begin docs update

MolloiLab · Nov 18, 2023 · 62b0618 · 62b0618
1 parent 2ba19e8
commit 62b0618
Show file tree

Hide file tree

Showing 4 changed files with 30 additions and 14 deletions.
diff --git a/docs/02_multi_threading.jl b/docs/02_multi_threading.jl
@@ -1,5 +1,5 @@
 ### A Pluto.jl notebook ###
-# v0.19.26
+# v0.19.32
 
 #> [frontmatter]
 #> title = "Multi-threading"
@@ -32,7 +32,7 @@ TableOfContents()
 # ╔═╡ ede39bd3-1e20-4426-a518-abe2f574aa91
 md"""
 ## Multi-threaded distance transforms
-One of the advantages of a Julia-based distance transform library is how accessible something like multi-threading is for complex algorithms. The `Felzenszwalb` distance transform is highly parallelizable and is set up to take advantage of multi-threaded hardware.
+One of the advantages of a Julia-based distance transform library is how accessible something like multi-threading is for complex algorithms. The felzenszwalb distance transform algorithm is highly parallelizable, which is what the CPU approach is based on, and is set up to take advantage of multi-threaded hardware by default.
 """
 
 # ╔═╡ 735aa1eb-2ce8-4ae2-8181-76056aa560de
@@ -54,11 +54,8 @@ array = [
 # ╔═╡ c8bd961f-14d7-4724-92f0-3d8cb24b239d
 bool_array = boolean_indicator(array)
 
-# ╔═╡ a0b300ff-2835-4440-a34d-cd1d3e5ed346
-tfm = Felzenszwalb()
-
 # ╔═╡ e38b8b56-a2cb-46a8-b5c8-9577a846ae1f
-sq_euc_transform = transform(bool_array, tfm, threads)
+sq_euc_transform = transform(bool_array)
 
 # ╔═╡ 6001d3d3-a067-4df3-a613-873fb4168c11
 md"""
@@ -69,7 +66,7 @@ md"""
 array3D = rand([0, 1], 5, 5, 3)
 
 # ╔═╡ 86c7fa96-b05d-42c8-b86b-331993df95fc
-sq_euc_transform3D = transform(boolean_indicator(array3D), tfm, threads)
+sq_euc_transform3D = transform(boolean_indicator(array3D))
 
 # ╔═╡ b7651361-7332-4215-84dd-5cf55c61768d
 md"""
@@ -92,14 +89,14 @@ begin
 
 		# Squared Euclidean DT
 		x2 = boolean_indicator(rand([0, 1], n, n))
-		sedt = @benchmark transform($x2, $tfm)
+		sedt = @benchmark transform($x2; threaded = false)
 
 		append!(sedt_mean, BenchmarkTools.mean(sedt).time)
 		append!(sedt_std, BenchmarkTools.std(sedt).time)
 
 		# Squared Euclidean DT threaded
 		x3 = boolean_indicator(rand([0, 1], n, n))
-		sedtP = @benchmark transform($x3, $tfm, $threads)
+		sedtP = @benchmark transform($x3; threaded = true)
 
 		append!(sedtP_mean, BenchmarkTools.mean(sedtP).time)
 		append!(sedtP_std, BenchmarkTools.std(sedtP).time)
@@ -145,7 +142,6 @@ One might wonder why, if multi-threading is so beneficial, GPUs wouldn't be util
 # ╟─5104bf97-e5bf-463e-a68a-cf1a9a7f77e8
 # ╠═c7914c8a-5c30-482d-9776-1121c1268b81
 # ╠═c8bd961f-14d7-4724-92f0-3d8cb24b239d
-# ╠═a0b300ff-2835-4440-a34d-cd1d3e5ed346
 # ╠═e38b8b56-a2cb-46a8-b5c8-9577a846ae1f
 # ╟─6001d3d3-a067-4df3-a613-873fb4168c11
 # ╠═ae31df46-a7c8-41f3-b8e3-d5eeb1ba317c

diff --git a/docs/03_gpu.jl b/docs/03_gpu.jl
@@ -1,5 +1,5 @@
 ### A Pluto.jl notebook ###
-# v0.19.26
+# v0.19.32
 
