Skip to main content

Simple RAG Tutorial

Welcome to your first RAG implementation! In this tutorial, you'll build a basic Retrieval-Augmented Generation system using LlamaIndex.

What You'll Build​

A simple RAG system that can:

  • Load and index documents
  • Answer questions based on document content
  • Provide source attribution

Prerequisites​

Step 1: Import Dependencies​

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings
from llama_index.embeddings.openai import OpenAIEmbedding
from llama_index.llms.openai import OpenAI
import os
from dotenv import load_dotenv

# Load environment variables
load_dotenv()

Step 2: Configure Settings​

# Configure LLM and embedding model
Settings.llm = OpenAI(
    model="gpt-3.5-turbo",
    api_key=os.getenv("OPENAI_API_KEY")
)

Settings.embed_model = OpenAIEmbedding(
    model="text-embedding-ada-002",
    api_key=os.getenv("OPENAI_API_KEY")
)

Step 3: Load Documents​

# Create a data directory and add some text files
# For this example, create a file called 'sample.txt' in ./data/

# Load documents from data directory
documents = SimpleDirectoryReader("./data").load_data()
print(f"Loaded {len(documents)} documents")

Step 4: Create Vector Index​

# Create vector store index
index = VectorStoreIndex.from_documents(documents)
print("Index created successfully!")

Step 5: Create Query Engine​

# Create query engine
query_engine = index.as_query_engine()

# Test query
response = query_engine.query("What is this document about?")
print(f"Response: {response}")

Complete Code​

Here's the full implementation:

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader, Settings
from llama_index.embeddings.openai import OpenAIEmbedding
from llama_index.llms.openai import OpenAI
import os
from dotenv import load_dotenv

def main():
    # Load environment variables
    load_dotenv()

    # Configure Settings
    Settings.llm = OpenAI(
        model="gpt-3.5-turbo",
        api_key=os.getenv("OPENAI_API_KEY")
    )

    Settings.embed_model = OpenAIEmbedding(
        model="text-embedding-ada-002",
        api_key=os.getenv("OPENAI_API_KEY")
    )

    # Load documents
    documents = SimpleDirectoryReader("./data").load_data()
    print(f"Loaded {len(documents)} documents")

    # Create index
    index = VectorStoreIndex.from_documents(documents)

    # Create query engine
    query_engine = index.as_query_engine()

    # Interactive query loop
    while True:
        query = input("\nAsk a question (or 'quit' to exit): ")
        if query.lower() == 'quit':
            break

        response = query_engine.query(query)
        print(f"\nAnswer: {response}")

if __name__ == "__main__":
    main()

Running the Code​

  1. Create data directory:

    mkdir data

  2. Add sample document: Create data/sample.txt with some content

  3. Run the script:

    python simple_rag.py

Expected Output​

Loaded 1 documents
Index created successfully!

Ask a question (or 'quit' to exit): What is this document about?

Answer: [Your document summary based on content]

What's Happening?​

  1. Document Loading: SimpleDirectoryReader loads text files
  2. Embedding: Documents are converted to vector embeddings
  3. Indexing: Vectors are stored in a searchable index
  4. Retrieval: Query finds relevant document chunks
  5. Generation: LLM generates answer using retrieved context

Next Steps​

🎉 Congratulations! You've built your first RAG system.

Continue learning:

Troubleshooting​

Common issues:

  • API key errors: Check your .env file
  • No documents found: Ensure files are in ./data/ directory
  • Import errors: Run pip install -r requirements.txt

Need help? Join our community discussions!