How apoorvvats181/ai-powered-book-recommendation-engine works
Project type and purpose
• This is a data‑driven, full‑stack book recommendation engine.
• It ingests cleaned book metadata (CSV files), performs NLP pipelines (classification, sentiment analysis, embeddings), stores vector representations in ChromaDB, and exposes an interactive Gradio dashboard for end users.
File structure
• Top‑level data files (books_cleaned.csv, books_with_categories.csv, books_with_emotions.csv) indicate an ETL/data‑prep stage.
• Multiple Jupyter notebooks (data‑exploration.ipynb, sentiment‑analysis.ipynb, text‑classification.ipynb, vector‑search.ipynb) show step‑by‑step pipeline development.
• gradio-dashboard.py is the entry point for the UI service.
• tag_descriptor.txt suggests token‑level tagging or additional metadata for classification.
README for extra insights
• Core ML components: OpenAI embeddings, Hugging Face BART (zero‑shot classification), RoBERTa (sentiment).
• Vector DB: ChromaDB.
• Framework: LangChain orchestrates searches and chaining model calls.
• UI: Gradio serves the front‑end interface.
Final advice
– Capture every pipeline stage as its own component, even if implemented in notebooks.
– Emphasize the orchestration role of LangChain between UI and ML modules.
– Don’t omit the vector database – it’s central to the semantic search capability.