news-summarizer

Description

This is a tool that aggregates news articles relevant to a certain topic and summarizes it, presenting the key points from each article.

Installation

You can install and run this project using Docker or setting up locally using uv as package manager.

Method 1: Docker (Recommended for convenience)

1. Install Docker
   Install Docker from here. Be sure to update and get the latest version.

2. Clone the Repository

   git clone https://github.com/limjoobin/news-summarizer.git
   

3. Build and run the container using docker compose

   docker compose up --build
   

   This will build the images and run the container according to the specifications in compose.yaml. When ready, the application will be running on http://localhost:8000.

4. Stop the application
   To stop the application (stop and remove all containers and network, and clean up the environment), do

   docker compose down
   

Method 2: Running locally

1. Install uv
   Install uv from here. We will use uv as package and project manager.

2. Clone the Repository

   git clone https://github.com/limjoobin/news-summarizer.git
   

3. Run the news-scraper service

   cd news-scraper/src
   uv run news_scraper.py
   

   If successful, the news-scraper service should be running at http://localhost:8001.

4. Run the vllm server
   Run the vllm server to serve the bart-large-cnn model for summarization. Be sure to install vllm dependencies beforehand.

   cd summarizer
   uv add vllm
   uv run vllm serve facebook/bart-large-cnn --host 127.0.0.1 --port 8003
   

   Note: The vllm server must be run on a Linux environment. If using Windows, run it from WSL2. If successful, the vllm server should be running at http://localhost:8003

5. Run the summarizer service

   cd summarizer/src
   uv run main.py
   

   If successful, the summarizer service should be running at http://localhost:8002

6. Run the app frontend

   cd app
   uv run app.py
   

   If successful, the app frontend should be running at http://localhost:7860.

Usage

Visit the app frontend to use the service. Do note that there is a delay for each summary generated from the scraping and summarization.