Skip to content

Adjustments and for lab01-lab03 and script adjustments for lab04#29

Merged
madiepev merged 22 commits intoMicrosoftLearning:mainfrom
v-vfarias:main
Feb 27, 2026
Merged

Adjustments and for lab01-lab03 and script adjustments for lab04#29
madiepev merged 22 commits intoMicrosoftLearning:mainfrom
v-vfarias:main

Conversation

@v-vfarias
Copy link
Contributor

Updated env references for trail_guide_agent.py, run_batch_tests.py and interact_with_agent.py.

Updated instructions to add model name and agents name within double quotes.

Updated script evaluate_agent.py to create the .txt file once the evaluation is finished.

Updated the workflow to read the evaluate_results.txt directly from the repo instead of re-running the evaluation (for current testing purposes).

Add agent responses JSON, refactor Bicep resources, enable hosted age…
- Changed the model used in the trail guide agent to be configurable via the MODEL_NAME environment variable, defaulting to "gpt-4.1".
- Updated the dataset name in the evaluation script to "trail-guide-evaluation-dataset_2" and adjusted the dataset path accordingly.
- Removed the `agent-responses.json` file as it is no longer needed.
- Updated Bicep template to deploy model deployments on the AI account.
- Modified `trail_guide_agent.py` to load environment variables directly from the `.env` file.
- Enhanced `evaluate_agent.py` for better error handling and logging, including clearer output for evaluation results.
- Updated dataset names and paths in `evaluate_agent.py` for consistency.
- Improved the structure and readability of the evaluation process in `evaluate_agent.py`.
- Simplified environment variable loading in test scripts (`interact_with_agent.py` and `run_batch_tests.py`).
@madiepev madiepev merged commit 9037e42 into MicrosoftLearning:main Feb 27, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants