B4:Bilder hinzugefügt

.github/pull_request_template.md

@@ -11,7 +11,7 @@ Please follow these Steps & Guidelines for your Pull Request:
 [] make sure that your PR contains only files / file changes within your project subdirectory. 
 [] make sure that all your file names consist only of lower case letters, numbers, or [-_.] - with only one . for the file extension
 [] delete all unnecessary files from your PR (including cat.jpg and kitty.jpg)
-[] check the image file sizes. All Images have to be as small as possible. See [./doc/disk_usage.md](./doc/disk_usage.md) about this issues, there's a script: [./bin/images-smaller.sh](./bin/images-smaller.sh)
+[] check the image file sizes. All Images have to be as small as possible. See [./doc/disk_usage.md](https://github.com/htw-imi-showtime/showtime-website/blob/main/doc/disk_usage.md) about this issues, there's a script: [./bin/images-smaller.sh](https://github.com/htw-imi-showtime/showtime-website/blob/main/bin/images-smaller.sh)
 
 **Note that if you fail to do so, I may simply reduce the file sizes brute force using 
 the script above - WITHOUT CHECKING THE QUALITY**
@@ -40,6 +40,10 @@ the script above - WITHOUT CHECKING THE QUALITY**
 
 ### More Information
 
+We are using the Forking Workflow.
+
 - [Best Practices for Pull Requests](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/getting-started/best-practices-for-pull-requests)
+- Atlassian Git Tutorial: [Forking Workflow](https://www.atlassian.com/git/tutorials/comparing-workflows/forking-workflow)
+- GitHub Docs: [Collaborating with pull requests](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests)
+- Documentation in [README.MD](https://github.com/htw-imi-showtime/showtime-website/blob/main/README.MD) and [/doc](https://github.com/htw-imi-showtime/showtime-website/blob/main/doc)
 
-- Documentation in README.MD and /doc

content/ss24/bachelor/b4-sempy/_index.md

@@ -1,44 +1,44 @@
 +++
 project_id = "B4"
-title = "Sempy"
-subtitle = "AI-Research Agent"
-claim = "SEMPY is an AI-Research Agent that supports customers of the idealo website to help them make better buying decission. This is tool takes different information from different websites and summarizes them and gives the best advice."
+title = "SEMPY AI-Research Agent"
+subtitle = ""
+claim = ""
 abstract = ""
 
 # Properties for displaying the project in the project list
-card_image = "Logo.jpg"
+card_image = "logo.jpg"
 # card_description is deprecated, will be replaced by abstract/and or claim
-card_description = "" 
+card_description = "SEMPY is an AI research agent that supports customers of the idealo website in making better buying decisions. This tool gathers information from various websites, summarizes it, and provides the best advice." 
 
 # Names are optional, team size is sufficient
 team = ["Son Xuan Hoang","Eldar Akhundzada", "Mathieu Wassmuth", "Paul Schiffner", "Yasin Dalowar"]
-supervisor = ["Jakob Panten", "Jana Dolan", "Tino Nitze"]
+supervisor = "Jakob Panten & Jana Dolan & Tino Nitze"
 draft = false
 
 website_link = ""
 source_link = ""
 
-tags = ["ai"]
-categories = ["External Stakeholder"]
-study_focus = ['Web Technology']
 +++
 
+
 {{<section title="Our Goal">}}
-The goal of this project is to offer an AI driven research agent on idealo's price comparison platform that supports users who have not yet made a decision on which product to buy and are looking for additional information. 
+In recent years, artificial intelligence technology has advanced significantly, and today many people use technologies such as OpenAI's GPT-4, Google's TensorFlow, and IBM Watson because these tools make life easier. Our goal is to simplify the lives of customers on idealo, helping them with their purchasing decisions and saving them a lot of time. We aim to leverage this technology to enhance the customer experience on the idealo site. Our SEMPY research aims to gather the most important product data and customer opinions in one place, avoiding the need to search on different platforms such as YouTube, Amazon, and Google. 
 
