LangChain Exploration Project

This repository contains my hands‑on exploration of LangChain and various vector database integrations.
I experimented with data ingestion, text splitting, embeddings and vector stores (FAISS, ChromaDB) to understand how LLM applications are built end‑to‑end.

---

📂 Project Structure

1. Data Ingestion & Transformation

• dataingestion.ipynb → Notebook for loading and preparing raw data.
• Hospital.pdf → Sample document used for ingestion.
• speech.txt → Sample text file for testing.
• Transformers & Splitters:
  • CharacterTextSplitter.ipynb
  • HTMLtextSplitter.ipynb
  • JsonSplitter.ipynb
  • text_splitter.ipynb

2. Embeddings

• HuggingFace.ipynb → Using Hugging Face models for embeddings.
• ollamaembedding.ipynb → Generating embeddings with Ollama models.

3. Vector Stores

• FAISS
  • Faiss.ipynb → Working with FAISS vector store.
  • faiss_index/, index.faiss, index.pkl → Saved FAISS index files.
• ChromaDB
  • chroma.ipynb → Working with ChromaDB vector store.
  • chroma_db/62c68d8c-5165-454a-9838-3417bce0a066/ → Local Chroma database folder.

4. Project Setup

• .gitignore → Ensures venv, indexes, and DB files aren’t committed unnecessarily.
• requirements.txt → Python dependencies for running notebooks and scripts.