test(bidirectional): Add integ test

[mkmeral]

Description

Adds integration tests for bidirectional streaming agents using Nova Sonic with a thread-based architecture that mimics real-world usage patterns.

Key Features

• Thread-based test context manager - Manages continuous audio streaming with background threads that simulate real microphone input
• Multi-turn conversation support - Enables natural back-and-forth conversations with follow-up questions
• Amazon Polly integration - Generates test audio via text-to-speech, eliminating need for physical audio devices or pre-recorded files
• Audio caching - Caches generated audio for fast test re-runs
• Automatic lifecycle management - Context manager handles agent start/end and thread cleanup

Architecture

The test infrastructure uses a context manager (BidirectionalTestContext) that manages two background threads:

1. Input thread - Continuously streams audio (real speech + silence padding) or sends text/other data
2. Event collection thread - Captures all model outputs (text, audio, tool calls)

This approach enables realistic testing of continuous streaming scenarios without requiring physical audio hardware.

Usage Example

async with BidirectionalTestContext(agent, audio_generator) as ctx:
    # Turn 1
    await ctx.say("What is 5 + 3?")
    await ctx.wait_for_response()
    
    # Turn 2 - follow-up question
    await ctx.say("Now multiply that by 2")
    await ctx.wait_for_response()
    
    # Validate
    text_outputs = ctx.get_text_outputs()
    assert "8" in " ".join(text_outputs)
    assert "16" in " ".join(text_outputs)

Files Added

• tests_integ/bidirectional_streaming/test_bidirectional_agent.py - Integration tests
• tests_integ/bidirectional_streaming/utils/test_context.py - Thread-based context manager
• tests_integ/bidirectional_streaming/utils/audio_generator.py - Polly integration with caching
• tests_integ/bidirectional_streaming/conftest.py - Pytest fixtures
• tests_integ/bidirectional_streaming/__init__.py - Module init
• tests_integ/bidirectional_streaming/utils/__init__.py - Utils module init

Related Issues

Documentation PR

Type of Change

New feature

Testing

How have you tested the change?

• Ran hatch run prepare - all checks pass
• Ran integration tests: pytest tests_integ/bidirectional_streaming/ - 2 tests pass
• Verified multi-turn conversations work correctly
• Verified tool execution (calculator) works
• Verified audio I/O streaming works
• Verified session lifecycle management works
• No linting errors
• No type checking errors

Test Results

tests_integ/bidirectional_streaming/test_bidirectional_agent.py::test_single_turn_conversation PASSED
tests_integ/bidirectional_streaming/test_bidirectional_agent.py::test_multi_turn_conversation PASSED

2 passed in 20.52s

Verify that the changes do not break functionality or introduce warnings in consuming repositories: agents-docs, agents-tools, agents-cli

• I ran hatch run prepare

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[JackYPCOnline]

TODO: check tool call after we emit those events.