Clickbait-and-Switch-AI

Fullstack service designed to analyze news articles from URLs. It leverages various NLP models to provide insights such as summarization, sentiment analysis, bias classification, etc.

Comment

Clickbait & Switch AI

Clickbait & Switch AI is a fullstack service designed to analyse news articles from URLs. It leverages various Natural Language Processing (NLP) models to provide insights such as summarization, sentiment analysis, bias classification, topic identification, and named entity recognition.

Demo: https://drive.google.com/file/d/1vZyMiMuTljLZm5mR-yb-wBmrmMH6-FMh/view?usp=sharing

Features

  • Article Fetching & Parsing: Downloads and extracts content from a given news article URL.
  • Text Summarization: Generates a concise summary of the article content.
  • Named Entity Recognition (NER): Identifies and extracts names of people mentioned in the article.
  • Sentiment Analysis: Determines the overall sentiment (e.g., positive, negative, neutral) of the article.
  • Bias Classification: Classifies the article's text based on potential bias (e.g., objective, biased).
  • Topic Classification: Identifies the main topics of the article (e.g., politics, technology, health).

Technology Stack

Backend

  • Python 3.x
  • FastAPI: For building the asynchronous API.
  • Uvicorn: ASGI server for FastAPI.
  • Hugging Face Transformers: For state-of-the-art NLP models:
    • Summarization: facebook/bart-large-cnn
    • Zero-Shot Classification (for bias & topic): facebook/bart-large-mnli
    • Sentiment Analysis: Default pipeline model
  • spaCy: For Named Entity Recognition (en_core_web_sm).
  • Newspaper3k: For article downloading and parsing.
  • Pytest: For running automated tests.

Frontend

  • Language: JavaScript / TypeScript
  • Framework/Library: React
  • Build Tool: Vite
  • Styling: Material UI
  • API Client: Axios
  • Test: Vitest, Reacting Testing Library

Project Structure

Clickbait-and-Switch-AI/
├── backend/
│   ├── app/
│   │   ├── routes/
│   │   │   └── analyse.py  # API endpoint definitions
│   │   └── __init__.py
│   ├── services/
│   │   └── nlp.py          # Core NLP processing logic
│   ├── tests/
│   │   ├── conftest.py     # Pytest fixtures
│   │   ├── test_api.py     # API endpoint tests
│   │   └── test_nlp_service.py # NLP service tests
│   ├── main.py             # FastAPI application entry point
│   ├── requirements.txt    # (Assumed) Python dependencies
│   └── ...                 # Other Python files
├── frontend/               # React + Vite Frontend application
│   ├── public/             # Static assets
│   ├── src/
│   │   ├── assets/         # Images, fonts, etc.
│   │   ├── components/     # Reusable React components
│   │   ├── pages/          # Page-level components
│   │   ├── services/       # API service calls (e.g., Axios instances)
│   │   ├── styles/         # Global styles, theme configurations (e.g., for Material UI)
│   │   ├── App.jsx         # Main App component
│   │   └── main.jsx        # Entry point for React application
│   ├── tests/              # Vitest/React Testing Library tests
│   ├── index.html          # Main HTML file for Vite
│   ├── package.json        # Project metadata and dependencies
│   ├── vite.config.js      # Vite configuration
│   └── ...                 # Other frontend files (e.g., .eslintrc.cjs, .prettierrc)
├── .gitignore              # Git ignore rules
└── README.md               # This file

Setup and Installation

Prerequisites

  • Python 3.8+
  • pip (Python package installer)
  • Node.js and npm or yarn (if a frontend is used)

Backend Setup

  1. Clone the repository:

    git clone <your-repository-url>
cd Clickbait-and-Switch-AI/backend

  2. Create and activate a virtual environment:

    python -m venv venv
# On Windows
venv\Scripts\activate
# On macOS/Linux
source venv/bin/activate

  3. Install Python dependencies: (Assuming you have a requirements.txt file in the backend directory)

    pip install -r requirements.txt

    Note: The first time you run the application, Hugging Face Transformers and spaCy may download the required NLP models. This might take some time and require an internet connection.

Frontend Setup

  1. Navigate to the frontend directory (if it exists):

    cd ../frontend

  2. Install frontend dependencies:

    npm install
# or
yarn install

Running the Application

Backend

From the Clickbait-and-Switch-AI/backend directory:

python main.py

The API will be accessible at http://localhost:8000.

Frontend

From the Clickbait-and-Switch-AI/frontend directory (if it exists):

npm run dev
# or
yarn dev

The frontend will typically be accessible at http://localhost:5173.

API Usage

The primary endpoint for analysis is:

  • Endpoint: POST /analyse/
  • Request Body (JSON): json { "url": "https://www.example-news.com/article-link" }
  • Response: A JSON object containing the analysis results (title, summary, named people, sentiment, bias classification, topic classification).

Running Tests (Backend)

From the Clickbait-and-Switch-AI/backend directory:

pytest

Running Tests (Frontend)

From the Clickbait-and-Switch-AI/frontend directory:

npm test
Share this project:

Updates