fixing lifecycle issues with Terraform moduel for AI Foundry

[JinLee794]

This pull request introduces several updates across configuration, deployment, and codebase files to improve naming consistency, environment setup, and agent lifecycle management. The most significant changes include renaming the project and repository, updating environment variable usage for Azure Voice Live integration, refining agent lifecycle classes to consistently track call connection IDs, and re-enabling automated documentation deployment.

Project Naming and Repository Updates

• Changed project name references from gbb-ai-audio-agent to art-voice-agent-accelerator in both the repository clone instructions and the azure.yaml configuration file for improved branding and clarity. [1] [2]

Environment and Configuration Enhancements

• Updated environment variable names and usage for Azure Voice Live integration in .env.sample and voice_config.py to use AZURE_SPEECH_ENDPOINT and AZURE_SPEECH_KEY, aligning with Azure's recommended naming conventions. Added AI_FOUNDRY_AGENT_ID for AI Foundry integration. [1] [2]
• Ensured BASE_URL in infrastructure configuration strips trailing slashes for consistency in URL handling.

Agent Lifecycle and Handler Improvements

• Refactored agent lifecycle handler classes (ThreadBridge, RouteTurnThread, and related usage) to consistently accept and track call_connection_id, ensuring accurate correlation and logging of call events. [1] [2] [3] [4]
• Removed the demo event processor and FastAPI webhook handler example from events/demo.py to clean up legacy code and reduce maintenance overhead.

Deployment and Documentation Automation

• Re-enabled the deploy job for GitHub Pages in docs.yml, updating the branch condition and ensuring documentation is automatically deployed for main and feature branches. Also added installation of system dependencies required for documentation builds. [1] [2]

Other Minor Updates

• Improved logging messages for Redis initialization to clarify cluster support and retry logic.
• Updated timing logic in AOAI pool initialization for more precise startup metrics and removed commented-out Voice Live pool initialization code. [1] [2]
• Fixed a minor markdown character issue in the workflow README.

[Copilot]

Pull Request Overview

This PR introduces comprehensive infrastructure improvements and testing modernization, primarily focusing on fixing Terraform module lifecycle issues with AI Foundry while standardizing test execution patterns across the codebase.

Key Changes

• Infrastructure Updates: Fixed AI Foundry Terraform module lifecycle issues and updated project naming from gbb-ai-audio-agent to art-voice-agent-accelerator
• Test Modernization: Converted async unit tests to synchronous execution using run_async() helper and eliminated async pytest fixtures
• Redis Enhancements: Added cluster support with automatic failover detection and address remapping for Azure Redis

Reviewed Changes

Copilot reviewed 69 out of 82 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/*.py	Converted async tests to sync execution pattern with run_async() helper
infra/terraform/modules/aifoundry/	Fixed lifecycle issues and simplified resource management
src/redis/manager.py	Added cluster support with MovedError handling and IP address remapping
docs/	Updated documentation structure and repository references
requirements-docs.txt	Added documentation dependencies for enhanced rendering

[Copilot]

The sounddevice module mock should be moved to a centralized test configuration file (like conftest.py) rather than being duplicated in individual test files. This prevents inconsistencies and makes dependency mocking more maintainable across the test suite.

[Copilot]

The run_async() helper function is duplicated across multiple test files. This utility should be moved to a shared test utilities module to avoid code duplication and ensure consistent async test execution patterns.

Suggested change

	def run_async(coro):
	"""Execute coroutine for pytest environments without asyncio plugin."""
	return asyncio.run(coro)
	from tests.utils import run_async

[Copilot]

The direct use of var.foundry_account_name without validation or normalization could cause resource creation issues if the variable contains invalid characters for Azure resource names. Consider adding validation or using the previous local.account_name_raw pattern that included normalization.

[Copilot]

Changing account_principal_ids from list(string) to map(string) is a breaking change that could affect existing Terraform deployments. Consider using a different variable name or maintaining backward compatibility to prevent deployment failures.

Suggested change

	description = "Map of principals to assign Cognitive Services access to the AI Foundry account (keys should be stable labels)."
	description = "List of principal IDs to assign Cognitive Services access to the AI Foundry account (legacy, use account_principal_id_map for new deployments)."
	type = list(string)
	default = []
	}
	variable "account_principal_id_map" {
	description = "Map of principals to assign Cognitive Services access to the AI Foundry account (keys should be stable labels). Use this for new deployments."

[pablosalvador10]

LGTM