🎓 University Information Retrieval Bot

An AI-powered chatbot assistant built to help students get accurate, document-backed information about universities in Pakistan. It leverages semantic search and LLMs to answer queries related to admissions, programs, faculty, campus life, and more — based on real university documents.

✅ Features

• 📄 PDF-based document ingestion and embedding
• 🔍 Semantic search using ChromaDB + HuggingFace embeddings
• 💬 LLM-based responses with grounded context
• 🧠 Metadata-aware document retrieval
• 🇵🇰 Built specifically for Pakistani universities and students

---

🛠️ Tech Stack

• Python
• LangChain
• ChromaDB
• HuggingFace Embeddings
• PyPDFLoader

---

🚀 How It Works

1. Document Parsing: University PDFs are loaded from the Books/ directory. Metadata is stored in JSON files under metadata/.
2. Vectorization: Documents are embedded and stored in ChromaDB with metadata.
3. Query Pipeline:
   • User submits a query (e.g., "What is QAU's fee structure?")
   • System retrieves top matching documents
   • The context is passed to an LLM
   • A response is generated based on the actual content

---

🧪 Example Query

> Who is the rector of NUST?

> 🛠️ Currently supports:
> - National University of Sciences and Technology (NUST)
> - Lahore University of Management Sciences (LUMS)
> - Quaid-i-Azam University (QAU)

More universities will be added soon.