn.prototype._update.call(this);var t=this._bounds,i=this._container,e=t.getSize(),n=Ki?2:1;Lt(i,t.min),i.width=n*e.x,i.height=n*e.y,i.style.width=e.x+"px",i.style.height=e.y+"px",Ki&&this._ctx.scale(2,2),this._ctx.translate(-t.min.x,-t.min.y),this.fire("update")}},_reset:function(){mn.prototype._reset.call(this),this._postponeUpdatePaths&&(this._postponeUpdatePaths=!1,this._updatePaths())},_initPath:function(t){this._updateDashArray(t),this._layers[n(t)]=t;var i=t._order={layer:t,prev:this._drawLast,next:null};this._drawLast&&(this._drawLast.next=i),this._drawLast=i,this._drawFirst=this._drawFirst||this._drawLast},_addPath:function(t){this._requestRedraw(t)},_removePath:function(t){var i=t._order,e=i.next,n=i.prev;e?e.prev=n:this._drawLast=n,n?n.next=e:this._drawFirst=e,delete t._order,delete this._layers[L.stamp(t)],this._requestRedraw(t)},_updatePath:function(t){this._extendRedrawBounds(t),t._project(),t._update(),this._requestRedraw(t)},_updateStyle:function(t){this._updateDashArray(t),this._requestRedraw(t)},_updateDashArray:function(t){if(t.options.dashArray){var i,e=t.options.dashArray.split(","),n=[];for(i=0;i')}}catch(t){return function(t){return document.createElement("<"+t+' xmlns="urn:schemas-microsoft.com:vml" class="lvml">')}}}(),vn={_initContainer:function(){this._container=ht("div","leaflet-vml-container")},_update:function(){this._map._animatingZoom||(mn.prototype._update.call(this),this.fire("update"))},_initPath:function(t){var i=t._container=gn("shape");pt(i,"leaflet-vml-shape "+(this.options.className||"")),i.coordsize="1 1",t._path=gn("path"),i.appendChild(t._path),this._updateStyle(t),this._layers[n(t)]=t},_addPath:function(t){var i=t._container;this._container.appendChild(i),t.options.interactive&&t.addInteractiveTarget(i)},_removePath:function(t){var i=t._container;ut(i),t.removeInteractiveTarget(i),delete this._layers[n(t)]},_updateStyle:function(t){var i=t._stroke,e=t._fill,n=t.options,o=t._container;o.stroked=!!n.stroke,o.filled=!!n.fill,n.stroke?(i||(i=t._stroke=gn("stroke")),o.appendChild(i),i.weight=n.weight+"px",i.color=n.color,i.opacity=n.opacity,n.dashArray?i.dashStyle=ei(n.dashArray)?n.dashArray.join(" "):n.dashArray.replace(/( *, *)/g," "):i.dashStyle="",i.endcap=n.lineCap.replace("butt","flat"),i.joinstyle=n.lineJoin):i&&(o.removeChild(i),t._stroke=null),n.fill?(e||(e=t._fill=gn("fill")),o.appendChild(e),e.color=n.fillColor||n.color,e.opacity=n.fillOpacity):e&&(o.removeChild(e),t._fill=null)},_updateCircle:function(t){var i=t._point.round(),e=Math.round(t._radius),n=Math.round(t._radiusY||e);this._setPath(t,t._empty()?"M0 0":"AL "+i.x+","+i.y+" "+e+","+n+" 0,23592600")},_setPath:function(t,i){t._path.v=i},_bringToFront:function(t){ct(t._container)},_bringToBack:function(t){_t(t._container)}},yn=Ji?gn:E,xn=mn.extend({getEvents:function(){var t=mn.prototype.getEvents.call(this);return t.zoomstart=this._onZoomStart,t},_initContainer:function(){this._container=yn("svg"),this._container.setAttribute("pointer-events","none"),this._rootGroup=yn("g"),this._container.appendChild(this._rootGroup)},_destroyContainer:function(){ut(this._container),q(this._container),delete this._container,delete this._rootGroup,delete this._svgSize},_onZoomStart:function(){this._update()},_update:function(){if(!this._map._animatingZoom||!this._bounds){mn.prototype._update.call(this);var t=this._bounds,i=t.getSize(),e=this._container;this._svgSize&&this._svgSize.equals(i)||(this._svgSize=i,e.setAttribute("width",i.x),e.setAttribute("height",i.y)),Lt(e,t.min),e.setAttribute("viewBox",[t.min.x,t.min.y,i.x,i.y].join(" ")),this.fire("update")}},_initPath:function(t){var i=t._path=yn("path");t.options.className&&pt(i,t.options.className),t.options.interactive&&pt(i,"leaflet-interactive"),this._updateStyle(t),this._layers[n(t)]=t},_addPath:function(t){this._rootGroup||this._initContainer(),this._rootGroup.appendChild(t._path),t.addInteractiveTarget(t._path)},_removePath:function(t){ut(t._path),t.removeInteractiveTarget(t._path),delete this._layers[n(t)]},_updatePath:function(t){t._project(),t._update()},_updateStyle:function(t){var i=t._path,e=t.options;i&&(e.stroke?(i.setAttribute("stroke",e.color),i.setAttribute("stroke-opacity",e.opacity),i.setAttribute("stroke-width",e.weight),i.setAttribute("stroke-linecap",e.lineCap),i.setAttribute("stroke-linejoin",e.lineJoin),e.dashArray?i.setAttribute("stroke-dasharray",e.dashArray):i.removeAttribute("stroke-dasharray"),e.dashOffset?i.setAttribute("stroke-dashoffset",e.dashOffset):i.removeAttribute("stroke-dashoffset")):i.setAttribute("stroke","none"),e.fill?(i.setAttribute("fill",e.fillColor||e.color),i.setAttribute("fill-opacity",e.fillOpacity),i.setAttribute("fill-rule",e.fillRule||"evenodd")):i.setAttribute("fill","none"))},_updatePoly:function(t,i){this._setPath(t,k(t._parts,i))},_updateCircle:function(t){var i=t._point,e=Math.max(Math.round(t._radius),1),n="a"+e+","+(Math.max(Math.round(t._radiusY),1)||e)+" 0 1,0 ",o=t._empty()?"M0 0":"M"+(i.x-e)+","+i.y+n+2*e+",0 "+n+2*-e+",0 ";this._setPath(t,o)},_setPath:function(t,i){t._path.setAttribute("d",i)},_bringToFront:function(t){ct(t._path)},_bringToBack:function(t){_t(t._path)}});Ji&&xn.include(vn),Le.include({getRenderer:function(t){var i=t.options.renderer||this._getPaneRenderer(t.options.pane)||this.options.renderer||this._renderer;return i||(i=this._renderer=this.options.preferCanvas&&Xt()||Jt()),this.hasLayer(i)||this.addLayer(i),i},_getPaneRenderer:function(t){if("overlayPane"===t||void 0===t)return!1;var i=this._paneRenderers[t];return void 0===i&&(i=xn&&Jt({pane:t})||fn&&Xt({pane:t}),this._paneRenderers[t]=i),i}});var wn=en.extend({initialize:function(t,i){en.prototype.initialize.call(this,this._boundsToLatLngs(t),i)},setBounds:function(t){return this.setLatLngs(this._boundsToLatLngs(t))},_boundsToLatLngs:function(t){return t=z(t),[t.getSouthWest(),t.getNorthWest(),t.getNorthEast(),t.getSouthEast()]}});xn.create=yn,xn.pointsToPath=k,nn.geometryToLayer=Wt,nn.coordsToLatLng=Ht,nn.coordsToLatLngs=Ft,nn.latLngToCoords=Ut,nn.latLngsToCoords=Vt,nn.getFeature=qt,nn.asFeature=Gt,Le.mergeOptions({boxZoom:!0});var Ln=Ze.extend({initialize:function(t){this._map=t,this._container=t._container,this._pane=t._panes.overlayPane,this._resetStateTimeout=0,t.on("unload",this._destroy,this)},addHooks:function(){V(this._container,"mousedown",this._onMouseDown,this)},removeHooks:function(){q(this._container,"mousedown",this._onMouseDown,this)},moved:function(){return this._moved},_destroy:function(){ut(this._pane),delete this._pane},_resetState:function(){this._resetStateTimeout=0,this._moved=!1},_clearDeferredResetState:function(){0!==this._resetStateTimeout&&(clearTimeout(this._resetStateTimeout),this._resetStateTimeout=0)},_onMouseDown:function(t){if(!t.shiftKey||1!==t.which&&1!==t.button)return!1;this._clearDeferredResetState(),this._resetState(),mi(),bt(),this._startPoint=this._map.mouseEventToContainerPoint(t),V(document,{contextmenu:Q,mousemove:this._onMouseMove,mouseup:this._onMouseUp,keydown:this._onKeyDown},this)},_onMouseMove:function(t){this._moved||(this._moved=!0,this._box=ht("div","leaflet-zoom-box",this._container),pt(this._container,"leaflet-crosshair"),this._map.fire("boxzoomstart")),this._point=this._map.mouseEventToContainerPoint(t);var i=new P(this._point,this._startPoint),e=i.getSize();Lt(this._box,i.min),this._box.style.width=e.x+"px",this._box.style.height=e.y+"px"},_finish:function(){this._moved&&(ut(this._box),mt(this._container,"leaflet-crosshair")),fi(),Tt(),q(document,{contextmenu:Q,mousemove:this._onMouseMove,mouseup:this._onMouseUp,keydown:this._onKeyDown},this)},_onMouseUp:function(t){if((1===t.which||1===t.button)&&(this._finish(),this._moved)){this._clearDeferredResetState(),this._resetStateTimeout=setTimeout(e(this._resetState,this),0);var i=new T(this._map.containerPointToLatLng(this._startPoint),this._map.containerPointToLatLng(this._point));this._map.fitBounds(i).fire("boxzoomend",{boxZoomBounds:i})}},_onKeyDown:function(t){27===t.keyCode&&this._finish()}});Le.addInitHook("addHandler","boxZoom",Ln),Le.mergeOptions({doubleClickZoom:!0});var