LangChain Lab: RAG, Chains, and Chat Model Patterns

Hands on patterns for building LLM applications with LangChain, written as small, focused, runnable scripts.

Contents

• RAG (RAGS/): retrieval augmented generation over classic literature (Alice in Wonderland, Dracula, Frankenstein, The Lord of the Rings). Document loading, chunking, embeddings, a ChromaDB vector database, and metadata filtered retrieval.
• Chains (chains/): chain basics and inner workings, then sequential, parallel, and conditional composition.
• Chat models (chat_models/): conversation loops, interactive chat, and persisting message history to Firebase.
• Prompt templates (prompt_templates/): parameterized prompting patterns.

Stack

LangChain · ChromaDB · Python