Semantic Book Recommender System

A semantic book recommendation system that combines content analysis with emotional tone filtering. Built with LangChain, ChromaDB, and Gradio.

Features

• Semantic search using book descriptions
• Emotion-based filtering (Joy, Surprise, Anger, Fear, Sadness)
• Category filtering (Fiction, Nonfiction, etc.)
• Interactive Gradio interface
• Docker container support
• Data processing pipeline from raw to processed data

Project Structure

├── data/
│ ├── raw/ # Original dataset
│ │ └── books.csv # Raw book data from 02/04/2020
│ ├── cleaned/ # Processed intermediate data
│ │ └── books_cleaned.csv # Cleaned version of raw data
│ └── processed/ # Final processed data
│ ├── books_with_categories.csv # Category-enriched data
│ ├── books_with_emotions.csv # Emotion-analyzed data
│ └── tagged_description.txt # Processed descriptions
├── notebooks/
│ ├── data-exploration.ipynb # Initial data analysis
│ ├── sentiment_analysis.ipynb # Emotion classification
│ ├── text_classification.ipynb # Category classification
│ └── vector-search.ipynb # Embedding experiments
├── app.py # Main application
├── Dockerfile # Container configuration
├── requirements.txt # Python dependencies
└── LICENSE # Apache 2.0 License

Install dependencies:

pip install -r requirements.txt

Docker Setup

docker build -t book-recommender .
docker run -p 7860:7860 book-recommender

Usage

1. Run the application:

python app.py

2. Access the interface at http://localhost:7860
3. Input parameters:
   • Book description (natural language)
   • Category filter (All/Fiction/Nonfiction/etc.)
   • Emotional tone filter (All/Happy/Surprising/etc.)

Data Pipeline

• Raw Data: Original books.csv (4.1MB, 02/2020)
• Cleaning:
  • Remove duplicates
  • Handle missing values
  • Standardize formats
• Processing:
  • Category classification
  • Emotion analysis
  • Description tagging
• Final Datasets:
  • books_with_emotions.csv (7.1MB)
  • tagged_description.txt (2.6MB)

Development Notebooks

• data-exploration.ipynb: Initial dataset analysis
• sentiment_analysis.ipynb: Emotion classification development
• text_classification.ipynb: Category tagging implementation
• vector-search.ipynb: Embedding and similarity search experiments

Result

License

Apache 2.0 License - See LICENSE for details.