This AI sample lacks risk & safety evaluation implementation

[carlotta94c]

As per email sent on 13th Dec with same subject, please double-check that this AI sample implements evaluations. In particular, what we are looking for is:

• Evaluation file(s) (might be a Jupyter notebook, a unit test script, etc.) that evaluates the solution against quality metrics
• Evaluation file(s) (might be a Jupyter notebook, a unit test script, etc.) that evaluates the solution against at least 2 safety metrics
• A descriptive section in your readme explaining how evaluation is implemented into the sample.

[pamelafox]

Step 1) Provision an AI project. See infra in https://github.com/Azure-Samples/azureai-basic-python/tree/main/infra

Step 2) Generate simulated QA pairs from the app using the adversial simulator.

See callback code in https://github.com/Azure-Samples/azure-search-openai-demo/pull/2233/files#diff-8da4e313bfcffa446d60e3717caf532fa6344fdf6c210d869717088828197476

https://learn.microsoft.com/en-us/azure/ai-studio/how-to/develop/simulator-interaction-data#run-an-adversarial-simulation

Step 3) Run the safety evaluators against the simulated answers:
https://learn.microsoft.com/en-us/azure/ai-studio/how-to/develop/evaluate-sdk#risk-and-safety-evaluators

[pamelafox]

You can use this endpoint:
https://pf-ragpostgres-5lf2-ca.wittysmoke-7b396797.eastus2.azurecontainerapps.io

[pamelafox]

cc @alfredodeza

[pamelafox]

All done!