This Python FastAPI project integrates LangChain for implementing the RAG (Retrieval Augmented Generation) pattern, leveraging a CosmosDB vector store for text retrieval and an Azure Storage Account for image retrieval. Please view the tutorial walk-through: LangChain RAG with React, FastAPI, Cosmos DB Vector Search: Part 2.
- Utilizes FastAPI for building a high-performance web API.
- Integrates LangChain for implementing the RAG pattern, enabling advanced natural language processing tasks.
- Utilizes CosmosDB vector store for efficient text retrieval based on semantic similarity.
- Retrieves images from an Azure Storage Account associated with the retrieved text.
- Python environment with FastAPI, LangChain, and Azure SDK installed.
- Access to Azure subscription for deploying CosmosDB and Azure Storage Account.
- Basic knowledge of FastAPI, LangChain, CosmosDB, and Azure Storage concepts.
- Clone the repository to your local machine.
- Set up CosmosDB for MongoDB VCore Vector Store and Azure Storage Account in your Azure subscription.
- Create .env file and populate:
- VECTORDB_ENDPOINT=''
- VECTORDB_API_KEY=''
- OPENAI_API_KEY=''
- MONGO_CONNECTION_STRING=''
- AZURE_STORAGE_CONNECTION_STRING=''
- AZURE_STORAGE_CONTAINER="images"
- Create pythonn env:
python -m venv venv
- Install Requirements:
venv\Scripts\activate
python -m pip install -r requirements.txt
- Run the FastAPI server:
python main.py
- View the Swagger Docs: http://127.0.0.1:8000/docs
- Tryout quick searh:
This project is licensed under the MIT License, granting permission for commercial and non-commercial use with proper attribution.
For any questions or issues, please open an issue on GitHub or reach out to the project maintainers.
This project is provided for educational and demonstration purposes only. Use at your own risk.