README.md

metadata

title: Eshop Chat
emoji: 🦀
colorFrom: indigo
colorTo: gray
sdk: docker
pinned: false

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

How to run on your localhost? (For MAC users)

• Install Docker for MAC app to make docer enginee available to vscode
• Install vscode
• Open the project in vscode
• Create .env file only on local machine (DO NOT ADD AND PUSH IT TO GIT)
• Add openai api key to .env file: Eg. OPENAI_API_KEY=sk-WSISHD232UDDUDJHHQQTT
• Type CMD+SHIFT+P and select Dev Containers: Rebuild Container. It will create docker container environment to run the code within the container. You don't have to install any project specific packages on you local machine.
• Use menu Terminal > New Terminal, to open a shell and type following command to run chainlit server
• If everything works, then you can access the application using http://localhost:7860/ . Please note that port 7860 is required by huggingface to make our application accessible using huggingface url (after git push).
• chainlit run app.py --port 7860

BestBuy Datasets

• https://github.com/BestBuyAPIs/open-data-set

This app only uses first 100 product information

• It uses products.json data to create normalized categories
• Google colab is used to explore the data: https://colab.research.google.com/drive/17iPtcrdEHnqcECF6xakGB0HaWG33PzFw?usp=sharing
• It generates product spec using the product and categories features, which is then used to create vector embeddings and save in vector store (Chroma vector store)