test(models): re-enable finish_reason unknown-maps-to-OTHER litellm test

[anxkhn]

Please ensure you have read the contribution guide before creating a pull request.

Link to Issue or Description of Change

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

• N/A (test-only hardening, no behavior change)

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

Problem:
test_finish_reason_unknown_maps_to_other in
tests/unittests/models/test_litellm.py is skipped with
@pytest.mark.skip(reason="LiteLLM finish_reason mapping behaviour changed").
That skip was added during the v2.0.0 GA transition, and the reason is now stale.
LiteLlm._map_finish_reason ends with:

return _FINISH_REASON_MAPPING.get(finish_reason_str, types.FinishReason.OTHER)

so any finish_reason that is not in _FINISH_REASON_MAPPING (which only has
length, stop, tool_calls, function_call, content_filter)
deterministically maps to FinishReason.OTHER, independent of LiteLLM's own
normalization. The test deliberately builds a mock choice/response that returns a
raw "totally_unknown_reason" to bypass LiteLLM's normalization and exercise
ADK's own fallback, and asserts the result is OTHER. This skipped test is the
only coverage of the _map_finish_reason -> OTHER fallback branch in the litellm
suite, so the skip has left a live production branch untested.

Solution:
Remove the single @pytest.mark.skip(...) line above the test. No production
change. The test passes as-is against current code, restoring coverage of the
OTHER fallback. I chose deletion over refreshing the mock because the extracted
test body already drives generate_content_async ->
_model_response_to_generate_content_response -> _map_finish_reason -> OTHER
correctly against the current source; no mock detail has drifted.

Diff (1 file, -1 line):

-@pytest.mark.skip(reason="LiteLLM finish_reason mapping behaviour changed")
 @pytest.mark.asyncio
 async def test_finish_reason_unknown_maps_to_other(
     mock_acompletion, lite_llm_instance

Testing Plan

Unit Tests:

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

pytest results (with uv run --no-sync):

• Re-enabled test (was skipped, now runs):
  pytest tests/unittests/models/test_litellm.py::test_finish_reason_unknown_maps_to_other -q
  -> 1 passed.
• Neighborhood:
  pytest tests/unittests/models/test_litellm.py -q -k "finish_reason or map_finish"
  -> 10 passed.
• Full file:
  pytest tests/unittests/models/test_litellm.py -q
  -> 318 passed (previously 317 passed, 1 skipped; the delta is exactly this
  re-enabled test).

I also confirmed the test genuinely guards the branch: temporarily changing the
production fallback from OTHER to STOP made this test fail
(AssertionError: - OTHER + STOP); reverting restored green. So it is an
effective regression test, not a vacuous one.

Manual End-to-End (E2E) Tests:

N/A. This is a test-only change (removal of a stale skip marker); there is no
runtime or user-facing behavior to test manually.

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. (N/A: a
  one-line skip-marker removal; the test already carries an explanatory
  docstring and comments.)
• I have added tests that prove my fix is effective or that my feature works.
  (Re-enables the sole existing coverage of the _map_finish_reason -> OTHER
  fallback.)
• New and existing unit tests pass locally with my changes.
• I have manually tested my changes end-to-end. (N/A: test-only change.)
• Any dependent changes have been merged and published in downstream modules.
  (None.)

Additional context

The mapping table and fallback live in src/google/adk/models/lite_llm.py
(_FINISH_REASON_MAPPING and _map_finish_reason). The test reaches the fallback
through LiteLlm.generate_content_async ->
_model_response_to_generate_content_response, which reads
first_choice.get("finish_reason") and passes it to _map_finish_reason.