fix(openai-agents): Patch models functions following library refactor #5449
+134
−135
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
alexander-alderman-webb
changed the title
Contributor
Semver Impact of This PR🟢 Patch (bug fixes) 📋 Changelog PreviewThis is how your changes will appear in the changelog. Bug Fixes 🐛Openai Agents
Other
Documentation 📚
Internal Changes 🔧
🤖 This preview updates automatically when you update the PR. |
Contributor
Codecov Results 📊❌ 25 failed | Total: 25 | Pass Rate: 0% | Execution Time: 1m 54s ❌ Patch coverage is 0.00%. Project has 17827 uncovered lines. Files with missing lines (184)
Generated by Codecov Action |
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
sl0thentr0py
approved these changes
Feb 12, 2026
ericapisani
approved these changes
Feb 12, 2026
Member
ericapisani
left a comment
ericapisani
left a comment
There was a problem hiding this comment.
Small things, but otherwise LGTM 🚀
| @wraps(original_fetch_response) | ||
| async def wrapped_fetch_response(*args: "Any", **kwargs: "Any") -> "Any": | ||
| response = await original_fetch_response(*args, **kwargs) | ||
| if hasattr(response, "model") and response.model: |
Member
There was a problem hiding this comment.
Nitpick: I think an alternative approach that you could take here in order to make this conditional more concise (by removing the and part of the conditional) is to do the following:
Suggested change
| if hasattr(response, "model") and response.model: | |
| if getattr(response, "model", None): |
This would also make things consistent with what you have on line 119 within the wrapped_get_response function.
Contributor
Author
There was a problem hiding this comment.
We need to rethink this response model business because we probably shouldn't be setting response models on agent spans either.
This was put in before I knew as much about agents, but there are libraries that let you vary the request model during agent execution as well (in particular Claude Code wrapped by the Claude Agent SDK).
Created an issue to track #5458
(and the plan would be to clean up as part of that issue).
Comment on lines
+136
to
+137
| # Uses explicit try/finally instead of context manager to ensure cleanup | ||
| # even if the consumer abandons the stream (GeneratorExit). |
Member
There was a problem hiding this comment.
This comment looks like it needs to be updated or removed as there's no try/finally below and the code uses a context manager 😅
Contributor
Author
There was a problem hiding this comment.
you're completely right, and it doesn't look like we catch GeneratorExit anywhere for that matter 😅 .
I'll keep this commit atomic since we're moving the contents of wrapped_get_model() to _get_model() as is now, and follow up with a PR after this first train is merged 🚅
Comment on lines
+182
to
+183
| if hasattr(streaming_response, "model") | ||
| and streaming_response.model |
Member
There was a problem hiding this comment.
Also a nitpick but similar to the comment I left above
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.
| ) -> "agents.Model": | ||
| return _get_model(turn_preparation.get_model, agent, run_config) | ||
|
|
||
| agents.run_internal.run_loop.get_model = new_wrapped_get_model |
There was a problem hiding this comment.
Model patch applied to wrong module
Medium Severity
In the v0.8+ branch, the wrapper is built from turn_preparation.get_model but assigned to agents.run_internal.run_loop.get_model. This leaves turn_preparation.get_model unpatched, so the new model path can bypass _get_model instrumentation in patches/models.py.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
The
AgentRunner._get_model()class method has been moved to the functionagents.run_internal.turn_preparation.get_model().Patch the new function by re-using existing wrapper logic.
Issues
Reminders
tox -e linters.feat:,fix:,ref:,meta:)