Book Recommendation System

Book repository hosts a sophisticated book recommendation system employing SQL and pandas-based methods. The system leverages two key strategies: (1) popularity-based recommendations, which rely on average ratings and total ratings of books, and (2) collaborative filtering, which utilizes the KNN algorithm to suggest books to users based on their interactions with similar users.

Dataset

The dataset used for this project can be downloaded from Kaggle. After downloading, extract the contents into a directory named data inside Book repository.

Models

• Pandas-based Models:

  • recommend_popularity_based_pd: Recommends popular books based on pandas DataFrame.
  • recommend_collaborative_filtering_preprocessing: Preprocesses the data for collaborative filtering.
  • recommend_collaborative_filtering_train: Trains the collaborative filtering model.
  • recommend_collaborative_filtering_test: Tests the collaborative filtering model by generating recommendations for a given user.

• SQL-based Models:

  • recommend_popularity_based_sql: Recommends popular books based on SQL database using SQLAlchemy.
  • recommend_collaborative_filtering_preprocessing_sql: Preprocesses the data for collaborative filtering using SQL.
  • recommend_collaborative_filtering_train_sql: Trains the collaborative filtering model.
  • recommend_collaborative_filtering_test_sql: Tests the collaborative filtering model by generating recommendations for a given user using SQL.

Files

• books.py: Contains the BookData class to load and store data using pandas.
• database.py: Contains SQLAlchemy models and functions to interact with the SQLite database.
• models_pd.py: Contains pandas-based recommendation models.
• models_sql.py: Contains SQL-based recommendation models.
• main.py: Demonstrates how to use both SQL and pandas-based approaches for book recommendation.
• app.py: Contains the FastAPI application for CRUD operations on books to manipulate the database and generate recommendations by giving user-id using a trained KNN model.
• utils.py: Contains load and save model functions.
• knn: The trained KNN model.

Installation

1. Clone the repository:

   git clone https://github.com/your_username/your_repository.git

2. Navigate to the project directory:

   cd Recommendation-System-Projects

3. Install the project dependencies:

   pip install -r requirements.txt

4. Train the model and see the results:

   python main.py

5. Access the FastAPI Swagger UI: Use uvicorn app:app --reload command under Book repository. Open your web browser and go to http://127.0.0.1:8000/docs to test the FastAPI application for CRUD operations on books and to generate recommendations using collaborative filtering.
6. Build Dockerfile:

   docker build -t recommendation-app .
   docker run -d -p 8000:8000 recommendation-app

   You can access http://localhost:8000/ after built.

CI/CD & Deployment

After each commit, GitHub Actions automatically initiates the building, testing, and deployment processes. Docker plays a crucial role in our CI/CD pipeline by containerizing the application. For more details, refer to .github/workflows/main.yml.

To configure DockerHub integration, specify your DOCKERHUB_USERNAME and DOCKERHUB_TOKEN in Settings -> Secrets and Variables -> Actions.

Once your Docker image is successfully built, navigate to DockerHub, open your image, select Public View, and copy the Docker Pull Command.

Then, execute the command in your terminal. Finally, run docker run $ImageName$ to access your application via the provided link.

License

This project is licensed under the MIT License.