Retrieve Augment Generate — Build Your First RAG Pipeline in Python | Episode 16

Retrieve Augment Generate — Build Your First RAG Pipeline in Python

This example demonstrates how to implement a complete Retrieve-Augment-Generate (RAG) pipeline in Python using ChromaDB for document retrieval and Claude for answer generation. By building a programming language knowledge base and querying it, you can retrieve relevant documents, augment prompts with context, and generate accurate answers grounded in your data rather than the model’s memory.

Code

from chromadb import Client
from some_claude_sdk import ClaudeClient  # hypothetical Claude SDK import

# Initialize ChromaDB client and Claude client
chroma_client = Client()
claude = ClaudeClient(api_key="your_api_key")

# Connect to or create a collection of programming language docs
collection = chroma_client.get_or_create_collection("programming_languages")

# Example documents to add (only run once)
docs = [
  {"id": "py", "text": "Python is a versatile, interpreted language known for readability."},
  {"id": "js", "text": "JavaScript runs in browsers and enables interactive web pages."},
  {"id": "rust", "text": "Rust is a systems language focusing on safety and performance."},
  {"id": "go", "text": "Go is a statically typed language designed at Google for concurrency."},
  {"id": "ts", "text": "TypeScript is a typed superset of JavaScript that compiles to plain JS."}
]
# collection.add_documents(docs)  # Uncomment to add docs initially

def rag_pipeline(question: str) -> str:
  # Step 1: Retrieve relevant documents from ChromaDB
  results = collection.query(query_texts=[question], n_results=3)
  context = "\n".join(doc['text'] for doc in results['documents'][0])

  # Step 2: Augment prompt with retrieved context
  prompt = f"""
  You are an AI assistant. Answer the question using only the context below.
  Context:
  {context}

  Question:
  {question}

  Answer:"""

  # Step 3: Generate answer with Claude using the augmented prompt
  response = claude.generate(prompt=prompt, max_tokens=200)
  return response.text.strip()

# Example questions
questions = [
  "What is unique about Rust programming language?",
  "How does TypeScript relate to JavaScript?",
  "What makes Go suitable for concurrent programming?"
]

for q in questions:
  answer = rag_pipeline(q)
  print(f"Q: {q}\nA: {answer}\n")

Key Points

• The RAG pipeline consists of retrieving relevant documents, augmenting the prompt with context, and generating an answer.
• ChromaDB efficiently indexes and retrieves documents relevant to the user’s query.
• Prompt augmentation ensures the language model answers based strictly on retrieved context, improving accuracy.
• Combining Claude with ChromaDB allows building AI apps that answer questions from your own data, not just model memory.
• This pattern forms the foundation for document-based Q&A and knowledge retrieval applications.