🧠 LangChain Docs Bot with Memory and Logging

This assistant answers developer questions using a curated set of LangChain .txt files. It uses FAISS for vector search, HuggingFace embeddings for chunk encoding, and Phi-2 for grounded LLM responses. The bot supports session-based memory and logs every Q&A to a file. Designed for CLI use.

---

🔧 Features

• Modular ingestion of LangChain .txt files
• Chunking and embedding via HuggingFace + FAISS
• Retrieval of top-matching chunks
• Prompt injection using a universal template
• Response generation via Phi-2
• Session memory for multi-turn Q&A
• Logging of all interactions

---

📁 File Structure

• ingest.py: Loads and chunks .txt docs into FAISS
• retrieve.py: Retrieves answers with memory + logging
• requirements.txt: All dependencies
• README.md: Project overview and usage guide

---

🚀 How to Run

pip install -r requirements.txt
python ingest.py
python retrieve.py

🔗 GitHub Repository

[LangChain Docs Bot on GitHub](https://github.com/paulsuv9/langchain-docs-bot)