Build Your First RAG App with LlamaIndex

1. 1
   Install LlamaIndex & Set API Key

   Install the necessary Python packages for LlamaIndex and its OpenAI integration. You will also need to set your OpenAI API key as an environment variable for authentication.
2. 2
   Prepare and Load Data

   LlamaIndex needs data to work with. Create a directory named 'data' and place a text file inside. Then, use the `SimpleDirectoryReader` to load all documents from that directory.
3. 3
   Create a Vector Index

   An index is a data structure that allows for efficient querying. The `VectorStoreIndex` creates vector embeddings of your document text, enabling semantic search. This step converts your documents into a queryable format.
4. 4
   Create a Query Engine

   A query engine is the primary interface for asking questions over your data. The `as_query_engine()` method builds a simple RAG pipeline that retrieves relevant context from the index and passes it to an LLM to generate an answer.
5. 5
   Query Your Data

   Use the query engine to ask a question. LlamaIndex will find relevant text chunks from your documents, feed them to the LLM as context, and return a synthesized answer based on your data, not the LLM's general knowledge.

import os
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader

# --- 1. Setup & API Key ---
# Ensure your OPENAI_API_KEY is set in your environment variables
if "OPENAI_API_KEY" not in os.environ:
    raise EnvironmentError("OPENAI_API_KEY environment variable not set.")

# --- 2. Prepare Data ---
# Create a 'data' directory if it doesn't exist
if not os.path.exists("data"):
    os.makedirs("data")

# Create a dummy text file with some content
file_path = "data/my_story.txt"
with open(file_path, "w") as f:
    f.write("Paul Graham is a computer scientist and entrepreneur. ")
    f.write("In 1995, he and Robert Morris founded Viaweb, the first software as a service company. ")
    f.write("Viaweb was acquired by Yahoo in 1998. In 2005, he co-founded the seed capital firm Y Combinator.")

print("Created dummy data file: data/my_story.txt")

# --- 3. Load Data ---
# Load the documents from the 'data' directory
documents = SimpleDirectoryReader("data").load_data()
print(f"Loaded {len(documents)} document(s).")

# --- 4. Index Data ---
# Create a VectorStoreIndex from the documents
# This will create vector embeddings and store them in-memory by default
print("Indexing documents...")
index = VectorStoreIndex.from_documents(documents)

# --- 5. Create Query Engine ---
# The query engine is the main interface for asking questions
query_engine = index.as_query_engine()
print("Query engine created.")

# --- 6. Query Data ---
# Ask a question about the content of the documents
query_text = "What did Paul Graham co-found in 2005?"
print(f"\nQuerying: '{query_text}'")
response = query_engine.query(query_text)

print("\nResponse:")
print(response)

# To run this script:
# 1. pip install llama-index llama-index-llms-openai
# 2. export OPENAI_API_KEY='your-openai-api-key'
# 3. python your_script_name.py