Semantic Book Recommender

An intelligent book recommendation system that uses semantic search and emotional analysis to suggest books based on user preferences.

Features

• Semantic Search: Find books using natural language descriptions
• Category Filtering: Filter recommendations by book categories
• Emotional Tone Analysis: Get recommendations based on emotional content (Happy, Surprising, Angry, Suspenseful, Sad)
• Interactive Dashboard: User-friendly Gradio interface
• AI-Powered: Uses OpenAI embeddings and LangChain for semantic understanding

Data

The system uses a dataset of 7,000+ books with metadata including:

• Book titles, authors, and descriptions
• Categories and genres
• Publication years and ratings
• Emotional analysis scores
• Book covers and thumbnails

Setup

1. Clone the repository

   git clone https://github.com/KatyaMarush/book_recommender.git
   cd book-recommender

2. Create virtual environment

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

3. Install dependencies

   pip install -r requirements.txt

4. Prepare Data Run the Jupyter notebooks in the correct order to prepare the required data files:

   # 1. Data exploration and initial processing
   jupyter notebook data-exploration.ipynb
   
   # 2. Vector search preparation
   jupyter notebook vector-search.ipynb
   
   # 3. Text classification for categories
   jupyter notebook text-classification.ipynb
   
   # 4. Sentiment analysis for emotions
   jupyter notebook sentiment-analysis.ipynb

   This will create the necessary processed files: books_cleaned.csv, books_with_categories.csv, books_with_emotions.csv, and tagged_description.txt.

5. Environment Variables Create a .env file with your OpenAI API key:

   OPENAI_API_KEY=your_api_key_here
   

Usage

Web Dashboard

After running all notebooks, launch the interactive Gradio dashboard:

python gradio-dashboard.py

Jupyter Notebooks (Run in Order)

1. data-exploration.ipynb: Initial data analysis and preprocessing
2. vector-search.ipynb: Vector search implementation
3. text-classification.ipynb: Text classification models
4. sentiment-analysis.ipynb: Emotional content analysis

Project Structure

book-recommender/
├── gradio-dashboard.py      # Main web interface
├── requirements.txt         # Python dependencies
├── config.py               # Configuration settings
├── test_recommender.py     # Unit tests
├── cover_placeholder.jpg   # Default book cover
├── *.ipynb                # Jupyter notebooks for analysis
└── data files (created by notebooks):
    ├── books_cleaned.csv
    ├── books_with_categories.csv
    ├── books_with_emotions.csv
    └── tagged_description.txt

Technologies Used

• Python: Core programming language
• Gradio: Web interface framework
• LangChain: AI/ML framework for semantic search
• OpenAI: Embeddings and language models
• Pandas: Data manipulation
• ChromaDB: Vector database for similarity search
• Transformers: NLP models for text analysis

How It Works

1. Data Processing: Books are analyzed for categories and emotional content
2. Semantic Embeddings: Book descriptions are converted to vector embeddings
3. Query Processing: User queries are embedded and compared to book vectors
4. Filtering: Results are filtered by category and emotional tone
5. Ranking: Books are ranked by relevance and emotional scores

Example Usage

1. Run the notebooks in order to prepare data
2. Launch the Gradio dashboard: python gradio-dashboard.py
3. Enter a book description: "A story about forgiveness"
4. Select a category: "Fiction"
5. Choose emotional tone: "Happy"
6. Get personalized book recommendations with covers and descriptions

License

This project is for educational and research purposes.

Contributing

Feel free to submit issues and enhancement requests!