llamaindex-patterns

Overview

LlamaIndex is a data framework for connecting LLMs with external data. It provides document loaders, vector stores, query engines, and chat engines for building RAG applications and knowledge-augmented agents.

Capabilities

• Load documents from 160+ sources (PDF, Notion, Slack, databases)
• Build vector indices for semantic search
• Create query engines with retrieval and synthesis
• Build conversational chat engines with memory
• Use agents with tool use and multi-step reasoning
• Implement advanced RAG patterns (routing, fusion, recursive)

When to Use

Trigger phrases:

• "llamaindex patterns"

• "LlamaIndex data framework — ingestion, indexing, query engines, chat engines, ag"

• Building RAG applications over custom data sources

• Needing structured document ingestion pipelines

• Wanting query engines with citations and source tracking

• Building chatbots over knowledge bases

• Implementing agentic RAG with tool use

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Pseudo Code

# Example workflow for this skill
def execute(input_data):
    # Step 1: Validate input
    if not input_data:
        raise ValueError("Input data is required")

    # Step 2: Process core logic
    result = process(input_data)

    # Step 3: Validate output
    validate_output(result)

    return result

Document Ingestion

from llama_index.core import VectorStoreIndex, SimpleDirectoryReader

# Load documents
documents = SimpleDirectoryReader("./data").load_data()

# Build index
index = VectorStoreIndex.from_documents(documents)

Query Engine

from llama_index.core import VectorStoreIndex
from llama_index.llms.openai import OpenAI

llm = OpenAI(model="gpt-4o", temperature=0)

index = VectorStoreIndex.from_documents(documents)
query_engine = index.as_query_engine(
    llm=llm,
    similarity_top_k=5,
    response_mode="compact",  # or "tree_summarize", "refine"
)

response = query_engine.query("What are the company's revenue streams?")
print(response.response)
print(response.source_nodes)  # Citations