Pn=Ze.extend({addHooks:function(){this._map.on("dblclick",this._onDoubleClick,this)},removeHooks:function(){this._map.off("dblclick",this._onDoubleClick,this)},_onDoubleClick:function(t){var i=this._map,e=i.getZoom(),n=i.options.zoomDelta,o=t.originalEvent.shiftKey?e-n:e+n;"center"===i.options.doubleClickZoom?i.setZoom(o):i.setZoomAround(t.containerPoint,o)}});Le.addInitHook("addHandler","doubleClickZoom",Pn),Le.mergeOptions({dragging:!0,inertia:!zi,inertiaDeceleration:3400,inertiaMaxSpeed:1/0,easeLinearity:.2,worldCopyJump:!1,maxBoundsViscosity:0});var bn=Ze.extend({addHooks:function(){if(!this._draggable){var t=this._map;this._draggable=new Be(t._mapPane,t._container),this._draggable.on({dragstart:this._onDragStart,drag:this._onDrag,dragend:this._onDragEnd},this),this._draggable.on("predrag",this._onPreDragLimit,this),t.options.worldCopyJump&&(this._draggable.on("predrag",this._onPreDragWrap,this),t.on("zoomend",this._onZoomEnd,this),t.whenReady(this._onZoomEnd,this))}pt(this._map._container,"leaflet-grab leaflet-touch-drag"),this._draggable.enable(),this._positions=[],this._times=[]},removeHooks:function(){mt(this._map._container,"leaflet-grab"),mt(this._map._container,"leaflet-touch-drag"),this._draggable.disable()},moved:function(){return this._draggable&&this._draggable._moved},moving:function(){return this._draggable&&this._draggable._moving},_onDragStart:function(){var t=this._map;if(t._stop(),this._map.options.maxBounds&&this._map.options.maxBoundsViscosity){var i=z(this._map.options.maxBounds);this._offsetLimit=b(this._map.latLngToContainerPoint(i.getNorthWest()).multiplyBy(-1),this._map.latLngToContainerPoint(i.getSouthEast()).multiplyBy(-1).add(this._map.getSize())),this._viscosity=Math.min(1,Math.max(0,this._map.options.maxBoundsViscosity))}else this._offsetLimit=null;t.fire("movestart").fire("dragstart"),t.options.inertia&&(this._positions=[],this._times=[])},_onDrag:function(t){if(this._map.options.inertia){var i=this._lastTime=+new Date,e=this._lastPos=this._draggable._absPos||this._draggable._newPos;this._positions.push(e),this._times.push(i),this._prunePositions(i)}this._map.fire("move",t).fire("drag",t)},_prunePositions:function(t){for(;this._positions.length>1&&t-this._times[0]>50;)this._positions.shift(),this._times.shift()},_onZoomEnd:function(){var t=this._map.getSize().divideBy(2),i=this._map.latLngToLayerPoint([0,0]);this._initialWorldOffset=i.subtract(t).x,this._worldWidth=this._map.getPixelWorldBounds().getSize().x},_viscousLimit:function(t,i){return t-(t-i)*this._viscosity},_onPreDragLimit:function(){if(this._viscosity&&this._offsetLimit){var t=this._draggable._newPos.subtract(this._draggable._startPos),i=this._offsetLimit;t.xi.max.x&&(t.x=this._viscousLimit(t.x,i.max.x)),t.y>i.max.y&&(t.y=this._viscousLimit(t.y,i.max.y)),this._draggable._newPos=this._draggable._startPos.add(t)}},_onPreDragWrap:function(){var t=this._worldWidth,i=Math.round(t/2),e=this._initialWorldOffset,n=this._draggable._newPos.x,o=(n-i+e)%t+i-e,s=(n+i+e)%t-i-e,r=Math.abs(o+e)0?s:-s))-i;this._delta=0,this._startTime=null,r&&("center"===t.options.scrollWheelZoom?t.setZoom(i+r):t.setZoomAround(this._lastMousePos,i+r))}});Le.addInitHook("addHandler","scrollWheelZoom",zn),Le.mergeOptions({tap:!0,tapTolerance:15});var Mn=Ze.extend({addHooks:function(){V(this._map._container,"touchstart",this._onDown,this)},removeHooks:function(){q(this._map._container,"touchstart",this._onDown,this)},_onDown:function(t){if(t.touches){if($(t),this._fireClick=!0,t.touches.length>1)return this._fireClick=!1,void clearTimeout(this._holdTimeout);var i=t.touches[0],n=i.target;this._startPos=this._newPos=new x(i.clientX,i.clientY),n.tagName&&"a"===n.tagName.toLowerCase()&&pt(n,"leaflet-active"),this._holdTimeout=setTimeout(e(function(){this._isTapValid()&&(this._fireClick=!1,this._onUp(),this._simulateEvent("contextmenu",i))},this),1e3),this._simulateEvent("mousedown",i),V(document,{touchmove:this._onMove,touchend:this._onUp},this)}},_onUp:function(t){if(clearTimeout(this._holdTimeout),q(document,{touchmove:this._onMove,touchend:this._onUp},this),this._fireClick&&t&&t.changedTouches){var i=t.changedTouches[0],e=i.target;e&&e.tagName&&"a"===e.tagName.toLowerCase()&&mt(e,"leaflet-active"),this._simulateEvent("mouseup",i),this._isTapValid()&&this._simulateEvent("click",i)}},_isTapValid:function(){return this._newPos.distanceTo(this._startPos)<=this._map.options.tapTolerance},_onMove:function(t){var i=t.touches[0];this._newPos=new x(i.clientX,i.clientY),this._simulateEvent("mousemove",i)},_simulateEvent:function(t,i){var e=document.createEvent("MouseEvents");e._simulated=!0,i.target._simulatedClick=!0,e.initMouseEvent(t,!0,!0,window,1,i.screenX,i.screenY,i.clientX,i.clientY,!1,!1,!1,!1,0,null),i.target.dispatchEvent(e)}});Vi&&!Ui&&Le.addInitHook("addHandler","tap",Mn),Le.mergeOptions({touchZoom:Vi&&!zi,bounceAtZoomLimits:!0});var Cn=Ze.extend({addHooks:function(){pt(this._map._container,"leaflet-touch-zoom"),V(this._map._container,"touchstart",this._onTouchStart,this)},removeHooks:function(){mt(this._map._container,"leaflet-touch-zoom"),q(this._map._container,"touchstart",this._onTouchStart,this)},_onTouchStart:function(t){var i=this._map;if(t.touches&&2===t.touches.length&&!i._animatingZoom&&!this._zooming){var e=i.mouseEventToContainerPoint(t.touches[0]),n=i.mouseEventToContainerPoint(t.touches[1]);this._centerPoint=i.getSize()._divideBy(2),this._startLatLng=i.containerPointToLatLng(this._centerPoint),"center"!==i.options.touchZoom&&(this._pinchStartLatLng=i.containerPointToLatLng(e.add(n)._divideBy(2))),this._startDist=e.distanceTo(n),this._startZoom=i.getZoom(),this._moved=!1,this._zooming=!0,i._stop(),V(document,"touchmove",this._onTouchMove,this),V(document,"touchend",this._onTouchEnd,this),$(t)}},_onTouchMove:function(t){if(t.touches&&2===t.touches.length&&this._zooming){var i=this._map,n=i.mouseEventToContainerPoint(t.touches[0]),o=i.mouseEventToContainerPoint(t.touches[1]),s=n.distanceTo(o)/this._startDist;if(this._zoom=i.getScaleZoom(s,this._startZoom),!i.options.bounceAtZoomLimits&&(this._zoomi.getMaxZoom()&&s>1)&&(this._zoom=i._limitZoom(this._zoom)),"center"===i.options.touchZoom){if(this._center=this._startLatLng,1===s)return}else{var r=n._add(o)._divideBy(2)._subtract(this._centerPoint);if(1===s&&0===r.x&&0===r.y)return;this._center=i.unproject(i.project(this._pinchStartLatLng,this._zoom).subtract(r),this._zoom)}this._moved||(i._moveStart(!0,!1),this._moved=!0),g(this._animRequest);var a=e(i._move,i,this._center,this._zoom,{pinch:!0,round:!1});this._animRequest=f(a,this,!0),$(t)}},_onTouchEnd:function(){this._moved&&this._zooming?(this._zooming=!1,g(this._animRequest),q(document,"touchmove",this._onTouchMove),q(document,"touchend",this._onTouchEnd),this._map.options.zoomAnimation?this._map._animateZoom(this._center,this._map._limitZoom(this._zoom),!0,this._map.options.zoomSnap):this._map._resetView(this._center,this._map._limitZoom(this._zoom))):this._zooming=!1}});Le.addInitHook("addHandler","touchZoom",Cn),Le.BoxZoom=Ln,Le.DoubleClickZoom=Pn,Le.Drag=bn,Le.Keyboard=Tn,Le.ScrollWheelZoom=zn,Le.Tap=Mn,Le.TouchZoom=Cn;var Zn=window.L;window.L=t,Object.freeze=$t,t.version="1.3.1+HEAD.ba6f97f",t.noConflict=function(){return window.L=Zn,this},t.Control=Pe,t.control=be,t.Browser=$i,t.Evented=ui,t.Mixin=Ee,t.Util=ai,t.Class=v,t.Handler=Ze,t.extend=i,t.bind=e,t.stamp=n,t.setOptions=l,t.DomEvent=de,t.DomUtil=xe,t.PosAnimation=we,t.Draggable=Be,t.LineUtil=Oe,t.PolyUtil=Re,t.Point=x,t.point=w,t.Bounds=P,t.bounds=b,t.Transformation=Z,t.transformation=S,t.Projection=je,t.LatLng=M,t.latLng=C,t.LatLngBounds=T,t.latLngBounds=z,t.CRS=ci,t.GeoJSON=nn,t.geoJSON=Kt,t.geoJson=sn,t.Layer=Ue,t.LayerGroup=Ve,t.layerGroup=function(t,i){return new Ve(t,i)},t.FeatureGroup=qe,t.featureGroup=function(t){return new qe(t)},t.ImageOverlay=rn,t.imageOverlay=function(t,i,e){return new rn(t,i,e)},t.VideoOverlay=an,t.videoOverlay=function(t,i,e){return new an(t,i,e)},t.DivOverlay=hn,t.Popup=un,t.popup=function(t,i){return new un(t,i)},t.Tooltip=ln,t.tooltip=function(t,i){return new ln(t,i)},t.Icon=Ge,t.icon=function(t){return new Ge(t)},t.DivIcon=cn,t.divIcon=function(t){return new cn(t)},t.Marker=Xe,t.marker=function(t,i){return new Xe(t,i)},t.TileLayer=dn,t.tileLayer=Yt,t.GridLayer=_n,t.gridLayer=function(t){return new _n(t)},t.SVG=xn,t.svg=Jt,t.Renderer=mn,t.Canvas=fn,t.canvas=Xt,t.Path=Je,t.CircleMarker=$e,t.circleMarker=function(t,i){return new $e(t,i)},t.Circle=Qe,t.circle=function(t,i,e){return new Qe(t,i,e)},t.Polyline=tn,t.polyline=function(t,i){return new tn(t,i)},t.Polygon=en,t.polygon=function(t,i){return new en(t,i)},t.Rectangle=wn,t.rectangle=function(t,i){return new wn(t,i)},t.Map=Le,t.map=function(t,i){return new Le(t,i)}}); \ No newline at end of file diff --git a/rmarkdown-libs/proj4/proj4.min.js b/rmarkdown-libs/proj4/proj4.min.js new file mode 100644 index 0000000..99865a5 --- /dev/null +++ b/rmarkdown-libs/proj4/proj4.min.js @@ -0,0 +1 @@ +!function(t,s){"object"==typeof exports&&"undefined"!