docs: add agentic retrieval README examples

[mahikaw]

Document CLI and Python examples for agentic retrieval with hosted NVIDIA endpoints so users can smoke test against an existing LanceDB index.

Description

Checklist

• I am familiar with the Contributing Guidelines.
• New or existing tests cover these changes.
• The documentation is up to date with these changes.

[greptile-apps]

Greptile Summary

This PR adds a new "Run agentic retrieval" section to nemo_retriever/README.md documenting CLI and Python API usage for the LLM-driven ReAct query loop against an existing LanceDB index.

• Two bash examples cover the full-run case and a reduced smoke-test case (--top-k 1 --agentic-react-max-steps 1 --agentic-backend-top-k 1), with NVIDIA_API_KEY exported before use.
• A Python code block uses the correct public imports (nemo_retriever.cli.query_workflow.agentic_query_documents, nemo_retriever.query.options.*) and correct parameter names verified against the live dataclass definitions; a comment notes the env-var requirement.

Confidence Score: 5/5

Documentation-only change; no executable code is modified.

All Python imports and dataclass field names in the new examples are verified against the live source files. The NVIDIA_API_KEY prerequisite is called out in both the prose and a code comment. No API contracts, logic paths, or configuration are touched.

No files require special attention.

Important Files Changed

Filename	Overview
nemo_retriever/README.md	Adds 83-line "Run agentic retrieval" section with CLI and Python examples; all imports and dataclass field names verified against source.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[User sets NVIDIA_API_KEY] --> B[retriever query --agentic OR agentic_query_documents]
    B --> C{GPU available?}
    C -- Yes --> D[Local GPU embedding actor]
    C -- No --> E[CPU/remote embedding via hosted NVIDIA endpoint]
    D --> F[LanceDB index lookup]
    E --> F
    F --> G[ReAct LLM loop via nvidia/llama-3.3-nemotron-super-49b-v1.5]
    G --> H{Max steps reached?}
    H -- No --> F
    H -- Yes --> I[Return results]

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
flowchart TD
    A[User sets NVIDIA_API_KEY] --> B[retriever query --agentic OR agentic_query_documents]
    B --> C{GPU available?}
    C -- Yes --> D[Local GPU embedding actor]
    C -- No --> E[CPU/remote embedding via hosted NVIDIA endpoint]
    D --> F[LanceDB index lookup]
    E --> F
    F --> G[ReAct LLM loop via nvidia/llama-3.3-nemotron-super-49b-v1.5]
    G --> H{Max steps reached?}
    H -- No --> F
    H -- Yes --> I[Return results]

Loading

_{Reviews (3): Last reviewed commit: "docs: clarify agentic endpoint defaults" | Re-trigger Greptile}