You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Sample coverage and quality of sample control (exercise against live service and have samples more interesting than CRUD ops)
We should show case streaming in our samples.
We should show case streaming for an Azure only tool: this will inform what deserialization convenience methods we should add to ResponsesClient/ResponsesAsyncClient
(strongly recommend using relevant SKILL.md for this) We should completely de-duplicate classes from the openai:openai-java package
(strongly recommend using relevant SKILL.md for this) We should add custom getters and setters wrapping types that are ModelA | ModelB in TSP:
I believe Vincent was working on this for C# and Java, so I would recommend reaching out
Union types are surfaced as private BinaryData field. We should have a typed public surface
Double check that wherever BinaryData is exposed to users, it's because we are using unknown in TSP
Enrich the READMEs with hero scenario driven code snippets.
Must do's (or as much as possible)
List sorted in decreasing order of importance:
tool: this will inform what deserialization convenience methods we should add toResponsesClient/ResponsesAsyncClientopenai:openai-javapackageModelA | ModelBin TSP:private BinaryData field. We should have a typed public surfaceBinaryDatais exposed to users, it's because we are usingunknownin TSPREADMEs with hero scenario driven code snippets.Nice to-have's
azure-ai-agentsandazure-ai-projectsHTTPClient integration with OpenAI library #47540Completely optional, but I think it might be a good idea... probably
client.tspfiles so we consume updates applied to other languages that we might otherwise miss.