Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
68 lines (48 loc) · 2.34 KB

CONTRIBUTING.md

File metadata and controls

68 lines (48 loc) · 2.34 KB
Raw

Contributing to LocalRAG-Kit

Thank you for your interest in contributing to LocalRAG-Kit! We welcome contributions from developers, security researchers, and AI enthusiasts.

🛠️ Development Setup

  1. Clone the Repository:

    git clone https://github.com/RitualDev-Lab/localrag-kit.git
cd localrag-kit

  2. Create and Activate a Virtual Environment:

    # Windows PowerShell
python -m venv .venv
.\.venv\Scripts\Activate.ps1

# Linux / macOS
python3 -m venv .venv
source .venv/bin/activate

  3. Install Dependencies:

    pip install -e ".[dev]"

🧪 Running Tests

Ensure all unit and integration tests pass before submitting code:

pytest -v

🏗️ Architecture Overview

LocalRAG-Kit is structured into clean, decoupled layers:

  • localrag/core/: File harvesting, pathspec ignore parsing, file parsers (Markdown, PDF, Code), and semantic boundary chunkers.
  • localrag/storage/: SQLite engine with SQLite FTS5 for BM25 keyword search, binary BLOB vector storage, NumPy SIMD cosine similarity, and change detection.
  • localrag/providers/: Pluggable embedding providers (Fast Offline Hasher, Ollama, OpenAI) and LLM providers (Ollama, OpenAI, Mock).
  • localrag/retrieval/: Reciprocal Rank Fusion (RRF) hybrid reranking, context synthesizer with token budgeting, and grounded prompt orchestration.
  • localrag/server/: High-performance FastAPI server with Server-Sent Events (SSE) streaming, REST endpoints, and embedded static web UI.
  • localrag/cli.py: Interactive CLI with Rich formatting (scan, index, search, ask, stats, serve).

📝 Contribution Guidelines

  1. Keep it 100% Offline & Zero-Config: Features must never require mandatory external cloud accounts or external database installations (like Chroma, Pinecone, or PostgreSQL). SQLite must remain the zero-infra backbone.
  2. Deterministic Chunker Line Numbers: Every parser and chunker must compute 1-based exact physical start and end line coordinates so citation provenance is verifiable.
  3. Cross-Platform Compatibility: Ensure paths use pathlib.Path and console outputs handle Windows CP1252 / UTF-8 safely.

📜 License

By contributing to LocalRAG-Kit, you agree that your contributions will be licensed under the MIT License.