[Azure AI Content Safety ] API Review

[azure-sdk]

New API Review meeting has been requested.

Service Name: Azure AI Content Safety Review Created By: Meng Ai Review Date: 03/28/2024 08:00 AM PT Release Plan: PR: Hero Scenarios Link: Not Provided Core Concepts Doc Link: Not Provided

Description:

Detailed meeting information and documents provided can be accessed here For more information that will help prepare you for this review, the requirements, and office hours, visit the documentation here

[mikekistler]

Notes from REST API Review 03/28/24

• Same concern with versioning of custom categories as we discussed previously for automaticReviewers
  • We should definitely avoid "automatic" versioning. If we did versioning it should be done with a "/versions" and versionId in the path
  • Managing this is going to be complicated for customers to understand and for the service to implement
  • Johan says Form Recognizer has some versioning support; Also OpenAI
• The simplest thing is to allow customers to give them names, and let users tell us why they need different versions with the same name.
• Just have one concept "category" and allow the user to create, update, delete and inference using this category
• Consider adding a "display name" that would give the category name for a model.
  • Inference would specify the model name(s)
• Why not use the approved status values for the category status field?
  • You can have any additional non-terminal states
  • "BuildSucceeded" -> "Succeeded", "BuildFailed" -> "Failed"
• Shouldn't return "DeleteInfo" -- if the delete fails then there's a bug in the service.
  • Hide this from the customer. (Try as hard as you can to hide it)

Looks good for preview but please address these concerns as soon as practical and definitely before GA.