 {{</section>}}
 
+{{<team-member image="Video12.gif" name="App Functionality">}}
+
 
 {{<section title="The team">}}
-The SEMPY team consists of 5 young IMI students that came together to build this project. Through this project, we were able to gain a lot of knowledge about working with AI. During the development, we learned a lot by helping each other.
+The SEMPY team consists of five young IMI students who came together to build this project. None of us had experience working with the OpenAI API, so it was quite a challenge for us. Through this project, we gained a lot of knowledge about working with AI and learned a lot by supporting each other during the development.
 
 {{</section>}} 
 
 {{<gallery>}}
-<!-- {{<team-member image="cat.jpg" name="Son Xuan Hoang">}} -->
-{{<team-member image="Paul.jpg" name="Paul Schiffner">}}
-<!-- {{<team-member image="cat.jpg" name="Yasin Dalowar">}} -->
-<!-- {{<team-member image="cat.jpg" name="Mathieu Wassmuth">}}
-{{<team-member image="cat.jpg" name="Eldar Akhundzada">}} -->
+{{<team-member image="son.jpg" name="Son Xuan Hoang">}} 
+{{<team-member image="paul.jpg" name="Paul Schiffner">}}
+{{<team-member image="yasin.jpg" name="Yasin Dalowar">}}
+{{<team-member image="mathieu.jpg" name="Mathieu Wassmuth">}}
+{{<team-member image="eldar.jpg" name="Eldar Akhundzada">}}
 {{</gallery>}}
 

content/ss24/bachelor/b4-sempy/features.md

@@ -1,11 +1,14 @@
 +++
 title = "Features"
 weight = 10
-draft = true
+draft = false
 +++
 
 {{<section title="Features">}}
+Our SEMPY Button offers a relaxed shopping experience that provides customers with the best recommendations based on the information from the three major platforms: idealo, Amazon, and YouTube. Whether the customer is interested in a Victorinox Huntsman, an MSR Hubba Hubba 2 NX tent, or Apple AirPods 3, the SEMPY-AI Tool caters to the customer's individual needs.
 
-Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nulla facilisis neque id vulputate malesuada. Quisque dignissim finibus urna sed sagittis. 
+Due to its crawling capabilities, the SEMPY-AI Tool analyzes customer reviews, pros and cons, Q&A, ratings, and YouTube transcripts from Idealo, Amazon, and YouTube. It then generates a text with all the important information from these sources in an overview. It emphasizes the key points mentioned by customers, helping prospective buyers make informed purchasing decisions. Furthermore, the SEMPY-AI Tool provides advice on which product is best suited for a person. 
+
+<-Video->
 
 {{</section>}}

content/ss24/bachelor/b4-sempy/process.md

@@ -1,16 +1,35 @@
 +++
 title = "Process"
 weight = 10
-draft = true
+draft = false
 +++
 
 {{<section title="Process">}}
-<!-- We startetd the project by brainstorming and noting down our ideas on a Miro Board. We had an idea, to do a chatbot like "Perplexity" and let the customers ask there bot about the products. Instead we decided to have the AI-Tool on the productpage of idealp and output the information for the customer. We wanted the tool to include pros and cons as well as a summary of the customer reviews to the product they are looking up. 
+At the beginning of the project, there was a brainstorming phase where we, as a group, wanted to define how our project should look at the end of the semester. We had ideas like Perplexity AI, ChatGPT, and similar AI tools. Initially, we thought of having a tool where customers could have a conversation on the Idealo website and ask everything about the product.
 
