rag-reranker

A lightweight cross-encoder reranker for LangChain Document objects.

Installation

pip install rag-reranker

Usage

from langchain_core.documents import Document
from rag_reranker import CrossEncoderReranker

docs = [
    Document(page_content="Paris is the capital of France."),
    Document(page_content="Berlin is the capital of Germany."),
]

reranker = CrossEncoderReranker()
results = reranker.rerank("what is the capital of france?", docs, top_k=1)

print(results[0].document.page_content, results[0].score)

Development

Build locally:

python -m pip install --upgrade build
python -m build

This produces distributable files in the dist/ folder.

Publish to PyPI

python -m pip install --upgrade twine
python -m twine check dist/*
python -m twine upload dist/*

You can use TestPyPI first:

python -m twine upload --repository testpypi dist/*