LlamaIndex sample: what you configure

Aligned with e2e-tests/dragent/llamaindex/.

workflow.yaml

Piece	What you see
llms: + datarobot-llm-component	Named LLM for the app; see LLM configuration.
workflow._type: llamaindex_agent	DRAgent uses the LlamaIndex integration.
workflow.llm_name	Must match llms.

myagent.py

You define the AgentWorkflow, agents (e.g. planner / writer), and how the final answer is read from streamed events—the sample shows one way to extract text. The {chat_history} placeholder in the user message string is filled by the integration when you opt in (see the example’s string patterns).

register.py wires the package for DRAgent.