 #> [frontmatter]
 #> title = "GPU"
@@ -91,6 +91,6 @@ One such example can be seen in the next tutorial on using distance transforms w
 # ╟─ed7fea36-60ca-47e0-9046-bf2565fdd41d
 # ╠═22cec81f-c503-43de-8297-ec0591e4e6c5
 # ╠═afeb0802-60c1-40e0-a9e5-b13ad6af27bc
-# ╟─b370e3a1-37ec-48aa-86e4-d053bf9bce3b
+# ╠═b370e3a1-37ec-48aa-86e4-d053bf9bce3b
 # ╠═b885869b-985e-4d73-acd9-b381e6272540
 # ╟─8488fa03-8ad4-4b81-8877-4ec75018a84f
diff --git a/docs/Manifest.toml b/docs/Manifest.toml
@@ -397,11 +397,11 @@ version = "1.15.1"
 
 [[deps.DistanceTransforms]]
 deps = ["GPUArraysCore", "KernelAbstractions", "LoopVectorization"]
-git-tree-sha1 = "78d24c6fabfa651c61bc3c85ea724beb7620d01f"
+git-tree-sha1 = "6e6e2569aea7e4470c251f713c103f3593a679fb"
 repo-rev = "master"
 repo-url = "https://github.com/Dale-Black/DistanceTransforms.jl.git"
 uuid = "71182807-4d06-4237-8dd0-bdafe4d097e2"
-version = "0.2.0"
+version = "0.2.1"
 
 [[deps.DistanceTransformsPy]]
 deps = ["CondaPkg", "PythonCall"]

diff --git a/index.jl b/index.jl
@@ -29,6 +29,11 @@ md"""
 ## Tutorials
 """
 
+# ╔═╡ 8b5d431a-f91c-422b-86cf-3ace891e742f
+md"""
+## Python
+"""
+
 # ╔═╡ 71604c4c-6d3a-4b22-bb47-8144604148be
 md"""
 ## Benchmarks
@@ -83,6 +88,11 @@ struct Article
 	image_url::String
 end
 
+# ╔═╡ e0ef1b4d-0076-410c-b325-32a7a274f487
+article_list_python = Article[
+	Article("Python", "https://colab.research.google.com/drive/1-CDqQgrBHoxNqs2IbMebMRxsp0m21jSa?usp=sharing", "https://img.freepik.com/free-vector/code-typing-concept-illustration_114360-3581.jpg?size=626&ext=jpg&ga=GA1.1.1694943658.1700350224&semt=sph"),
+];
+
 # ╔═╡ b7c92fa4-ed51-495d-ba82-5ae61b3de194
 article_list_quickstart = Article[
 	Article("Getting Started", "docs/01_getting_started.jl", "https://img.freepik.com/free-photo/futuristic-spaceship-takes-off-into-purple-galaxy-fueled-by-innovation-generated-by-ai_24640-100023.jpg"),
@@ -134,6 +144,13 @@ to_html(
     )
 )
 
+# ╔═╡ 75f02433-7ef3-45f4-ad2a-d610923824af
+to_html(
+    divv(:class => "flex flex-wrap justify-center items-start",
+        [article_card(article, "secondary"; data_theme = data_theme) for article in article_list_python]...
+    )
+)
+
 # ╔═╡ 1198c970-07a0-464a-b853-aa737634c0e5
 to_html(
     divv(:class => "flex flex-wrap justify-center items-start",
@@ -154,6 +171,8 @@ to_html(
 # ╟─9f5666d8-b86b-46d1-a97a-f7f708fdedb6
 # ╟─5e70dc3d-85b6-4149-9962-70215d3c7f1e
 # ╟─ace65d98-fe6b-45b7-990d-df8f59f6a5b3
+# ╟─8b5d431a-f91c-422b-86cf-3ace891e742f
+# ╟─75f02433-7ef3-45f4-ad2a-d610923824af
 # ╟─71604c4c-6d3a-4b22-bb47-8144604148be
 # ╟─1198c970-07a0-464a-b853-aa737634c0e5
 # ╟─b3349485-1a5f-4da6-be79-965fec5fbc13
@@ -166,6 +185,7 @@ to_html(
 # ╟─09b3d1e1-b44f-461d-abaf-2e5a1f6f6cc2
 # ╟─b7c92fa4-ed51-495d-ba82-5ae61b3de194
 # ╟─0d222d9d-7576-40a7-acb4-b0e27f9a9ef8
+# ╟─e0ef1b4d-0076-410c-b325-32a7a274f487
 # ╟─6cf265a0-5b72-4e5d-bb16-77ff69de566e
 # ╟─8d11bae1-d232-4c03-981a-cd9f1db24e6b
 # ╟─3bcd40f0-e8dd-48bd-8249-51b89f5766de