README.md

metadata

title: PaperIndex
emoji: 📚
colorFrom: blue
colorTo: purple
sdk: docker
sdk_version: latest
app_file: app.py
pinned: false

Paper Index - AI Paper Evaluation System

A comprehensive system for evaluating AI research papers using advanced language models.

Features

• Daily Paper Crawling: Automatically fetches papers from Hugging Face daily
• AI Evaluation: Uses Claude Sonnet to evaluate papers across multiple dimensions
• Interactive Dashboard: Beautiful web interface for browsing and evaluating papers
• Database Storage: Persistent storage of papers and evaluations
• Smart Navigation: Intelligent date navigation with fallback mechanisms

Hugging Face Spaces Deployment

This application is configured for deployment on Hugging Face Spaces.

Configuration

• Port: 7860 (Hugging Face Spaces standard)
• Health Check: /api/health endpoint
• Docker: Optimized Dockerfile for containerized deployment

Deployment Steps

1. Fork/Clone this repository to your Hugging Face account
2. Create a new Space on Hugging Face
3. Select Docker as the SDK
4. Set Environment Variables:
   • ANTHROPIC_API_KEY: Your Anthropic API key for Claude access
5. Deploy: The Space will automatically build and deploy

Environment Variables

ANTHROPIC_API_KEY=your_api_key_here
PORT=7860  # Optional, defaults to 7860

Local Development

Prerequisites

• Python 3.9+
• Anthropic API key

Installation

1. Clone the repository:

   git clone <repository-url>
   cd paperindex
   

2. Install dependencies:

   pip install -r requirements.txt
   

3. Set environment variables:

   export ANTHROPIC_API_KEY=your_api_key_here
   

4. Run the application:

   python app.py
   

5. Access the application:

   • Main interface: http://localhost:7860
   • API documentation: http://localhost:7860/docs

API Endpoints

Core Endpoints

• GET /api/daily - Get daily papers with smart navigation
• GET /api/paper/{paper_id} - Get paper details
• GET /api/eval/{paper_id} - Get paper evaluation
• GET /api/health - Health check endpoint

Evaluation Endpoints

• POST /api/papers/evaluate/{arxiv_id} - Start paper evaluation
• GET /api/papers/evaluate/{arxiv_id}/status - Get evaluation status

Cache Management

• GET /api/cache/status - Get cache statistics
• POST /api/cache/clear - Clear all cached data
• POST /api/cache/refresh/{date} - Refresh cache for specific date

Architecture

Frontend

• HTML/CSS/JavaScript: Modern, responsive interface
• Real-time Updates: Dynamic content loading
• Theme Support: Light/dark mode toggle

Backend

• FastAPI: High-performance web framework
• SQLite: Lightweight database for paper storage
• Async Processing: Background evaluation tasks
• Caching: Intelligent caching system for performance

AI Integration

• Claude Sonnet: Advanced paper evaluation
• Multi-dimensional Analysis: Comprehensive evaluation criteria
• Structured Output: JSON-based evaluation results

Database Schema

Papers Table

CREATE TABLE papers (
    arxiv_id TEXT PRIMARY KEY,
    title TEXT NOT NULL,
    authors TEXT NOT NULL,
    abstract TEXT,
    categories TEXT,
    published_date TEXT,
    evaluation_content TEXT,
    evaluation_score REAL,
    overall_score REAL,
    evaluation_tags TEXT,
    evaluation_status TEXT DEFAULT 'not_started',
    is_evaluated BOOLEAN DEFAULT FALSE,
    evaluation_date TIMESTAMP,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
    updated_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);

Evaluation Dimensions

The system evaluates papers across 12 key dimensions:

1. Task Formalization - Clarity of problem definition
2. Data & Resource Availability - Access to required data
3. Input-Output Complexity - Complexity of inputs/outputs
4. Real-World Interaction - Practical applicability
5. Existing AI Coverage - Current AI capabilities
6. Automation Barriers - Technical challenges
7. Human Originality - Creative contribution
8. Safety & Ethics - Responsible AI considerations
9. Societal/Economic Impact - Broader implications
10. Technical Maturity Needed - Development requirements
11. 3-Year Feasibility - Short-term potential
12. Overall Automatability - Comprehensive assessment

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

License

This project is licensed under the MIT License - see the LICENSE file for details.