https://reviewreward-production.up.railway.app

If you like my work, consider buying me a coffee! ☕️

Review Authenticity Checker API

This project is a Flask-based API designed to check the authenticity of text reviews. It uses machine learning models to determine if a review is genuine or fake. The API supports both manual text review and automated checks using Cohere's language model.

Features

• Text Review Authenticity Check: Determines if a given text review is real, fake, or computer-generated.
• Machine Learning Models: Utilizes Naive Bayes and SVM models trained on a dataset of deceptive and truthful reviews.
• Cohere Language Model Integration: Checks review authenticity using Cohere's language model for an additional layer of verification.

Getting Started

Prerequisites

• Python 3.11.5
• Flask
• Vercel CLI (for deployment)
• AWS S3 (optional, for remote model storage)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/review-authenticity-checker.git
   cd review-authenticity-checker

2. Create a virtual environment:

   python -m venv .venv
   source .venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Download NLTK stopwords:

   python -m nltk.downloader stopwords

Configuration

1. Cohere API Key: Set up your Cohere API key as an environment variable to keep it secure:

   export COHERE_API_KEY="your-cohere-api-key"

2. AWS S3 Configuration (Optional): If you store your model files in AWS S3, configure your AWS credentials.

Usage

1. Run the Flask app:

   flask run

2. API Endpoint:

   The API has one main endpoint:

   • /api/review (POST): Check the authenticity of a text review.

     Request:

     {
       "productName": "Product Name",
       "category": "Category",
       "brand": "Brand",
       "purchaseDate": "2023-08-01",
       "purchasePrice": "100",
       "productReview": "This is my review text.",
       "shoppingLink": "http://shoppinglink.com"
     }

     Response:

     {
       "message": "Original"  // or "Fake"
     }

Deployment

1. Create .vercelignore file:

   .venv
   

2. Generate requirements.txt:

   pip freeze > requirements.txt

3. Deploy to Vercel:

   vercel

   Follow the prompts to complete the deployment process.

Project Structure

your-project/
├── .vercelignore
├── requirements.txt
├── app.py
├── templates/
├── static/
└── ...

Acknowledgements

• Cohere for their language model API.
• NLTK for natural language processing tools.