RAG API with FastAPI and Docker

This project is a complete, containerized prototype of a Retrieval-Augmented Generation (RAG) application. It uses an LLM to answer questions based on a custom knowledge base, served via a high-performance FastAPI backend.

Features

• Retrieval-Augmented Generation (RAG): Provides answers to user questions grounded in a specific set of documents, reducing hallucinations and providing context-aware responses.

• FastAPI Backend: A modern, asynchronous, and high-performance API to serve the RAG chain.

• Dockerized Environment: The entire application is containerized with Docker and orchestrated with Docker Compose for easy, consistent, and reproducible setup.

• Scalable & Deployable: Built with deployment in mind, ready to be pushed to any cloud service that supports containers.

Tech Stack

• Backend: FastAPI

• ASGI Server: Uvicorn

• AI Framework: LangChain

• LLM Provider: OpenAI

• Vector Store: ChromaDB

• Containerization: Docker & Docker Compose

Getting Started

Follow these instructions to get the project up and running on your local machine.

Prerequisites

• Docker and Docker Compose installed on your system.

• An OpenAI API Key.

Installation & Setup

1. Clone the repository:

   git clone <your-repository-url>  
   cd <your-project-directory>
   

2. Create the environment file:

   Create a file named .env in the root of the project directory and add your OpenAI API key:

   OPENAI_API_KEY=sk-YourSecretApiKeyHere
   API_KEY=secretkey
   

3. Add your data:

   Place the PDF or text files you want to use as your knowledge base inside the /data directory.

4. Build the knowledge base: Run the ingestion script. This will process your documents and create the local vector store in a ./chroma_db directory.

   docker-compose run --rm --build rag-app python ingest.py
   

   Note: The --rm flag automatically removes the container after the script finishes.

5. Run the application: Start the FastAPI application using Docker Compose.

   docker-compose up -d
   

   The -d flag runs the container in detached mode. The application will be available at http://localhost:8000.

API Usage

You can interact with the API through its documentation, which is automatically generated by FastAPI.

• Interactive Docs (Swagger): http://localhost:8000/docs

• Alternative Docs (ReDoc): http://localhost:8000/redoc

Example Request with curl

Here is an example of how to send a query to the /generate endpoint from your terminal:

curl -X 'POST' \
  'http://localhost:8000/generate' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -H 'x-api-key: secretkey' \
  -d '{
  "query": "What is the best way to handle path parameters in FastAPI?"
}'

Deployment

This application has been successfully deployed using Azure Container Apps and is live at the following URL:

➡️ Live Application URL: [RAG-APP]

Usage

Use the RAG app by clicking on the generate endpoint and then the "try it out" button. Then enter your question and the "x-api-key" which is just "secretkey"

Data

The data that has been ingested are the following papers:

• Attention Is All You Need
• Language Models are Few-Shot Learners
• Denoising Diffusion Probabilistic Models
• High-Resolution Image Synthesis with Latent Diffusion Models
• Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• Neural Networks are Decision Trees
• Segment Anything

Project Structure

/
|-- .env                # Environment variables (OpenAI API Key)
|-- .gitignore          # Files to ignore for git
|-- Dockerfile          # Instructions to build the application container
|-- docker-compose.yml  # Defines the services for local development
|-- ingest.py           # Script to process data and build the vector store
|-- requirements.txt    # Python dependencies
|-- README.md           # This file
|
|-- /app/               # Main application source code
|   |-- main.py         # FastAPI application and endpoints
|   |-- rag_logic.py    # RAG chain creation and logic
|
|-- /data/              # Source documents for the knowledge base 
|   |-- /chroma_db/         # Persisted ChromaDB vector store (created by ingest.py)
|   |-- README.md