2026 01 27 fix tutorial cloud evaluations implementation

[jonburchel]

This pull request significantly refactors the evaluation script for the Modern Workplace Assistant sample, transitioning it from a custom, manual evaluation workflow to a streamlined, cloud-based evaluation using the Azure AI Projects SDK and built-in evaluators. The update also aligns the requirements with stable SDK versions.

Key changes include:

Migration to Cloud Evaluation Workflow:

• The evaluate.py script is completely rewritten to use Azure AI Projects and OpenAI cloud evaluation APIs, replacing all custom validation, test data loading, and result calculation logic with standardized, cloud-based evaluation flows.

Simplification and Modernization:

• Removed all legacy code for loading test questions, validating responses, batch evaluation, and manual result reporting; now the script programmatically creates agents, defines evaluation criteria, runs cloud-based evaluations, and retrieves results.
• Updated the script to use environment variables for configuration and added clear usage instructions.

Requirements Update:

• Updated requirements.txt to pin azure-ai-projects to version 1.0.0 and azure-ai-agents to 1.2.0b6, ensuring compatibility with the new evaluation workflow.

[sdgilley]

⚠️ Manual review required

This PR requires manual review due to the following issues:

• Modified file has deleted cells: samples/python/enterprise-agent-tutorial/1-idea-to-prototype/evaluate.py

Please review and address these issues before merging.