code

Code Examples

Seven self-contained Python projects demonstrating the patterns covered in this guide.

Prerequisites

pip install anthropic python-dotenv

Each project has its own requirements.txt with additional dependencies.

Projects

Project	What It Demonstrates	Key Dependencies
01-basic-rag	Full RAG pipeline from scratch	anthropic, sentence-transformers, numpy
02-advanced-rag	HyDE, reranking, parent-child chunking	anthropic, sentence-transformers, qdrant-client
03-langgraph-agent	ReAct agent with LangGraph	anthropic, langgraph, langchain-anthropic
04-mcp-server	Custom MCP server	anthropic, mcp
05-eval-pipeline	LLM-as-judge evaluation pipeline	anthropic, pandas
06-semantic-cache	Semantic caching with embeddings	anthropic, sentence-transformers, redis
07-structured-output	Pydantic structured extraction	anthropic, instructor, pydantic

Setup

1. Copy .env.example to .env in any project directory
2. Add your ANTHROPIC_API_KEY
3. Install dependencies: pip install -r requirements.txt
4. Run the example: python main.py

Design Philosophy

• Each project is standalone — no shared code between projects
• Under 200 lines per main file so the full logic fits in one read
• Real runnable code, not pseudocode
• Uses the Anthropic Python SDK as the primary interface