ragl

Overview

ragl is a local-first RAG pipeline using ChromaDB for storage and Ollama for LLM inference. It supports document ingestion, chunking, vector storage, and querying via a FastAPI interface.

Features

• Document ingestion with metadata
• Vector search using ChromaDB
• Flexible embeddings (Ollama, SentenceTransformers, or custom models)
• FastAPI interface for querying
• Dockerized for ease

Setup & Usage

Clone the Repository

git clone https://github.com/jth500/ragl.git
cd ragl

Running with Docker

Ensure Docker is installed, then build and run the container:

docker-compose up --build -d

Ingesting data (CSV for now)

1. Copy CSV files into the container:

   docker cp my_data.csv ragl:/app/data/my_data.csv

2. Enter the Docker container shell:

   docker exec -it ragl /bin/sh

3. Run the ingestion script:

   python -m ragl.scripts.ingest_csv /app/data text_col_name /app/chroma_db 

Querying the API

curl -X POST "http://127.0.0.1:8000/ask" \
     -H "Content-Type: application/json" \
     -d '{"query": "What do we think of Tottenham?", "top_k": 3, "model": "llama3.2"}'

Roadmap

• Support for additional document formats (PDF, JSON, etc.)
• Cleaner data ingestion