JuliaML
Adds support for image generation and responses endpoints
- Responses endpoint support : #68
- Adds support for image generation
Long ago I had added support for image generation to my local branch, but failed to create a PR about it. Then today, I really wanted support for the responses endpoint so I could play around with function calling. So that logic is now added and works for web search, image tagging, vector search, reasoning, and tool calling! Examples included in docstrings.
-Stefan
Fantastic, thanks for doing this! I'll take a look and get it merged this weekend if everything checks out
@stefanjwojcik Looks like the streaming test hangs locally for me, haven't figured out why quite yet. If it doesn't hang for you that would he helpful to know so I can rule some things out
Thanks for testing. Yes, I realized some tests were a little funky - hoping to square those away this weekend.
Hey @roryl23 - only errors I'm getting are coming from the assistants endpoints - looks like the package is still running assistants V1 Beta logic. Updated setup is here: https://platform.openai.com/docs/assistants/migration.
Just adding a bit of context to make it easier to review: streaming now works without error, and all the new functionality vis-a-vis the responses endpoint works in my tests. The issue with the Assistants endpoint is not a result of the changes in this PR, but rather a function of OpenAI changes to that endpoint.
Ah, okay. Perfect, I'll take a look at this and get it merged no later than the weekend
Hi there @roryl23 and @stefanjwojcik Is there a chance this will make it into the next release of the package?
@sabinevidal yes! I'll cut a release this weekend. I just need to resolve the Assistants endpoint issue which is probably just disabling the test, for now at least.