Skip to content

Latest commit

 

History

History
88 lines (63 loc) · 3.03 KB

README.md

File metadata and controls

88 lines (63 loc) · 3.03 KB
Raw

News-RSS Project

Overview

The News-RSS project is a sophisticated tool designed to scrape and parse RSS feeds, leveraging the power of Large Language Models (LLMs) to extract valuable artifacts from articles. This includes information such as content placement, publication dates, author names, and other useful attributes.

arch-schema.png

There is a part of crawler platform that collect's scrapped news to storage (postgres) or queue (rabbitmq) with following pre-processing steps:

  • Extract rss item with content and source URL of article;
    • Check source URL into cache - does article has been already parsed?
    • If content is not empty replaced any trash tags from text content;
    • If content is empty send request to LLM to extract content from HTML document;
  • Send extracted data to storage/queue.

Features

  • RSS feed scraping and parsing
  • LLM-powered content analysis
  • Storing parsed data to storage/queue
  • Docker-based deployment for easy setup and scalability
  • Compile with feature to enable:
    • native/llm crawler;
    • local/redis caching;
    • rabbitmq/postgres storage.

Quick Start

  1. Clone the repository:

    git clone http://<gitlab-domain-address>/data-lake/news-rss.git
cd news-rss

  2. Build docker image from sources with passed features:

    To build with default features (local caching, native crawler, rabbitmq):

    docker build -t new-rss:latest .

    To build with features add parameter --build-arg FEATURES='--features cache-redis'

    For example to build with redis and llm crawler use this command:

    docker build --build-arg FEATURES='--features cache-redis --feature crawler-llm' -t new-rss:latest .

    There are following available features:

    • cache-redis Use remote redis cache service;
    • publish-offline Store scraped and parsed feeds content to postgres instead rabbitmq;
    • crawler-llm Use llm to scrape and parse html data of source article instead native crawler.

  3. Edit or create a new .env file in the project root and add your configuration:

    POSTGRES_USER=postgres
POSTGRES_PASSWORD=postgres

REDIS_USERNAME=redis
REDIS_PASSWORD=redis
REDIS_USER_PASSWORD=redis

  4. Edit configs file config/production.toml to launch docker compose services:

    The main parameter into topics.rss.target_url - you may pass multiple rss URL sources by column like this target_url=https://feeds.feedburner.com/ndtvnews-world-news,https://feeds.skynews.com/feeds/rss/world.xml

  5. Start the application using Docker Compose:

    docker compose --env-file .env up -d news-rss <other-needed-services>

    Into <other-needed-services> pass services names needed for current features. For example if you build project docker image with default features you have not to use postgres and redis services.

  6. The application should now be running. Check the logs with:

    docker compose logs -f