[jhackshaw]

Send the thinking config on the interactions API. build_generation_config dropped thinking_config and logged nothing. Now thinking_level and include_thoughts are sent, thinking_budget logs a warning, and non-streaming responses keep their thought summaries.

Link to Issue or Description of Change

1. Link to an existing issue (if applicable):

• Closes: thinking_config is silently dropped with use_interactions_api=True #7924

2. Or, if no issue exists, describe the change:

Problem:
With Gemini(use_interactions_api=True), nothing from thinking_config reaches the request:

• thinking_level is dropped, so the model thinks at its default level, and an unsupported level is accepted instead of returning a 400.
• include_thoughts is dropped, so summaries never come back.
• thinking_budget is dropped without the warning that unsupported sampling parameters get.

The non-streaming path also discarded ThoughtStep.summary, so summaries only reached SSE mode.

Solution:

• thinking_level → generation_config['thinking_level'], lowercased. THINKING_LEVEL_UNSPECIFIED is skipped.
• include_thoughts → generation_config['thinking_summaries'] (True → "auto", False → "none").
• thinking_budget has no interactions equivalent, so it gets a one-time warning like the sampling parameters.
• Non-streaming: each ThoughtStep.summary text becomes a Part(text=..., thought=True), as _handle_thought_summary does for streaming.

No version check is needed: both fields are on interactions.GenerationConfigParam in every supported google-genai release (>= 2.30).

Testing Plan

Unit Tests:

• I have added or updated unit tests for my change.
• All unit tests pass locally.

New tests in tests/unittests/models/test_interactions_utils.py cover the level and summaries mapping, the budget warning, and non-streaming summary parts.

$ pytest tests/unittests/models/test_interactions_utils.py
216 passed

Without the fix, 10 of these fail.

Manual End-to-End (E2E) Tests:

An LlmAgent on gemini-3.8-flash (Vertex AI) with use_interactions_api=True and thinking_level=HIGH. ADK's DEBUG request log:

	Generation Config	thought tokens (reasoning prompt, 3 runs)
main	{}	136, 138, 120
this branch	{"thinking_level": "high"}	350, 358, 309

Checklist

• I have read the CONTRIBUTING.md document.
• I have performed a self-review of my own code.
• I have commented my code, particularly in hard-to-understand areas.
• I have added tests that prove my fix is effective or that my feature works.
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end.
• Any dependent changes have been merged and published in downstream modules.

[adk-bot]

Thank you @jhackshaw for your contribution! 🎉

Your changes have been successfully imported and merged via Copybara in commit c20b90a.

Closing this PR as the changes are now in the main branch.