DRAG, a Document Search API

A lightweight, cost-effective Retrieval-Augmented Generation (RAG) API built for efficient document search and context-aware LLM responses.

Tech Stack

• Framework: FastAPI
• Orchestration: LlamaIndex
• Database: PostgreSQL with pgvector extension
• Dependency Management: uv
• Tests: Pytest

Architecture

Local Development & Setup

1. Clone the repository:

   git clone https://github.com/dalentri/drag.git
   cd drag

2. Install dependencies using uv:

   uv sync

3. Configure environment variables: Create a .env file in the root directory and add your credentials:

   OPEN_API_KEY=your_open_api_key_here
   DATABASE_URL=postgresql://user:password@localhost:5432/dbname
   MODEL=model_name

4. Add custom documents to data/raw_docs directory (optional)

5. Run the development server:

   uv run fastapi dev app/main.py

Running Evaluations

To begin running evaluations, ensure that the backend is running using the uv run fastapi dev app/main.py command.

To then run the evaluations:

PYTHONPATH=. uv run evals/run_evals.py

Running Tests

Drag comes equipped with unit tests for the API's functionality.

To run tests with Pytest:

pytest tests/test_api.py