How abhishek445566/book-recommendation-system works
Main Components
• Data Ingestion & Preprocessing: Represent the input data process where raw book data is downloaded (from Kaggle) and cleaned. This corresponds to the “data-exploration.ipynb” notebook.
• Zero-Shot Text Classification: Include a component for classifying books as “fiction” or “non-fiction” using LLMs. This is contained in the “text-classification.ipynb” notebook.
• Sentiment Analysis: Set apart the module that extracts sentiment or emotional tone from text (e.g., suspense, joy, sadness) as shown in the “sentiment-analysis.ipynb” notebook.
• Web Application / User Interface: Illustrate the front-end built using Gradio (represented in “gradio-dashboard.py”) that brings all functionalities together for end users.
• External Services & Configuration: Don’t forget to include the dependency on the OpenAI API (with the configuration via the .env file) and any other external APIs if applicable.
Relationships and Interactions
• The Data Ingestion module feeds cleaned data to the Vectorization module.
• The output of vectorization (i.e., vector embeddings) is stored in or accessed from the vector database for semantic search.
• The same cleaned data is processed by the Classification and Sentiment Analysis modules, which annotate the books with additional metadata (e.g., category labels and sentiment scores).
• The Gradio dashboard serves as the entry point for users; it sends natural language queries that are processed by the semantic search component and may incorporate results from the classification and sentiment analysis modules.
Architectural Patterns and Technologies
• Data Pipeline: Emphasize a modular data pipeline where each notebook represents a distinct processing module.
• Service Integration: Note the integration with external LLM services via the OpenAI API and how configuration is managed (via the .env file).
• Web Frontend Framework: Mark the Gradio component that creates the web-based user interface, providing an interactive dashboard for book recommendations.
• Converge these processing outputs into a combined “Book Enrichment” stage.
• Finally, connect the enriched data to the “Gradio Dashboard” (gradio-dashboard.py) which interacts with the user.