=typeof module?module.exports=s():"function"==typeof define&&define.amd?define(s):t.proj4=s()}(this,function(){"use strict";function k(t,s){if(t[s])return t[s];for(var i,a=Object.keys(t),h=s.toLowerCase().replace(H,""),e=-1;++eMath.PI&&(r-=2*Math.PI),h=Math.sin(o),n=Math.cos(o),e=h*h,{x:((a=i/Math.sqrt(1-s*e))+l)*n*Math.cos(r),y:(a+l)*n*Math.sin(r),z:(a*(1-s)+l)*h}}function c(t,s,i,a){var h,e,n,r,o,l,M,c,u,f,m,p,d,y=t.x,_=t.y,x=t.z?t.z:0,g=Math.sqrt(y*y+_*_),b=Math.sqrt(y*y+_*_+x*x);if(g/i<1e-12){if(p=0,b/i<1e-12)return d=-a,{x:t.x,y:t.y,z:t.z}}else p=Math.atan2(_,y);for(h=x/b,l=(e=g/b)*(1-s)*(n=1/Math.sqrt(1-s*(2-s)*e*e)),M=h*n,m=0;m++,r=s*(o=i/Math.sqrt(1-s*M*M))/(o+(d=g*l+x*M-o*(1-s*M*M))),f=(u=h*(n=1/Math.sqrt(1-r*(2-r)*e*e)))*l-(c=e*(1-r)*n)*M,l=c,M=u,1e-24=this.text.length)return;t=this.text[this.place++]}switch(this.state){case K:return this.neutral(t);case 2:return this.keyword(t);case 4:return this.quoted(t);case 5:return this.afterquote(t);case 3:return this.number(t);case-1:return}},e.prototype.afterquote=function(t){if('"'===t)return this.word+='"',void(this.state=4);if($.test(t))return this.word=this.word.trim(),void this.afterItem(t);throw new Error("havn't handled \""+t+'" in afterquote yet, index '+this.place)},e.prototype.afterItem=function(t){return","===t?(null!==this.word&&this.currentObject.push(this.word),this.word=null,void(this.state=K)):"]"===t?(this.level--,null!==this.word&&(this.currentObject.push(this.word),this.word=null),this.state=K,this.currentObject=this.stack.pop(),void(this.currentObject||(this.state=-1))):void 0},e.prototype.number=function(t){if(!tt.test(t)){if($.test(t))return this.word=parseFloat(this.word),void this.afterItem(t);throw new Error("havn't handled \""+t+'" in number yet, index '+this.place)}this.word+=t},e.prototype.quoted=function(t){'"'!==t?this.word+=t:this.state=5},e.prototype.keyword=function(t){if(Y.test(t))this.word+=t;else{if("["===t){var s=[];return s.push(this.word),this.level++,null===this.root?this.root=s:this.currentObject.push(s),this.stack.push(this.currentObject),this.currentObject=s,void(this.state=K)}if(!$.test(t))throw new Error("havn't handled \""+t+'" in keyword yet, index '+this.place);this.afterItem(t)}},e.prototype.neutral=function(t){if(Z.test(t))return this.word=t,void(this.state=2);if('"'===t)return this.word="",void(this.state=4);if(tt.test(t))return this.word=t,void(this.state=3);if(!$.test(t))throw new Error("havn't handled \""+t+'" in neutral yet, index '+this.place);this.afterItem(t)},e.prototype.output=function(){for(;this.placeD?Math.tan(i):0,c=Math.pow(M,2),u=Math.pow(c,2),f=1-this.es*Math.pow(h,2);n/=Math.sqrt(f);var m=Gt(i,h,e,this.en),p=this.a*(this.k0*n*(1+r/6*(1-c+o+r/20*(5-18*c+u+14*o-58*c*o+r/42*(61+179*u-u*c-479*c)))))+this.x0,d=this.a*(this.k0*(m-this.ml0+h*a*n/2*(1+r/12*(5-c+9*o+4*l+r/30*(61+u-58*c+270*o-330*c*o+r/56*(1385+543*u-u*c-3111*c))))))+this.y0}else{var y=e*Math.sin(a);if(Math.abs(Math.abs(y)-1)D?Math.tan(s):0,e=this.ep2*Math.pow(a,2),n=Math.pow(e,2),r=Math.pow(h,2),o=Math.pow(r,2),l=1-this.es*Math.pow(i,2),M=_*Math.sqrt(l)/this.k0,u=s-(l*=h)*(c=Math.pow(M,2))/(1-this.es)*.5*(1-c/12*(5+3*r-9*e*r+e-4*n-c/30*(61+90*r-252*e*r+45*o+46*e-c/56*(1385+3633*r+4095*o+1574*o*r)))),nt(this.long0+M*(1-c/6*(1+2*r+e-c/20*(5+28*r+24*o+8*e*r+6*e-c/42*(61+662*r+1320*o+720*o*r))))/a)):(u=z*et(x),0)):(p=.5*((m=Math.exp(_/this.k0))-1/m),d=this.lat0+x/this.k0,y=Math.cos(d),l=Math.sqrt((1-Math.pow(y,2))/(1+Math.pow(p,2))),u=Math.asin(l),x<0&&(u=-u),0==p&&0===y?0:nt(Math.atan2(p,y)+this.long0)),t.x=f,t.y=u,t},names:["Transverse_Mercator","Transverse Mercator","tmerc"]},is={init:function(){if(void 0===this.es||this.es<=0)throw new Error("incorrect elliptical usage");this.x0=void 0!==this.x0?this.x0:0,this.y0=void 0!==this.y0?this.y0:0,this.long0=void 0!==this.long0?this.long0:0,this.lat0=void 0!==this.lat0?this.lat0:0,this.cgb=[],this.cbg=[],this.utg=[],this.gtu=[];var t=this.es/(1+Math.sqrt(1-this.es)),s=t/(2-t),i=s;this.cgb[0]=s*(2+s*(-2/3+s*(s*(116/45+s*(26/45+-2854/675*s))-2))),this.cbg[0]=s*(s*(2/3+s*(4/3+s*(-82/45+s*(32/45+4642/4725*s))))-2),i*=s,this.cgb[1]=i*(7/3+s*(s*(-227/45+s*(2704/315+2323/945*s))-1.6)),this.cbg[1]=i*(5/3+s*(-16/15+s*(-13/9+s*(904/315+-1522/945*s)))),i*=s,this.cgb[2]=i*(56/15+s*(-136/35+s*(-1262/105+73814/2835*s))),this.cbg[2]=i*(-26/15+s*(34/21+s*(1.6+-12686/2835*s))),i*=s,this.cgb[3]=i*(4279/630+s*(-332/35+-399572/14175*s)),this.cbg[3]=i*(1237/630+s*(-24832/14175*s-2.4)),i*=s,this.cgb[4]=i*(4174/315+-144838/6237*s),this.cbg[4]=i*(-734/315+109598/31185*s),i*=s,this.cgb[5]=i*(601676/22275),this.cbg[5]=i*(444337/155925),i=Math.pow(s,2),this.Qn=this.k0/(1+s)*(1+i*(.25+i*(1/64+i/256))),this.utg[0]=s*(s*(2/3+s*(-37/96+s*(1/360+s*(81/512+-96199/604800*s))))-.5),this.gtu[0]=s*(.5+s*(-2/3+s*(5/16+s*(41/180+s*(-127/288+7891/37800*s))))),this.utg[1]=i*(-1/48+s*(-1/15+s*(437/1440+s*(-46/105+1118711/3870720*s)))),this.gtu[1]=i*(13/48+s*(s*(557/1440+s*(281/630+-1983433/1935360*s))-.6)),i*=s,this.utg[2]=i*(-17/480+s*(37/840+s*(209/4480+-5569/90720*s))),this.gtu[2]=i*(61/240+s*(-103/140+s*(15061/26880+167603/181440*s))),i*=s,this.utg[3]=i*(-4397/161280+s*(11/504+830251/7257600*s)),this.gtu[3]=i*(49561/161280+s*(-179/168+6601661/7257600*s)),i*=s,this.utg[4]=i*(-4583/161280+108847/3991680*s),this.gtu[4]=i*(34729/80640+-3418889/1995840*s),i*=s,this.utg[5]=-.03233083094085698*i,this.gtu[5]=.6650675310896665*i;var a=Tt(this.cbg,this.lat0);this.Zb=-this.Qn*(a+function(t,s){for(var i,a=2*Math.cos(s),h=t.length-1,e=t[h],n=0;0<=--h;)i=a*e-n+t[h],n=e,e=i;return Math.sin(s)*i}(this.gtu,2*a))},forward:function(t){var s=nt(t.x-this.long0),i=t.y,i=Tt(this.cbg,i),a=Math.sin(i),h=Math.cos(i),e=Math.sin(s),n=Math.cos(s);i=Math.atan2(a,n*h),s=Math.atan2(e*h,Rt(a,h*n)),s=Lt(Math.tan(s));var r,o,l=Dt(this.gtu,2*i,2*s);return i+=l[0],s+=l[1],o=Math.abs(s)<=2.623395162778?(r=this.a*(this.Qn*s)+this.x0,this.a*(this.Qn*i+this.Zb)+this.y0):r=1/0,t.x=r,t.y=o,t},inverse:function(t){var s,i,a,h,e,n,r,o=(t.x-this.x0)*(1/this.a),l=(t.y-this.y0)*(1/this.a);return l=(l-this.Zb)/this.Qn,o/=this.Qn,r=Math.abs(o)<=2.623395162778?(l+=(s=Dt(this.utg,2*l,2*o))[0],o+=s[1],o=Math.atan(zt(o)),i=Math.sin(l),a=Math.cos(l),h=Math.sin(o),e=Math.cos(o),l=Math.atan2(i*e,Rt(h,e*a)),o=Math.atan2(h,e*a),n=nt(o+this.long0),Tt(this.cgb,l)):n=1/0,t.x=n,t.y=r,t},names:["Extended_Transverse_Mercator","Extended Transverse Mercator","etmerc"]},as={init:function(){var t=function(t,s){if(void 0===t){if((t=Math.floor(30*(nt(s)+Math.PI)/Math.PI)+1)<0)return 0;if(60D?this.ns=Math.log(a/r)/Math.log(h/o):this.ns=s,isNaN(this.ns)&&(this.ns=s),this.f0=a/(this.ns*Math.pow(h,this.ns)),this.rh=this.a*this.f0*Math.pow(l,this.ns),this.title||(this.title="Lambert Conformal Conic"))},forward:function(t){var s=t.x,i=t.y;Math.abs(2*Math.abs(i)-Math.PI)<=D&&(i=et(i)*(z-2*D));var a,h,e=Math.abs(Math.abs(i)-z);if(DD?this.ns0=(this.ms1*this.ms1-this.ms2*this.ms2)/(this.qs2-this.qs1):this.ns0=this.con,this.c=this.ms1*this.ms1+this.ns0*this.qs1,this.rh=this.a*Math.sqrt(this.c-this.ns0*this.qs0)/this.ns0)},forward:function(t){var s=t.x,i=t.y;this.sin_phi=Math.sin(i),this.cos_phi=Math.cos(i);var a=Vt(this.e3,this.sin_phi,this.cos_phi),h=this.a*Math.sqrt(this.c-this.ns0*a)/this.ns0,e=this.ns0*nt(s-this.long0),n=h*Math.sin(e)+this.x0,r=this.rh-h*Math.cos(e)+this.y0;return t.x=n,t.y=r,t},inverse:function(t){var s,i,a,h,e,n;return t.x-=this.x0,t.y=this.rh-t.y+this.y0,a=0<=this.ns0?