Shakti28/Entertainment-Recommendation-System

🎬 Entertainment Recommendation System

Welcome to the Entertainment Recommendation System — a personalized content discovery tool that suggests Anime, Movies, and Web Series based on your preferences. Built with Streamlit, Sentence-BERT, and FAISS, this system provides fast and accurate recommendations through a clean and intuitive UI.

🚀 Features

• —🔍 Content-based Recommendations for anime, movies, and web series.
• —⚡ Real-time Suggestions with semantic search using Sentence-BERT.
• —🧠 Efficient similarity search via FAISS.
• —🎨 Custom tab bar UI with category-specific background and icons.
• —🧾 Error logging and clean modular codebase for easy debugging and scaling.

📂 Project Structure

Entertainment-Recommendation-System/
│
├── app.py                      # Main Streamlit app
├── Dockerfile                  # Docker build configuration
├── requirements.txt            # Python dependencies
├── .env                        # (Optional) Environment variables
│
├── config/
|       └── global_paths.py     # Global path management
|
├── data/
│   ├── raw_data/
│   │   ├── anime/
│   │   ├── movie/
│   │   └── web_series/
│   └── processed_data/
│       ├── anime/
│       ├── movie/
│       └── web_series/
│
├── artifacts/
│   ├── anime/
│   ├── movie/
│   └── web_series/
│
├── icon/                       # Icons for tab bar
├── image/                      # Background images
├── notebook/
│   └── *.ipynb                 # Preprocessing notebooks
│
└── src/
    ├── recommender/
    │   └── base.py             # Recommendation logic
    └── utils/
        └── logger.py           # Logging utility

📥 Datasets Used

• —Anime & Web Series: Collected from Kaggle public datasets.
• —Movie: MovieLens 32M Dataset

Organized into raw_data/ and processed into processed_data/ for each content type.

🧪 Preprocessing

To prepare your data and generate the model artifacts:

1. 1.Use your own script or notebook to clean and combine metadata into a tags column.
2. 2.Generate Sentence-BERT embeddings and FAISS index.
3. 3.Save the following files in each artifacts/{content_type} folder:
4. 4.data.pkl
5. 5.sbert_model.pkl
6. 6.faiss_index.index

🔧 Setup Instructions

1. Clone the Repository

git clone https://github.com/shakti-sarada/Entertainment-Recommendation-System.git
cd Entertainment-Recommendation-System

2. Install Dependencies

pip install -r requirements.txt

3. Run the App

streamlit run app.py

Visit http://localhost:8501 in your browser.

🐳 Docker Support

You can also run this project inside a Docker container:

docker build -t recommender-app .
docker run -p 8501:8501 recommender-app

📒 Logging

All user actions, errors, and recommendations are logged in the logs/recommender.log file for monitoring and debugging purposes.

❗ Troubleshooting

• —Ensure all .pkl and .index files are placed correctly inside the artifacts folder.
• —Python ≥ 3.8 is recommended.
• —Check that streamlit and sentence-transformers are installed properly.

📄 License

This project is licensed under the MIT License.

🙏 Acknowledgments

• —Kaggle Datasets for anime and web series metadata.
• —GroupLens for the MovieLens 32M dataset.
• —Streamlit, FAISS, and SentenceTransformers libraries.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference