🚀 Langchain Notebook Setup

This project demonstrates how to use Langchain and OpenAI in Jupyter notebooks for two main workflows:

• lang.ipynb: Automatically break down a complex question into multiple sub-questions using a Pydantic model and a system prompt.
• RAG-lang.ipynb: Retrieval-Augmented Generation (RAG) workflow, where the LLM answers a question using relevant context retrieved from a set of documents.

lang.ipynb: Sub-question Decomposition

It leverages a Pydantic model to enforce structured output (a list of sub-questions), and uses a system prompt together with an actual question (e.g., "What are the main components of an LLM-powered autonomous agent system?") to instruct the LLM to decompose the query into actionable sub-tasks.

What does the code do?

The notebook (lang.ipynb) loads your OpenAI API key from a .env file and uses the Langchain library to:

• 🧩 Define a Pydantic model for structured output, ensuring the response is always a list of sub-questions.
• 📝 Set up a system prompt that instructs the AI to break down any main question into sub-questions.
• ❓ Send an actual question (e.g., "What are the main components of an LLM-powered autonomous agent system?") to the OpenAI model, together with the system prompt.
• 📋 Receive and print a list of sub-questions, each representing a smaller part of the original query.

You can modify the Pydantic model or the system prompt, and use this pattern to decompose any complex query into actionable sub-tasks using LLMs.

RAG-lang.ipynb: Retrieval-Augmented Generation

This notebook demonstrates how to use a simple in-memory FAISS vectorstore and OpenAI to answer questions using retrieved context:

• 📄 Store example documents in a vector database (FAISS) with OpenAI embeddings.
• 🔍 Retrieve relevant context for a user question.
• 🧠 Format a system prompt with the retrieved context and the user's question.
• 🤖 Use the OpenAI model to generate a concise answer based on the context.

You can adapt this workflow to use your own documents for more advanced RAG applications.

1️⃣ Create and Activate Virtual Environment

🛠️ Open your terminal and run:

python -m venv venv
venv\Scripts\activate

2️⃣ Install Required Packages

📦 Install dependencies with:

pip install -r requirements.txt

💡 Or, to install manually:

pip install langchain-openai langchain-core python-dotenv

3️⃣ Set Up OpenAI API Key

🔑 Create a file named .env in the project folder with this content:

OPENAI_API_KEY=sk-<your-openai-key-here>

4️⃣ Start Jupyter Notebook

📓 Launch Jupyter with:

jupyter notebook

➡️ Open lang.ipynb and run the cells to get started!