------Bild----
+<-Text1->
+<-picture of our first idea-> 
 
-To work efficiently we defined tasks we wanted to do for the week and than split up in two groups. To give our project a structure we made a road map with a few milestones.We started out by getting familiar with Python and the Git workflow.After that we started building a class that takes information from the idealo page and pass the information from the site to an class that uses opaenAI's GPT to summarize the information. Furthermore we created a main class that combines the ai class and idealo class to see our first result. Following we  -->
+However, the problem was that we had a monthly budget, and if our project were to be used in the future, the monthly budget would explode. After a long discussion, we concluded that we would generate the output of the OpenAI API and place it on the website. After we had a rough idea how our project should look, we started to plan on which language we will program and which tools we goanna use. Our supervisors recommended us to use GitHub for the project because it easier to program as a group on GitHub. Following a lengthy discussion, we decided to use the programming language “Python” because it is easier to work with the OpenAI API. Following a long discussion, we decided to divide our group into two teams. One team was responsible for the AI-Class, and the other for parsing information from the Idealo site. This arrangement worked really well, and we soon had an AI-Class capable of responding to our questions. Additionally, we were able to parse the information from the Idealo site and save it as a JSON file. We connected the AI-Class and the IdealoParser-Class with the main program, and we already saw our 
+first results.
+<-Text2->
+We gave the API the prompt to read the JSON file and write a small summary, including pros and cons, customer reviews, target audience, and some facts about the product. We had a dummy HTML where we could test our output. Initially, we had only one box containing all the outputs in a single JSON. However, our supervisor advised us to create a separate JSON for each section to organize them better. 
 
+During the middle of the semester, we decided to crawl information from different platforms like YouTube and Amazon. Again, we had two teams responsible for this task. For YouTube, we utilized the YouTube API from Google. With our code, we were able to search for videos on YouTube and download their transcripts. This process was relatively straightforward, but to ensure quality reviews and realistic transcripts, we filtered videos based on likes, views, subscribers, and manually created transcripts. After working on the YouTube Transcript class for a few weeks, we successfully downloaded realistic transcripts from the videos and saved them as JSON files. We then fed this data into the API. Meanwhile, the other team managed to crawl customer reviews from Amazon and saved them as a JSON. 
 
+Now we had an idea of how we wanted our product to look and how we would present it at the IMI-Showtime. We considered downloading the HTML code from the idealo website and using it as a blueprint. However, the code was 17,000-18,000 lines long, which would have been a lot of work to manage. Our supervisors suggested having weekly meetings with various employees from idealo who could provide valuable input for our project. After one such meeting, an idealo employee suggested using a Chrome extension instead of downloading thousands of lines of code and connecting them all. We took care of the Chrome extension and were able to see our output on it.
+
+<-Text3->
+
+We were really happy to see our first result, which was a huge achievement for all of us. Of course, the input wasn't as good as we expected, but it was still a milestone for us. After presenting this to our supervisors, they gave us a design tip. Instead of printing the output on the website, we should have a button that opens a modal window with the output. We finally had a fixed goal for our project and a clear vision 
+of how it should look
+<-Text4->
+
+We worked on that idea and improved our prompts, resulting in this final output.
+<-Text5->
+
+If you click on the button, it will generate information about this product. We made two versions of our project. Since generating the information takes time, we created one version where the generated texts are already available and just open in the modal, and another version that generates the texts live. The advantage of pre-generating is that we save a lot of costs, and the customer can get the information immediately.
+
+<-Text5->
 
 {{</section>}}

doc/update.md

@@ -9,6 +9,15 @@ in branches, see [pull_request.md](./pull_request.md)
 2. update your fork and your local clone
 Remember, both are just git repositories. You can start in either one.
 
+
+## Further Reading
+
+We are using the Forking Workflow that is common to contribute to open source repos.
+
+- Atlassian Git Tutorial: [Forking Workflow](https://www.atlassian.com/git/tutorials/comparing-workflows/forking-workflow)
+- GitHub Docs: [Collaborating with pull requests](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests)
+
+## Updating your Fork & Clone
 ###   a. start in fork (on github)
 
 1. Go to the main page of your fork.