TruthTracer is a sophisticated news article analysis tool designed to detect potentially misleading content by cross-referencing information with multiple sources.

TruthTracer combines web scraping and natural language processing to provide an objective assessment of news article reliability.

The system works by:

1. Analysing a primary news article's content
2. Finding related articles from other publications
3. Cross-referencing information between sources
4. Identifying potential discrepancies, omissions, or misleading content
5. Providing a detailed report on the article's reliability

• Article Content Extraction: Automatically extracts the meaningful content from news articles, removing ads, navigation, and other noise
• Reference Article Discovery: Finds related articles from other sources covering the same topic
• Cross-Reference Analysis: Compares how different sources report on the same events
• Misleading Content Detection: Identifies potentially misleading statements, omissions, or distortions
• Clean API Interface: RESTful API for easy integration with other systems
• Modern Web Interface: User-friendly frontend for analysing articles

TruthTracer consists of several components:

• Scraping Module: Extracts content from news websites using both static and dynamic scraping techniques
• Processing Module: Cleans, normalises, and analyses text content
• Google Module: Handles search operations to find reference articles
• Utilities: Common utility functions for text processing, URL handling, etc.
• API: FastAPI backend that coordinates the analysis process
• Frontend: Simple, elegant web interface for submitting articles and viewing results

• Python 3.9+
• Ollama
• Chrome/Chromium (for dynamic web scraping)

1. Clone the repository:

   git clone https://github.com/alexjohnyoung/truthtracer.git
   cd truthtracer
   

2. Create and activate 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. Setup ChromeDriver (required for local API deployment):

   • Download the appropriate version of ChromeDriver that matches your Chrome/Chromium version
   • For Windows: Place the chromedriver.exe in your PATH or in the project's root directory
   • For macOS/Linux: Place the chromedriver executable in your PATH or in the project's root directory

uvicorn main:app --host 0.0.0.0 --port 8000

The frontend is served via a web server.

Navigate to the frontend directory and run:

python serve.py

This will start a web server on port 8080 and automatically open your browser.

To use a different port, run with the '--port' argument.

The API provides endpoints for starting an analysis and checking its status:

• GET /analyse-start?url={article_url}&max_references={number} - Start analysing an article
• GET /analyse-status/{analysis_id} - Check analysis status and get results

• src/ - Main source code directory
  • utils/ - Utility functions and helpers
  • scraping/ - Web scraping modules
  • processing/ - Text processing and analysis modules
  • google/ - Google search functionality
• frontend/ - Web interface
• main.py - API entry point

• Big thank you to my supervisors at Munster Technological University for their guidance and support
• All contributors to the open-source projects that make this possible