README.md

metadata

title: ChatFed Re-Ranker Service
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
license: mit

ReRanker Module

This is an LLM-based generation service designed to be deployed as a modular component of a broader RAG system. The service runs on a docker container and exposes a gradio UI on port 7860 as well as an MCP endpoint.

Configuration

1. The module requires an API key (set as an environment variable) for a model provider to run. Make sure to set the appropriate environment variables:

• HuggingFace: HF_TOKEN

2. Inference provider and model settings are accessible via params.cfg

MCP Endpoint

Available Tools

rerank_context

Re-ranks a list of context dicts (each with 'page_content' & 'metadata') using a cross-encoder and returns the top_n sorted results.

Input Schema:

Parameter	Type	Description
query	string	The search query to rank contexts against
contexts	string	List of context dictionaries to be re-ranked

Example Usage:

{
  "query": "your search query here",
  "contexts": "your context data here"
}

---

This tool uses a cross-encoder model to improve the relevance ranking of retrieved contexts based on their similarity to the input query.