Summary

This repository offers a personalized paper recommendation system for OpenClaw/Feishu, powered by AMiner, arXiv, and LLMs. It targets researchers and users seeking efficient academic paper discovery, providing a natural-language-first interface with scholar-aware cold starts and unified profile building for enhanced relevance.

How It Works

The system accepts scholar or topic-based inputs, unifying them into a ResearchProfile. It then performs retrieval, AMiner enrichment, summarization, and card rendering. The core novelty lies in its natural-language-first approach, scholar-aware cold starts leveraging AMiner signals, and graceful degradation when optional internal components are unavailable, ensuring robustness.

Quick Start & Requirements

• Install: pip install -r requirements.txt within a Python virtual environment.
• Configure: Copy config.example.yaml to config.yaml, setting aminer.token and LLM API details. Environment variables can also be used.
• Run: Execute via python3 scripts/handle_trigger.py --base-dir . --config config.yaml --text "<query>".
• Prerequisites: Python 3.x, AMiner API token, OpenAI-compatible LLM API key/endpoint.
• OpenClaw Install: Copy repo to ~/.openclaw/skills/aminer-rec5.

Highlighted Details

• Natural-language-first paper recommendation.
• Scholar-aware cold start from AMiner person signals.
• Unified topic and scholar profile building.
• arXiv retrieval plus AMiner enrichment.
• Structured summaries and recommendation reasons.

Maintenance & Community

This is a public-shareable cut of an internal system. No specific community links or contributor details are provided. A Star History link is available.

Licensing & Compatibility

• License: MIT.
• Compatibility: Permissive for commercial use and closed-source linking.

Limitations & Caveats

This public version omits internal secrets and local artifacts. Functionality dependent on optional internal hooks will degrade gracefully. The sole supported external interface is scripts/handle_trigger.py; other scripts are internal and subject to change.