… pipeline activate it prior to streaming results.

…lementation details

…bled agents

… STT services

This commit updates the test suite to ensure comprehensive coverage for the recent changes to how voice configurations (TTS model, settings) are managed and resolved within the voice agent system.

Key changes:

- **`tests/voice/test_pipeline.py`**:
    - Added tests for `VoicePipeline.get_effective_voice_configuration` to verify correct prioritization of workflow-provided vs. pipeline-default configurations.
    - Updated existing pipeline run tests (`run_single_turn`, `run_multi_turn`) to ensure they correctly use `VoiceConfiguration` provided by the workflow. This involved enhancing `FakeTTS` and `FakeWorkflow`.
    - Added new tests for pipeline-workflow interaction scenarios, covering cases where model name and settings are mixed and matched between workflow and pipeline configurations. A `MockModelProvider` was introduced for better control over TTS model resolution in tests.

- **`tests/voice/test_workflow.py`**:
    - Added tests for `VoiceWorkflowBase.get_voice_configuration` to ensure it correctly fetches configurations from its `current_agent`.
    - For `SingleAgentVoiceWorkflow`:
        - Tested retrieval of `VoiceConfiguration` from its agent.
        - Tested that `current_agent` is correctly updated after a simulated handoff.
        - Verified that `on_run` and `on_agent_change` callbacks are invoked as expected. Mocked `RunResultStreaming` for handoff simulation.

- **`tests/voice/test_voiceagent.py` (New file)**:
    - Created to test components from `src/agents/voice/voiceagent.py`.
    - Added tests for `VoiceAgentMixin` to ensure proper storage and retrieval of `VoiceConfiguration`.
    - Added tests for `with_voice_configuration` and `with_voice_config` helper functions, verifying they correctly create new agent instances with the specified voice configurations and preserve original agent details.

- **`tests/voice/test_model.py` (New file)**:
    - Created to test components from `src/agents/voice/model.py`.
    - Added tests for `VoiceConfiguration.get_effective_model()`, covering scenarios for model resolution (name vs. instance) and caching behavior.

- **`tests/voice/fake_models.py`**:
    - Enhanced `FakeTTS` to store `last_run_settings` and accept a `name` for easier instance identification in tests.

The existing tests in `tests/voice/test_openai_tts.py`, `tests/voice/test_input.py`, and `tests/voice/test_openai_stt.py` were reviewed and found not to require changes for this specific set of refactorings.

Co-authored-by: google-labs-jules[bot] <161369871+google-labs-jules[bot]@users.noreply.github.com>

…ponents

…rios

…ng behavior

…nheritance behavior