Handle token usage chunks in OpenAI streamed response

[timtimjnvr]

This pull request addresses and resolves a case related to issue #28850.

During chat completion requests streamed from OpenAI models hosted on OpenWebUI, the API can add token usage chunks in the response data.

This update modifies the handling of these responses to account for token usage chunks, preventing potential errors when these are encountered.

Release Notes

• Fixed Open Web UI compatibility with Zed when Open Web UI is sending usage stats chuncks in streamed response.

[cla-bot]

Thank you for your pull request and welcome to our community. We could not parse the GitHub identity of the following contributors: Timothée JANVIER.
This is most likely caused by a git client misconfiguration; please make sure to:

1. check if your git client is configured with an email to sign commits git config --list | grep email
2. If not, set it up using git config --global user.email email@example.com
3. Make sure that the git commit email is configured in your GitHub account settings, see https://github.com/settings/emails

[cla-bot]

We require contributors to sign our Contributor License Agreement, and we don't have @timtimjnvr on file. You can sign our CLA at https://zed.dev/cla. Once you've signed, post a comment here that says '@cla-bot check'.

[timtimjnvr]

@cla-bot check

[cla-bot]

The cla-bot has been summoned, and re-checked this pull request!

[zed-industries-bot]

	Warnings
⚠️	fix: handle token usage chuncks in open ai streamed response ^ Write PR titles using sentence case.
⚠️	This PR is missing release notes. Please add a "Release Notes" section that describes the change: Release Notes: - Added/Fixed/Improved ... If your change is not user-facing, you can use "N/A" for the entry: Release Notes: - N/A

Have feedback on this plugin? Let's hear it!

Generated by 🚫 dangerJS against dbb2873

[imumesh18]

This will create an issue in case of the stop event as then also we get the usage data. Now the problem is that we are just returning after usage data. Now this will break the whole event chain and the thread will be in broken state. You might wanna move this to else condition of choices check.

[timtimjnvr]

In the documentation I understand that the stop event ("finished_reason": "stop") is a field of a choice inside the choices array of the chunck object (https://platform.openai.com/docs/api-reference/chat-streaming/.

usage field seems to be sent in a different object chunck with empty choices:

"choices: A list of chat completion choices. Can contain more than one elements if n is greater than 1. Can also be empty for the last chunk if you set stream_options: {"include_usage": true}." (source)

However your proposal if some implementation have both fields (choices & usage) returned in the same event.

Thanks for the feedback 🙌

[imumesh18]

I believe there was another pr merged yesterday which fixes this: #32982