Skip to content

AK391/dailypapersHN

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 

Repository files navigation

Daily Papers HN

Daily Papers HN is a Python-based web application that displays academic papers in a Hacker News-like interface. It utilizes the Hugging Face Daily Papers API to fetch and present papers in a sortable, paginated list.

Screen Shot 2024-09-22 at 8 02 02 PM

Web Demo

You can try out the Daily Papers HN application online without any local setup:

Daily Papers HN Demo

This demo is hosted on Hugging Face Spaces and provides the full functionality of the application.

Technical Overview

  • Backend: Python 3.7+
  • Web Framework: Gradio
  • API Interaction: Requests library
  • Data Processing: Custom PaperManager class
  • Hosting: Hugging Face Spaces
  • Development Environment: Cursor with Claude 3.5-sonnet AI assistance

Key Components

  1. PaperManager Class: Handles paper fetching, sorting, and rendering

    • Implements custom sorting algorithms: Hot, New, and Rising
    • Manages pagination and paper formatting
  2. API Integration:

    • Endpoint: https://huggingface.co/api/daily_papers
    • Fetches up to 100 papers per request
  3. Sorting Algorithms:

    • Hot: score = upvotes / ((time_diff_hours + 2) ** 1.5)
    • Rising: score = upvotes / (time_diff_hours + 1)
    • New: Based on publishedAt timestamp
  4. Gradio Interface:

    • Utilizes gr.Blocks for layout
    • Implements custom CSS for Hacker News-like styling
    • Responsive design with dark mode support

Core Functionality

  1. Paper Fetching:

    def fetch_papers(self):
        response = requests.get(f"{API_URL}?limit=100")
        self.raw_papers = response.json()
        self.sort_papers()
  2. Sorting:

    def sort_papers(self):
        if self.sort_method == "hot":
            self.papers = sorted(self.raw_papers, key=self.calculate_score, reverse=True)
        # Similar for "new" and "rising"
  3. Rendering:

    def render_papers(self):
        current_papers = self.papers[start:end]
        papers_html = "".join([self.format_paper(paper, idx) for idx, paper in enumerate(current_papers)])
        return f"<table class='itemlist'>{papers_html}</table>"
  4. Pagination:

    def next_page(self):
        if self.current_page < self.total_pages:
            self.current_page += 1
        return self.render_papers()

Setup and Running

  1. Clone the repository:

    git clone https://github.com/your-username/dailypapersHN.git
    cd dailypapersHN
    
  2. Set up a virtual environment:

    python -m venv venv
    source venv/bin/activate  # On Windows: venv\Scripts\activate
    
  3. Install dependencies:

    pip install -r requirements.txt
    
  4. Run the application:

    python app.py
    
  5. Access the web interface at http://127.0.0.1:7860

Development Notes

  • The application uses Gradio's gr.Blocks for a flexible layout.
  • Custom CSS is implemented for styling, including dark mode support.
  • Error handling is in place for API requests and data processing.
  • The PaperManager class is designed for easy extension and modification of sorting algorithms.

Future Enhancements

  • Implement caching to reduce API calls
  • Add unit tests for sorting algorithms and rendering functions
  • Explore asynchronous paper fetching for improved performance

Contributing

Contributions are welcome! Please fork the repository and submit a pull request with your changes.

Deployment

The application is deployed on Hugging Face Spaces, which provides a serverless environment for hosting Gradio apps. To deploy your own version:

  1. Fork this repository to your GitHub account.
  2. Create a new Space on Hugging Face, linking it to your forked repository.
  3. Configure the Space to use the app.py file as the entry point.
  4. Hugging Face Spaces will automatically deploy and update the app based on your repository changes.

Note: This application was developed using Cursor with Claude 3.5-sonnet AI assistance. This README was also generated with the assistance of Claude 3.5-sonnet in Cursor AI.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages