feat(responses): add cancel() support for streaming responses

[giulio-leone]

Summary

Adds the ability to cancel streaming responses mid-stream via a new cancel() method on ResponseStream and AsyncResponseStream.

Fixes #2643

Changes

• Added response_id property to ResponseStream and AsyncResponseStream — returns the response ID from the accumulated snapshot (available after the first response.created event)
• Added cancel() method to ResponseStream and async cancel() to AsyncResponseStream — calls the cancel API endpoint and closes the stream
• Added current_snapshot property to ResponseStreamState to cleanly expose the internal snapshot
• Passed a cancel_response callback from Responses.stream() and AsyncResponses.stream() to the stream managers and stream classes, using a callback pattern to avoid coupling stream classes to the resource layer
• Added 11 unit tests covering: response_id availability, error cases, callback invocation, and stream closure

Usage

# Sync
with client.responses.stream(model="gpt-4o", input="Write a long essay") as stream:
    for event in stream:
        if should_stop(event):
            cancelled = stream.cancel()
            break

# Async
async with client.responses.stream(model="gpt-4o", input="Write a long essay") as stream:
    async for event in stream:
        if should_stop(event):
            cancelled = await stream.cancel()
            break

Backwards Compatibility

• All new parameters use None defaults — fully backwards compatible
• No changes to existing method signatures or behavior

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: b77dd8cfd6

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[Copilot]

Pull request overview

Adds first-class cancellation support for streaming Responses by exposing response_id and a new cancel() API on both sync and async streaming helpers, wired through the Responses.stream() / AsyncResponses.stream() managers.

Changes:

• Add response_id property on ResponseStream / AsyncResponseStream (backed by a new ResponseStreamState.current_snapshot accessor).
• Add cancel() / async cancel() on stream objects that invokes the Responses cancel endpoint via a passed callback and closes the stream.
• Add unit tests covering response_id, cancel preconditions, callback invocation, and snapshot exposure.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 3 comments.

File	Description
src/openai/lib/streaming/responses/_responses.py	Adds cancel callback plumbing, response_id, cancel() methods, and exposes current_snapshot.
src/openai/resources/responses/responses.py	Wires self.cancel into stream managers so streaming objects can call the cancel endpoint.
tests/lib/responses/test_response_stream_cancel.py	Adds unit tests for cancellation behavior and snapshot/ID availability.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[giulio-leone]

Intervention note for this PR:

Current blocker appears to be missing CI execution rather than failing jobs:

• No checks reported on this branch.

Suggested unblock sequence:

1. Trigger CI workflows for this PR (or confirm why workflows are not attached for this branch/fork).
2. Once checks appear, fix any failing jobs and re-run.
3. Request review only after a green check suite is visible.

If useful, I can run a follow-up status sweep as soon as checks are attached.

[giulio-leone]

This PR is ready for review — all CI checks pass, no merge conflicts, and all review threads have been resolved. Ready to merge when approved. 🚀

[giulio-leone]

Hi! 👋 Gentle ping — this PR is rebased, CI passes, and ready for review. Happy to address any feedback. Thanks!

[giulio-leone]

Hi! Gentle ping — this PR is rebased, CI passes, and ready for review. Happy to address any feedback. Thanks!