SignalRank-RAG

SignalRank-RAG is an open-source framework for building and evaluating Retrieval-Augmented Generation (RAG) systems.

The project focuses on:

• Document ingestion
• Chunking strategies
• Embedding pipelines
• Retrieval systems
• Reranking
• Evaluation
• Production deployment

Currently Implemented

• Recursive TXT and Markdown ingestion
• Portable relative source references
• Deterministic, content-sensitive document IDs
• Empty and unsupported-file handling
• 12 automated tests

Development Setup

python -m pip install -e ".[dev]"
python -m pytest -v

Roadmap

• Project scaffold
• Exception handling
• Logging framework
• Configuration management
• Data ingestion
• Chunking
• Embeddings
• Vector stores
• Retrieval
• Reranking
• Evaluation
• API deployment