📚 LLM-Based Semantic Book Recommender System

A semantic book recommendation system that uses Large Language Models, OpenAI embeddings, and LangChain to deliver personalized book suggestions based on your input description, emotional tone, and category.

⚠️ Note on Deployment
Due to limited OpenAI API credits, I have not provided a live Hugging Face demo link for this project.

While the app is fully functional and tested, real-time queries depend on OpenAI's paid API access.

If you'd like to try it out locally, you're welcome to clone the repo and add your own OpenAI API key.

Your understanding and cooperation are deeply appreciated. 🙏 Just trust me that I have deployed on huggingface spaces 😅

---

🧠 Key Features

• 🔍 Semantic Search using OpenAI Embeddings and ChromaDB
• ❤️ Emotion-aware recommendations powered by a fine-tuned emotion classifier
• 🧾 Category mapping to simplify complex metadata
• 🖼️ Interactive Gradio dashboard with thumbnails & smart captions
• ☁️ Hosted on Hugging Face Spaces for free & public access

---

🗂️ Project Structure

📦 book_recommendor ├── app.py # Gradio app entry point (main dashboard) ├── books_with_emotions.csv # Enriched book metadata with emotions ├── tagged_description.txt # Raw text for Chroma vector store ├── cover-not-found.jpg # Default image if book thumbnail is missing ├── requirements.txt # Python dependencies └── README.md # Project documentation

---

🛠️ How It Works

1. Vector Embedding: tagged_description.txt is processed with OpenAIEmbeddings to create a vector store using Chroma.
2. Query Input: User inputs a sentence like "A heartwarming story about forgiveness".
3. Similarity Search: Top matching books are retrieved semantically.
4. Tone & Category Filter: Filters results based on emotional tone and book category.
5. Output Display: Shows book cover, author(s), and a short summary using Gradio.

---

🤖 ML & NLP Components

• Text Embedding: langchain_openai.OpenAIEmbeddings
• Vector Search: Chroma from langchain.vectorstores
• Emotion Detection: j-hartmann/emotion-english-distilroberta-base via Hugging Face Transformers
• Frontend UI: Gradio with dropdown filters for category and tone

---

📦 Setup & Run Locally

1. Clone the Repository

git clone https://github.com/MuhammadHaweras/Book_Recommendor.git
cd book_recommendor

pip install -r requirements.txt

Add OpenAI Key (Create a .env file)

Run the App
python app.py

🏷️ Tags
#LangChain #LLM #BookRecommender #Gradio #OpenAI #ChromaDB #EmotionClassification #HuggingFaceSpaces