(s=Math.sqrt(t.x*t.x+t.y*t.y),1):(s=-Math.sqrt(t.x*t.x+t.y*t.y),-1),(h=0)!==s&&(h=Math.atan2(a*t.x,a*t.y)),a=s*this.ns0/this.a,n=this.sphere?Math.asin((this.c-a*a)/(2*this.ns0)):(i=(this.c-a*a)/this.ns0,this.phi1z(this.e3,i)),e=nt(h/this.ns0+this.long0),t.x=e,t.y=n,t},names:["Albers_Conic_Equal_Area","Albers","aea"],phi1z:function(t,s){var i,a,h,e,n=Zt(.5*s);if(tMath.PI&&(a=Math.PI),i=(2*s+Math.sin(2*s))/Math.PI,12*z*this.a)return;return i=s/this.a,a=Math.sin(i),h=Math.cos(i),e=this.long0,Math.abs(s)<=D?n=this.lat0:(n=Zt(h*this.sin_p12+t.y*a*this.cos_p12/s),r=Math.abs(this.lat0)-z,e=nt(Math.abs(r)<=D?0<=this.lat0?this.long0+Math.atan2(t.x,-t.y):this.long0-Math.atan2(-t.x,t.y):this.long0+Math.atan2(t.x*a,s*this.cos_p12*h-t.y*this.sin_p12*a))),t.x=e,t.y=n,t}return o=Ft(this.es),l=Qt(this.es),M=Wt(this.es),c=Xt(this.es),Math.abs(this.sin_p12-1)<=D?(u=this.a*Ut(o,l,M,c,z),s=Math.sqrt(t.x*t.x+t.y*t.y),n=Kt((u-s)/this.a,o,l,M,c),e=nt(this.long0+Math.atan2(t.x,-1*t.y))):Math.abs(this.sin_p12+1)<=D?(u=this.a*Ut(o,l,M,c,z),s=Math.sqrt(t.x*t.x+t.y*t.y),n=Kt((s-u)/this.a,o,l,M,c),e=nt(this.long0+Math.atan2(t.x,t.y))):(s=Math.sqrt(t.x*t.x+t.y*t.y),p=Math.atan2(t.x,t.y),f=Ht(this.a,this.e,this.sin_p12),d=Math.cos(p),_=-(y=this.e*this.cos_p12*d)*y/(1-this.es),x=3*this.es*(1-_)*this.sin_p12*this.cos_p12*d/(1-this.es),v=1-_*(b=(g=s/f)-_*(1+_)*Math.pow(g,3)/6-x*(1+3*_)*Math.pow(g,4)/24)*b/2-g*b*b*b/6,m=Math.asin(this.sin_p12*Math.cos(b)+this.cos_p12*Math.sin(b)*d),e=nt(this.long0+Math.asin(Math.sin(p)*Math.sin(b)/Math.cos(m))),w=Math.sin(m),n=Math.atan2((w-this.es*v*this.sin_p12)*Math.tan(m),w*(1-this.es))),t.x=e,t.y=n,t},names:["Azimuthal_Equidistant","aeqd"]},Ps={init:function(){this.sin_p14=Math.sin(this.lat0),this.cos_p14=Math.cos(this.lat0)},forward:function(t){var s,i,a,h=t.x,e=t.y,n=nt(h-this.long0),r=Math.sin(e),o=Math.cos(e),l=Math.cos(n);return(0<(s=this.sin_p14*r+this.cos_p14*o*l)||Math.abs(s)<=D)&&(i=this.a*o*Math.sin(n),a=this.y0+this.a*(this.cos_p14*r-this.sin_p14*o*l)),t.x=i,t.y=a,t},inverse:function(t){var s,i,a,h,e,n,r;return t.x-=this.x0,t.y-=this.y0,s=Math.sqrt(t.x*t.x+t.y*t.y),i=Zt(s/this.a),a=Math.sin(i),h=Math.cos(i),n=this.long0,Math.abs(s)<=D?r=this.lat0:(r=Zt(h*this.sin_p14+t.y*a*this.cos_p14/s),e=Math.abs(this.lat0)-z,n=Math.abs(e)<=D?nt(0<=this.lat0?this.long0+Math.atan2(t.x,-t.y):this.long0-Math.atan2(-t.x,t.y)):nt(this.long0+Math.atan2(t.x*a,s*this.cos_p14*h-t.y*this.sin_p14*a))),t.x=n,t.y=r,t},names:["ortho"]},Ss=1,Ns=2,ks=3,Es=4,qs=5,Is=6,Os=1,As=2,Gs=3,js=4,zs={init:function(){this.x0=this.x0||0,this.y0=this.y0||0,this.lat0=this.lat0||0,this.long0=this.long0||0,this.lat_ts=this.lat_ts||0,this.title=this.title||"Quadrilateralized Spherical Cube",this.lat0>=z-U/2?this.face=qs:this.lat0<=-(z-U/2)?this.face=Is:Math.abs(this.long0)<=U?this.face=Ss:Math.abs(this.long0)<=z+U?this.face=0=Math.abs(t.y)?y.value=Os:0<=t.y&&t.y>=Math.abs(t.x)?(y.value=As,s-=z):t.x<0&&-t.x>=Math.abs(t.y)?(y.value=Gs,s=s<0?s+Q:s-Q):(y.value=js,s+=z),c=Q/12*Math.tan(s),e=Math.sin(c)/(Math.cos(c)-1/Math.sqrt(2)),n=Math.atan(e),(r=1-(a=Math.cos(s))*a*(h=Math.tan(i))*h*(1-Math.cos(Math.atan(1/Math.cos(n)))))<-1?r=-1:1a.y)--s;else{if(!(Ls[s+1][0]<=a.y))break;++s}var h=Ls[s],i=function(t,s,i,a){for(var h=s;a;--a){var e=t(h);if(h-=e,Math.abs(e) - https://liuyanguu.github.io/categories/ - 2021-01-06T00:00:00+00:00 - - - - https://liuyanguu.github.io/categories/github/ - 2021-01-06T00:00:00+00:00 - - - + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/categories/data-visualization/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/tags/geodata/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/tags/geographic/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + 2023-05-29T00:00:00+00:00 + https://liuyanguu.github.io/post/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/categories/spatial-data/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/tags/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/ + 2023-05-29T00:00:00+00:00 + + https://liuyanguu.github.io/categories/github/ 2021-01-06T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2021/01/06/rmarkdown-to-github-pages/ 2021-01-06T00:00:00+00:00 - - - - https://liuyanguu.github.io/ - 2021-01-06T00:00:00+00:00 - - - - https://liuyanguu.github.io/categories/data-visualization/ - 2020-06-14T00:00:00+00:00 - - - - https://liuyanguu.github.io/tags/geodata/ - 2020-06-14T00:00:00+00:00 - - - - https://liuyanguu.github.io/tags/ - 2020-06-14T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/ 2020-06-14T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2020/06/12/ggplot-us-state-and-china-province-heatmap/ 2020-06-12T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/heatmap/ 2020-06-12T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/array/ 2020-01-11T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/data-engineering/ 2020-01-11T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/ 2020-01-11T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/drake/ 2019-09-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2019/09/15/drake-powerful-tool-for-automatic-reproducible-workflow/ 2019-09-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/markdown/ 2019-09-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/reproducible-workflow/ 2019-09-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2019/07/28/some-experience-on-writing-r-package/ 2019-07-28T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/r-package/ 2019-07-28T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/machine-learning/ 2019-07-18T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/shap/ 2019-07-18T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/ 2019-07-18T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/shapforxgboost/ 2019-07-18T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/xgboost/ 2019-07-18T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/shiny/ 2019-02-24T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/shiny/ 2019-02-24T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/ 2019-02-24T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/dart/ 2018-11-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/ 2018-11-15T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/ 2018-10-14T00:00:00+00:00 - - - - https://liuyanguu.github.io/tags/geographic/ - 2018-07-20T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/r/ 2018-07-20T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/spatial/ 2018-07-20T00:00:00+00:00 - - - - https://liuyanguu.github.io/categories/spatial-data/ - 2018-07-20T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/ 2018-07-20T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/ 2018-07-16T00:00:00+00:00 - - - - https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/ - 2018-07-09T00:00:00+00:00 - - - + https://liuyanguu.github.io/tags/catalog/ 2018-07-01T00:00:00+00:00 - - - + https://liuyanguu.github.io/categories/catalog/ 2018-07-01T00:00:00+00:00 - - - + https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/ 2018-07-01T00:00:00+00:00 - - - + https://liuyanguu.github.io/about-hugo/ - - - + https://liuyanguu.github.io/about-me/ - diff --git a/tags/array/index.html b/tags/array/index.html index 17488bc..6ab278e 100644 --- a/tags/array/index.html +++ b/tags/array/index.html @@ -1,83 +1,81 @@ - - - - - - - Array | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Data Engineering - - -
    -

    Working with 3D array as long-format data in R

    -
    - Yang Liu - / - 2020-01-11 -
    -
    - - - - -
    Thumbnail
    - - - - - - check.and.install.pkgs <- function(pkgs){ new.packages <- pkgs[!pkgs %in% installed.packages()[,"Package"]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c("data.table", "reshape2", "scatterplot3d")) Happy New… -     -
    -
    - -
    - -
    - - - - - - - - - - - - - - + + + + + diff --git a/tags/array/index.xml b/tags/array/index.xml index 200a777..cd840da 100644 --- a/tags/array/index.xml +++ b/tags/array/index.xml @@ -13,9 +13,9 @@ Sat, 11 Jan 2020 00:00:00 +0000 https://liuyanguu.github.io/post/2020/01/11/working-with-3d-array-as-long-format-data-in-r/ - Example: a 3-D array with dimension 4x3x2 Melt into long Recover the array check.and.install.pkgs &lt;- function(pkgs){ new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;)) Happy New Year! + Example: a 3-D array with dimension 4x3x2 Melt into long Recover the array check.and.install.pkgs &lt;- function(pkgs){ new.packages &lt;- pkgs[!pkgs %in% installed.packages()[,&quot;Package&quot;]] if(length(new.packages)) install.packages(new.packages, dependencies = TRUE) suppressPackageStartupMessages(invisible(lapply(pkgs, library, character.only = TRUE))) } check.and.install.pkgs(c(&quot;data.table&quot;, &quot;reshape2&quot;, &quot;scatterplot3d&quot;)) Happy New Year! Recently I spent some time working with array in R. -I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. +I believe it is a bad idea to work with array using for loop, which is both slow and error-prone. We can just melt it into a long data, do the work, and arrange back into array in the end if needed. diff --git a/tags/array/page/1/index.html b/tags/array/page/1/index.html index 2c9dbfe..a8e49bd 100644 --- a/tags/array/page/1/index.html +++ b/tags/array/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/array/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/array/ + + + + + + diff --git a/tags/catalog/index.html b/tags/catalog/index.html index f69f560..3e54eab 100644 --- a/tags/catalog/index.html +++ b/tags/catalog/index.html @@ -23,7 +23,7 @@
    Catalog - 2021-03-07 + 2023-05-30 @@ -72,9 +72,7 @@

    Catalog of my old blog< - +

    @@ -96,7 +94,6 @@

    Catalog of my old blog< -
    @@ -133,7 +130,6 @@

    Catalog of my old blog< } - diff --git a/tags/catalog/index.xml b/tags/catalog/index.xml index 1bb599d..85e9fb8 100644 --- a/tags/catalog/index.xml +++ b/tags/catalog/index.xml @@ -14,7 +14,7 @@ https://liuyanguu.github.io/post/2018/07/01/catalogue-of-my-old-blog/ Introducing my new blog written solely in R Markdown When I realized it was so convenient to write blog directly using R Markdown, I searched if there is a specific tool for it. And I found Blogdown, an R package developed by Yihui Xie, who also developed R Markdown. - Catalog As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018. +Catalog As a summary, I would like to create a catalog for the main topics I wrote on google blogger before July 2018. diff --git a/tags/catalog/page/1/index.html b/tags/catalog/page/1/index.html index d548de0..67c0173 100644 --- a/tags/catalog/page/1/index.html +++ b/tags/catalog/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/catalog/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/catalog/ + + + + + + diff --git a/tags/dart/index.html b/tags/dart/index.html index e9df3da..825d5d4 100644 --- a/tags/dart/index.html +++ b/tags/dart/index.html @@ -1,82 +1,80 @@ - - - - - - - DART | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Machine Learning - - -
    -

    Study shrinkage and DART in xgboost modeling using a simple dataset

    -
    - Yang Liu - / - 2018-11-15 -
    -
    - - - - - - - - + + + + + + + DART | Welcome to my blog + + + + + + + + + + +
    +
    + + + + +
    +
    + + + Machine Learning + + +
    +

    Study shrinkage and DART in xgboost modeling using a simple dataset

    +
    + Yang Liu + / + 2018-11-15 +
    +
    + + + + + + + + It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… - -
    -
    - -
    - -
    - - - -
    +[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/dart/index.xml b/tags/dart/index.xml index 88d1b39..65fc5b5 100644 --- a/tags/dart/index.xml +++ b/tags/dart/index.xml @@ -13,8 +13,8 @@ Thu, 15 Nov 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/ - Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -Data set.seed(123) n0 &lt;- 5 X &lt;- data. + Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. +Data set.seed(123) n0 &lt;- 5 X &lt;- data.frame(x1 = runif(n0), x2 = runif(n0)) Y &lt;- c(1, 5, 20, 50, 100) cbind(X, Y) ## x1 x2 Y ## 1 0. diff --git a/tags/dart/page/1/index.html b/tags/dart/page/1/index.html index 014f39d..8cd24a7 100644 --- a/tags/dart/page/1/index.html +++ b/tags/dart/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/dart/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/dart/ + + + + + + diff --git a/tags/drake/index.html b/tags/drake/index.html index d2568e3..e0c3d0a 100644 --- a/tags/drake/index.html +++ b/tags/drake/index.html @@ -23,7 +23,7 @@
    Drake - 2021-03-07 + 2023-05-30 @@ -71,9 +71,7 @@

    @@ -95,7 +93,6 @@

    @@ -132,7 +129,6 @@

    \ No newline at end of file + + + + https://liuyanguu.github.io/tags/drake/ + + + + + + diff --git a/tags/geodata/index.html b/tags/geodata/index.html index 063ae7a..8d6b51e 100644 --- a/tags/geodata/index.html +++ b/tags/geodata/index.html @@ -1,119 +1,149 @@ - - - - - - - Geodata | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Data Visualization - - -
    -

    US and China Admin1 COVID19 mortality and incidence rate

    -
    - Yang Liu - / - 2020-06-14 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + Geodata | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + Data Visualization + + +
    +

    Plot multiple countries on the world map

    +
    + Yang Liu + / + 2023-05-29 +
    +
    + + + + +
    Thumbnail
    + + + + + + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +[…] (You may download the world map from Github) +suppressPackageStartupMessages({ library("data.table") library("ggplot2")… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    US and China Admin1 COVID19 mortality and incidence rate

    +
    + Yang Liu + / + 2020-06-14 +
    +
    + + + + +
    Thumbnail
    + + + + + Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. Code hosted on my Github repo. -[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here("Code/COVID_make_map.R")) # US dt_JUH_US… -     -
    -
    - -
    -
    - - - Data Visualization - - -
    -

    ggplot US state and China province heatmap

    -
    - Yang Liu - / - 2020-06-12 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here("Code/COVID_make_map.R")) # US dt_JUH_US… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    ggplot US state and China province heatmap

    +
    + Yang Liu + / + 2020-06-12 +
    +
    + + + + +
    Thumbnail
    + + + + + (Updated on 2020-06-12, First posted on 2019-04-17) It sounds easy but turned out not as straight-forward as I thought. I will show: […] 50-state (including Alaska and Hawaii) United States thematic map, with map scale and with state abbreviations […] China thematic map, with map scale, with names of provinces in either English or Chinese -[…] The dt1 below can be used by both… -     -
    -
    - -
    - -
    - - - -
    +[…] The dt1 below can be used by both… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/geodata/index.xml b/tags/geodata/index.xml index 2c55bff..67173f0 100644 --- a/tags/geodata/index.xml +++ b/tags/geodata/index.xml @@ -6,14 +6,25 @@ Recent content in Geodata on Welcome to my blog Hugo -- gohugo.io en-us - Sun, 14 Jun 2020 00:00:00 +0000 + Mon, 29 May 2023 00:00:00 +0000 + + Plot multiple countries on the world map + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + Mon, 29 May 2023 00:00:00 +0000 + + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +The most straightforward way (You may download the world map from Github) +suppressPackageStartupMessages({ library(&quot;data.table&quot;) library(&quot;ggplot2&quot;) library(&quot;rgdal&quot;) library(&quot;raster&quot;) library(&quot;rgeos&quot;) library(&quot;here&quot;) library(&quot;ggthemes&quot;) }) # download data from GADM directly and row-bind the spatial polygons data frame cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;) download. + + US and China Admin1 COVID19 mortality and incidence rate https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/ Sun, 14 Jun 2020 00:00:00 +0000 https://liuyanguu.github.io/post/2020/06/14/us-and-china-admin1-covid19-mortality-and-incidence-rate/ - US China Compared to lung cancer mortality Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. + US China Compared to lung cancer mortality Similar to the last heatmap post I just revised, here I wrapped up the function and showing COVID-19 data just downloaded from JHU CSSE Github page. The function can take a given dataset and plot designated variable. Code hosted on my Github repo. US # major function, can download from Github repo Blogdown/hugo-xmag/Code source(here::here(&quot;Code/COVID_make_map.R&quot;)) # US dt_JUH_US &lt;- get.JHU.us.state() make_heatmap(data = dt_JUH_US, geo_data = get_state_name(), state_var = &quot;Province_State&quot;, fill_var = &quot;Mortality_Rate&quot;, label_var = &quot;abb&quot;) us_maps &lt;- lapply(c(&quot;Mortality_Rate&quot;, &quot;Incident_Rate&quot;, &quot;Testing_Rate&quot;, &quot;Hospitalization_Rate&quot;), make_heatmap, data = dt_JUH_US, geo_data = get_state_name(), state_var = &quot;Province_State&quot;, label_var = &quot;abb&quot;) plot_grid &lt;- gridExtra::grid. diff --git a/tags/geodata/page/1/index.html b/tags/geodata/page/1/index.html index 43b5734..d4d2a3e 100644 --- a/tags/geodata/page/1/index.html +++ b/tags/geodata/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/geodata/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/geodata/ + + + + + + diff --git a/tags/geographic/index.html b/tags/geographic/index.html index 0f92dd9..8b83a90 100644 --- a/tags/geographic/index.html +++ b/tags/geographic/index.html @@ -1,69 +1,101 @@ - - - - - - - Geographic | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Data Visualization - - -
    -

    Spatial data in R: Dividing raster layers into equal-area rings

    -
    - Yang Liu - / - 2018-07-20 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + Geographic | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + Data Visualization + + +
    +

    Plot multiple countries on the world map

    +
    + Yang Liu + / + 2023-05-29 +
    +
    + + + + +
    Thumbnail
    + + + + + + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +[…] (You may download the world map from Github) +suppressPackageStartupMessages({ library("data.table") library("ggplot2")… +     +
    +
    + +
    +
    + + + Data Visualization + + +
    +

    Spatial data in R: Dividing raster layers into equal-area rings

    +
    + Yang Liu + / + 2018-07-20 +
    +
    + + + + +
    Thumbnail
    + + + + + This data visualization example include: * Import .img file as a raster * Turn raster into a data.frame of points (coordinates) and values @@ -71,20 +103,18 @@

    - -

    +
    + +
    + +
    + + + +
    - - - - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    -
    - - - - - + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    + + + + + + - - - - - + + + + + diff --git a/tags/geographic/index.xml b/tags/geographic/index.xml index 90bce6f..7c1b1aa 100644 --- a/tags/geographic/index.xml +++ b/tags/geographic/index.xml @@ -6,14 +6,25 @@ Recent content in Geographic on Welcome to my blog Hugo -- gohugo.io en-us - Fri, 20 Jul 2018 00:00:00 +0000 + Mon, 29 May 2023 00:00:00 +0000 + + Plot multiple countries on the world map + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + Mon, 29 May 2023 00:00:00 +0000 + + https://liuyanguu.github.io/post/2023/05/29/plot-multiple-countries-on-the-world-map/ + This post will show it’s quite easy to download and plot the administrative areas of multiple countries on the world map. I will also showcase a bug that puzzled me for a long time and I recently figured out: strange connecting lines among countries! +The most straightforward way (You may download the world map from Github) +suppressPackageStartupMessages({ library(&quot;data.table&quot;) library(&quot;ggplot2&quot;) library(&quot;rgdal&quot;) library(&quot;raster&quot;) library(&quot;rgeos&quot;) library(&quot;here&quot;) library(&quot;ggthemes&quot;) }) # download data from GADM directly and row-bind the spatial polygons data frame cnames &lt;- c(&quot;Haiti&quot;, &quot;Togo&quot;, &quot;Uganda&quot;, &quot;Ghana&quot;, &quot;South Africa&quot;, &quot;Angola&quot;) download. + + Spatial data in R: Dividing raster layers into equal-area rings https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/ Fri, 20 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/ - Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: + Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: * Import .img file as a raster * Turn raster into a data.frame of points (coordinates) and values * Dividing the points into 100 equal-area rings diff --git a/tags/geographic/page/1/index.html b/tags/geographic/page/1/index.html index 241303a..3c11a5d 100644 --- a/tags/geographic/page/1/index.html +++ b/tags/geographic/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/geographic/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/geographic/ + + + + + + diff --git a/tags/heatmap/index.html b/tags/heatmap/index.html index 1c625b8..d3174ac 100644 --- a/tags/heatmap/index.html +++ b/tags/heatmap/index.html @@ -1,87 +1,85 @@ - - - - - - - Heatmap | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - Data Visualization - - -
    -

    ggplot US state and China province heatmap

    -
    - Yang Liu - / - 2020-06-12 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + Heatmap | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + Data Visualization + + +
    +

    ggplot US state and China province heatmap

    +
    + Yang Liu + / + 2020-06-12 +
    +
    + + + + +
    Thumbnail
    + + + + + (Updated on 2020-06-12, First posted on 2019-04-17) It sounds easy but turned out not as straight-forward as I thought. I will show: […] 50-state (including Alaska and Hawaii) United States thematic map, with map scale and with state abbreviations […] China thematic map, with map scale, with names of provinces in either English or Chinese -[…] The dt1 below can be used by both… -     -
    -
    - -
    - -
    - - - -
    +[…] The dt1 below can be used by both… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/heatmap/page/1/index.html b/tags/heatmap/page/1/index.html index f37dd1e..520a604 100644 --- a/tags/heatmap/page/1/index.html +++ b/tags/heatmap/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/heatmap/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/heatmap/ + + + + + + diff --git a/tags/index.html b/tags/index.html index 51027d3..612a773 100644 --- a/tags/index.html +++ b/tags/index.html @@ -23,7 +23,7 @@
    Tags - 2021-03-07 + 2023-05-30 @@ -72,14 +72,14 @@

