RAGxplorer 🦙🦺

RAGxplorer is an interactive streamlit tool to support the building of Retrieval Augmented Generation (RAG) applications by visualizing document chunks and the queries in the embedding space.

[!NOTE] This is an experimental re-factored version.

Demo 🔎

⚠️ Due to infra limitations, this freely hosted demo may occasionally go down. The best experience is to clone this repo, and run it locally.

Features ✨

• Document Upload: Users can upload PDF documents.
• Chunk Configuration: Options to configure the chunk size and overlap
• Choice of embedding model: all-MiniLM-L6-v2 or text-embedding-ada-002
• Vector Database Creation: Builds a vector database using Chroma
• Query Expansion: Generates sub-questions and hypothetical answers to enhance the retrieval process.
• Interactive Visualization: Utilizes Plotly to visualise the chunks.

Local Installation ⚙️

To run RAGxplorer, ensure you have Python installed, and then install the necessary dependencies:

pip install -r requirements-local-deployment.txt

[!TIP] ⚠️ Do not use requirements.txt. That is so the free streamlit deployment can run. That file includes an additional pysqlite3-binary dependency.

⚠️ If it helps with troubleshooting, this application was built using Python 3.11

Usage 🏎️

1. Setup OPENAI_API_KEY (required) and ANYSCALE_API_KEY (if you need anyscale). Copy the .streamlit/secrets.example.toml file to .streamlit/secrets.toml and fill in the values.
2. To start the application, run:

   streamlit run app.py
   

3. You may need to comment out/remove line 5-7 in app.py.

   __import__('pysqlite3')
   import sys
   sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')
   

[!NOTE] This repo is currently linked to the streamlit demo, and these lines were added due to the runtime in the free streamlit deployment env. See here.

Contributing 👋

Contributions to RAGxplorer are welcome. Please read our contributing guidelines (WIP) for details.

License 👀

This project is licensed under the MIT license - see the LICENSE file for details.

Acknowledgments 💙

• DeepLearning.AI and Chroma for the inspiration and code labs in their Advanced Retrival course.
• The Streamlit community for the support and resources.