    Tags

    geodata - (2) + (3)
  • geographic - (1) + (2)
  • @@ -135,7 +135,7 @@

    Tags

    xgboost - (4) + (3)
    • @@ -162,7 +162,6 @@

    Tags

    -
    @@ -199,7 +198,6 @@

    Tags

    } - diff --git a/tags/index.xml b/tags/index.xml index e6dab18..fea55ea 100644 --- a/tags/index.xml +++ b/tags/index.xml @@ -6,16 +6,25 @@ Recent content in Tags on Welcome to my blog Hugo -- gohugo.io en-us - Sun, 14 Jun 2020 00:00:00 +0000 + Mon, 29 May 2023 00:00:00 +0000 Geodata https://liuyanguu.github.io/tags/geodata/ - Sun, 14 Jun 2020 00:00:00 +0000 + Mon, 29 May 2023 00:00:00 +0000 https://liuyanguu.github.io/tags/geodata/ + + Geographic + https://liuyanguu.github.io/tags/geographic/ + Mon, 29 May 2023 00:00:00 +0000 + + https://liuyanguu.github.io/tags/geographic/ + + + Heatmap https://liuyanguu.github.io/tags/heatmap/ @@ -97,15 +106,6 @@ - - Geographic - https://liuyanguu.github.io/tags/geographic/ - Fri, 20 Jul 2018 00:00:00 +0000 - - https://liuyanguu.github.io/tags/geographic/ - - - R https://liuyanguu.github.io/tags/r/ diff --git a/tags/markdown/index.html b/tags/markdown/index.html index ece4b44..2975ec0 100644 --- a/tags/markdown/index.html +++ b/tags/markdown/index.html @@ -23,7 +23,7 @@
    Markdown - 2021-03-07 + 2023-05-30 @@ -105,9 +105,7 @@

    How

    - +
    @@ -129,7 +127,6 @@

    How -
    @@ -166,7 +163,6 @@

    How } - diff --git a/tags/markdown/index.xml b/tags/markdown/index.xml index 2295511..9152c08 100644 --- a/tags/markdown/index.xml +++ b/tags/markdown/index.xml @@ -23,10 +23,10 @@ RMarkdown file could be very slow to generate if lots of calculations are involv Mon, 16 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/ - Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: + Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: * Hierarchical clustering, dendrogram and heat map based on normalized odds ratios * The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend -* Heatmap is made by heatmap. +* Heatmap is made by heatmap.2 from gplots using the built dendrogram diff --git a/tags/markdown/page/1/index.html b/tags/markdown/page/1/index.html index 130a9b8..2e7d26a 100644 --- a/tags/markdown/page/1/index.html +++ b/tags/markdown/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/markdown/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/markdown/ + + + + + + diff --git a/tags/page/1/index.html b/tags/page/1/index.html index 1977e09..2006a69 100644 --- a/tags/page/1/index.html +++ b/tags/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/ + + + + + + diff --git a/tags/page/2/index.html b/tags/page/2/index.html index a27ce01..c05d2d1 100644 --- a/tags/page/2/index.html +++ b/tags/page/2/index.html @@ -23,7 +23,7 @@
    Tags - 2021-03-07 + 2023-05-30 @@ -72,14 +72,14 @@

    Tags

     geodata - (2) + (3)
  • geographic - (1) + (2)
  • @@ -135,7 +135,7 @@

    Tags

    xgboost - (4) + (3)
    • @@ -162,7 +162,6 @@

    Tags

    -
    @@ -199,7 +198,6 @@

    Tags

    } - diff --git a/tags/page/3/index.html b/tags/page/3/index.html index 1c6b235..812727a 100644 --- a/tags/page/3/index.html +++ b/tags/page/3/index.html @@ -23,7 +23,7 @@
    Tags - 2021-03-07 + 2023-05-30 @@ -72,14 +72,14 @@

    Tags

    geodata - (2) + (3)
  • geographic - (1) + (2)
  • @@ -135,7 +135,7 @@

    Tags

    xgboost - (4) + (3)
    • @@ -162,7 +162,6 @@

    Tags

    -
    @@ -199,7 +198,6 @@

    Tags

    } - diff --git a/tags/r/index.html b/tags/r/index.html index d44167f..4b44678 100644 --- a/tags/r/index.html +++ b/tags/r/index.html @@ -23,7 +23,7 @@
    R - 2021-03-07 + 2023-05-30 @@ -111,9 +111,7 @@

    How - +

    @@ -135,7 +133,6 @@

    How -
    @@ -172,7 +169,6 @@

    How } - diff --git a/tags/r/index.xml b/tags/r/index.xml index cbc59aa..b483201 100644 --- a/tags/r/index.xml +++ b/tags/r/index.xml @@ -13,7 +13,7 @@ Fri, 20 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/20/spatial-data-in-r-dividing-raster-layers-into-equal-area-rings/ - Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: + Saturation in ten cities with equal-area rings R Code for one city Results for the ring saturations Average saturation in each ring Methodology Original Code This data visualization example include: * Import .img file as a raster * Turn raster into a data.frame of points (coordinates) and values * Dividing the points into 100 equal-area rings @@ -27,10 +27,10 @@ Mon, 16 Jul 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/07/16/how-to-draw-heatmap-with-colorful-dendrogram/ - Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: + Data Version 1: Color both the branches and labels Version 2: color only the labels. Version 3: If there is no color, and we do not reorder the branches This data visualization example include: * Hierarchical clustering, dendrogram and heat map based on normalized odds ratios * The dendrogram was built separately to give color to dendrogram’s branches/labels based on cluster using dendextend -* Heatmap is made by heatmap. +* Heatmap is made by heatmap.2 from gplots using the built dendrogram diff --git a/tags/r/page/1/index.html b/tags/r/page/1/index.html index df80e38..94ac9df 100644 --- a/tags/r/page/1/index.html +++ b/tags/r/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/r/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/r/ + + + + + + diff --git a/tags/shap/index.html b/tags/shap/index.html index 460ed0b..c04a9ad 100644 --- a/tags/shap/index.html +++ b/tags/shap/index.html @@ -1,118 +1,116 @@ - - - - - - - SHAP | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - R package - - -
    -

    SHAP for XGBoost in R: SHAPforxgboost

    -
    - Yang Liu - / - 2019-07-18 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + SHAP | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + R package + + +
    +

    SHAP for XGBoost in R: SHAPforxgboost

    +
    + Yang Liu + / + 2019-07-18 +
    +
    + + + + +
    Thumbnail
    + + + + + I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… -     -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    SHAP Visualization in R (first post)

    -
    - Yang Liu - / - 2018-10-14 -
    -
    - - - - -
    Thumbnail
    - - - - - +install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… +     +
    +
    + +
    +
    + + + Machine Learning + + +
    +

    SHAP Visualization in R (first post)

    +
    + Yang Liu + / + 2018-10-14 +
    +
    + + + + +
    Thumbnail
    + + + + + Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R […] This is the example I used in the package SHAPforxgboost -# Example use iris suppressPackageStartupMessages({… -     -
    -
    - -
    - -
    - - - -
    +# Example use iris suppressPackageStartupMessages({… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/shap/index.xml b/tags/shap/index.xml index df5ea10..771ad87 100644 --- a/tags/shap/index.xml +++ b/tags/shap/index.xml @@ -13,9 +13,9 @@ Thu, 18 Jul 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/ - The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. + The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. +install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. @@ -24,7 +24,7 @@ install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_gith Sun, 14 Oct 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/ - Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: + Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R diff --git a/tags/shap/page/1/index.html b/tags/shap/page/1/index.html index 08bdf21..bc93a40 100644 --- a/tags/shap/page/1/index.html +++ b/tags/shap/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/shap/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/shap/ + + + + + + diff --git a/tags/shapforxgboost/index.html b/tags/shapforxgboost/index.html index 59ce76d..7fb0c2a 100644 --- a/tags/shapforxgboost/index.html +++ b/tags/shapforxgboost/index.html @@ -1,85 +1,83 @@ - - - - - - - SHAPforxgboost | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - R package - - -
    -

    SHAP for XGBoost in R: SHAPforxgboost

    -
    - Yang Liu - / - 2019-07-18 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + SHAPforxgboost | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + R package + + +
    +

    SHAP for XGBoost in R: SHAPforxgboost

    +
    + Yang Liu + / + 2019-07-18 +
    +
    + + + + +
    Thumbnail
    + + + + + I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… -     -
    -
    - -
    - -
    - - - -
    +install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/shapforxgboost/index.xml b/tags/shapforxgboost/index.xml index 208cc43..4fdc660 100644 --- a/tags/shapforxgboost/index.xml +++ b/tags/shapforxgboost/index.xml @@ -13,9 +13,9 @@ Thu, 18 Jul 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/ - The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. + The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. +install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. diff --git a/tags/shapforxgboost/page/1/index.html b/tags/shapforxgboost/page/1/index.html index a5c7c4b..c53cd65 100644 --- a/tags/shapforxgboost/page/1/index.html +++ b/tags/shapforxgboost/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/shapforxgboost/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/shapforxgboost/ + + + + + + diff --git a/tags/shiny/index.html b/tags/shiny/index.html index 2dd11ba..0d02e08 100644 --- a/tags/shiny/index.html +++ b/tags/shiny/index.html @@ -23,7 +23,7 @@
    Shiny - 2021-03-07 + 2023-05-30 @@ -71,9 +71,7 @@

    Shiny in Blogdown

    - +
    @@ -95,7 +93,6 @@

    Shiny in Blogdown

    -
    @@ -132,7 +129,6 @@

    Shiny in Blogdown

    } - diff --git a/tags/shiny/index.xml b/tags/shiny/index.xml index 4d5aa01..8b47ae1 100644 --- a/tags/shiny/index.xml +++ b/tags/shiny/index.xml @@ -13,10 +13,10 @@ Sun, 24 Feb 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/02/24/shiny-in-blogdown/ - How to embed Shiny My Shiny app example How to embed Shiny Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here. + How to embed Shiny My Shiny app example How to embed Shiny Since Blogdown is for static websites, it cannot run Shiny in rmarkdown directly. According to discussion here and document here. The only way to do it is using iframe and write outside the chunk: &lt;iframe src="https://liuyanguu.shinyapps.io/bcl_app/" width=1000 height=800"&gt;&lt;/iframe&gt; -There is also a built-in function in knitr to do the same thing and write in the chunk. +There is also a built-in function in knitr to do the same thing and write in the chunk. The only difference is that we can only set height, and the shiny app would appear slightly different with the siderbarPanel at the top and the mainPanel beneath it. diff --git a/tags/shiny/page/1/index.html b/tags/shiny/page/1/index.html index 3b2808c..edb64f7 100644 --- a/tags/shiny/page/1/index.html +++ b/tags/shiny/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/shiny/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/shiny/ + + + + + + diff --git a/tags/spatial/index.html b/tags/spatial/index.html index b74deb1..112a482 100644 --- a/tags/spatial/index.html +++ b/tags/spatial/index.html @@ -23,7 +23,7 @@
    Spatial - 2021-03-07 + 2023-05-30 @@ -77,9 +77,7 @@

    @@ -101,7 +99,6 @@

    @@ -138,7 +135,6 @@

    \ No newline at end of file + + + + https://liuyanguu.github.io/tags/spatial/ + + + + + + diff --git a/tags/xgboost/index.html b/tags/xgboost/index.html index 4148ff0..c6fe7b5 100644 --- a/tags/xgboost/index.html +++ b/tags/xgboost/index.html @@ -1,177 +1,145 @@ - - - - - - - XGBoost | Welcome to my blog - - - - - - - - - - -
    -
    - - - - -
    -
    - - - R package - - -
    -

    SHAP for XGBoost in R: SHAPforxgboost

    -
    - Yang Liu - / - 2019-07-18 -
    -
    - - - - -
    Thumbnail
    - - - - - + + + + + + + XGBoost | Welcome to my blog + + + + + + + +     + + +
    +
    + + + + +
    +
    + + + R package + + +
    +

    SHAP for XGBoost in R: SHAPforxgboost

    +
    + Yang Liu + / + 2019-07-18 +
    +
    + + + + +
    Thumbnail
    + + + + + I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… -     -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    Study shrinkage and DART in xgboost modeling using a simple dataset

    -
    - Yang Liu - / - 2018-11-15 -
    -
    - - - - - - - - +install.packages("SHAPforxgboost") # or devtools::install_github("liuyanguu/SHAPforxgboost") Why SHAP values SHAP’s main advantages are local explanation and consistency in global model… + +
    +
    + +
    +
    + + + Machine Learning + + +
    +

    Study shrinkage and DART in xgboost modeling using a simple dataset

    +
    + Yang Liu + / + 2018-11-15 +
    +
    + + + + + + + + It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… - -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    SHAP Visualization in R (first post)

    -
    - Yang Liu - / - 2018-10-14 -
    -
    - - - - -
    Thumbnail
    - - - - - +[…] set.seed(123) n0 <- 5 X <- data.frame(x1 = runif(n0), x2 = runif(n0)) Y <- c(1, 5, 20, 50, 100) cbind(X,… +     +
    +
    + +
    +
    + + + Machine Learning + + +
    +

    SHAP Visualization in R (first post)

    +
    + Yang Liu + / + 2018-10-14 +
    +
    + + + + +
    Thumbnail
    + + + + + Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R […] This is the example I used in the package SHAPforxgboost -# Example use iris suppressPackageStartupMessages({… -     -
    -
    - -
    -
    - - - Machine Learning - - -
    -

    eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting

    -
    - Yang Liu - / - 2018-07-09 -
    -
    - - - - -
    Thumbnail
    - - - - - - I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset. The ensemble method is powerful as it combines the predictions from multiple machine learning algorithms together to… -     -
    -
    - -
    - -
    - - - -
    +# Example use iris suppressPackageStartupMessages({… + +
    +
    + +
    + +
    + + + +
    + + + + +
    + + +
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    + +
    +
    + + - -
    - - -
    @2021 Yang Liu github.com/liuyanguu. Powered by Github.
    - -
    - - - - - - - - - - - + + + + + diff --git a/tags/xgboost/index.xml b/tags/xgboost/index.xml index 3236ff2..ade3201 100644 --- a/tags/xgboost/index.xml +++ b/tags/xgboost/index.xml @@ -13,9 +13,9 @@ Thu, 18 Jul 2019 00:00:00 +0000 https://liuyanguu.github.io/post/2019/07/18/visualization-of-shap-for-xgboost/ - The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. + The SHAPforxgboost package Why SHAP values Local explanation Consistency in global feature importance SHAP plots Summary plot Dependence plot Interaction effects SHAP force plot References The SHAPforxgboost package I wrote the R package SHAPforxgboost to cover all the plotting functions illustrated in this post. This post serves as the vignette for the package. Please install from CRAN or Github. -install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. +install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_github(&quot;liuyanguu/SHAPforxgboost&quot;) Why SHAP values SHAP’s main advantages are local explanation and consistency in global model structure. @@ -24,8 +24,8 @@ install.packages(&quot;SHAPforxgboost&quot;) # or devtools::install_gith Thu, 15 Nov 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/11/15/xgboost-dart-example/ - Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. -Data set.seed(123) n0 &lt;- 5 X &lt;- data. + Data Shrinkage DART: Dropout - MART skip_drop rate_drop one_drop It is always a good idea to study the packaged algorithm with a simple example. Inspired by my colleague Kodi’s excellent work showing how xgboost handles missing values, I tried a simple 5x2 dataset to show how shrinkage and DART influence the growth of trees in the model. +Data set.seed(123) n0 &lt;- 5 X &lt;- data.frame(x1 = runif(n0), x2 = runif(n0)) Y &lt;- c(1, 5, 20, 50, 100) cbind(X, Y) ## x1 x2 Y ## 1 0. @@ -34,18 +34,9 @@ Data set.seed(123) n0 &lt;- 5 X &lt;- data. Sun, 14 Oct 2018 00:00:00 +0000 https://liuyanguu.github.io/post/2018/10/14/shap-visualization-for-xgboost/ - Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: + Example 1 SHAP summary plot Alternative ways: SHAP dependence plot SHAP interaction effect plot SHAP force plot Example 2 Summary plot Dependence plot for each feature Force plot Stack plot by clustering groups Update 19/07/21: Since my R Package SHAPforxgboost has been released on CRAN, I updated this post using the new functions and illustrate how to use these functions using two datasets. For more information, please refer to: SHAP visualization for XGBoost in R - - eXtreme Gradient Boosting (XGBoost): Better than random forest or gradient boosting - https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/ - Mon, 09 Jul 2018 00:00:00 +0000 - - https://liuyanguu.github.io/post/2018/07/09/extreme-gradient-boosting-xgboost-better-than-random-forest-or-gradient-boosting/ - Overview About the Data 1. Extreme Gradient Boosting 2. Gradient boosting 3. Random Forest 4. Lasso 5. Best Subset Compare MSEs Original Code are stored on github Overview I recently had the great pleasure to meet with Professor Allan Just and he introduced me to eXtreme Gradient Boosting (XGBoost). I have extended the earlier work on my old blog by comparing the results across XGBoost, Gradient Boosting (GBM), Random Forest, Lasso, and Best Subset. - - diff --git a/tags/xgboost/page/1/index.html b/tags/xgboost/page/1/index.html index dc6f0f0..d29c0f2 100644 --- a/tags/xgboost/page/1/index.html +++ b/tags/xgboost/page/1/index.html @@ -1 +1,10 @@ -https://liuyanguu.github.io/tags/xgboost/ \ No newline at end of file + + + + https://liuyanguu.github.io/tags/xgboost/ + + + + + +