refactor(providers): use extractReasoningFromDelta helper for reasoni…#588
Merged
edelauna merged 5 commits intoJun 15, 2026
Merged
Conversation
…ng streams Replace duplicated reasoning_content extraction logic across 7 providers (deepseek, mimo, openai, opencode-go, qwen-code, requesty, unbound) with the shared extractReasoningFromDelta helper. This also adds the OpenRouter- style 'reasoning' field fallback to each provider for free.
daewoongoh
requested review from
JamesRobert20,
edelauna,
hannesrudolph,
navedmerchant and
taltas
as code owners
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (7)
🚧 Files skipped from review as they are similar to previous changes (3)
📝 WalkthroughWalkthroughProvider streaming handlers (DeepSeek, MiMo, OpenAI, Opencode-Go, Qwen-Code, Requesty, Unbound) now import and use ChangesConsolidate Reasoning Extraction Helper Adoption
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~28 minutes Possibly related issues
Possibly related PRs
Suggested reviewers
Poem
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Codecov Report✅ All modified and coverable lines are covered by tests. 📢 Thoughts on this report? Let us know! |
Add streaming tests for openai, requesty, and unbound covering: - delta.reasoning_content yields a reasoning chunk - delta.reasoning fallback yields a reasoning chunk when reasoning_content is absent (OpenRouter-style) This raises line coverage on the changed reasoning paths from 50%.
edelauna
reviewed
Jun 13, 2026
edelauna
left a comment
edelauna
left a comment
Contributor
There was a problem hiding this comment.
Thanks for adding this, had one question related to test coverage
Comment on lines
+242
to
+259
| it("falls back to delta.reasoning when reasoning_content is absent", async () => { | ||
| mockCreate.mockImplementationOnce(async () => ({ | ||
| [Symbol.asyncIterator]: async function* () { | ||
| yield { choices: [{ delta: { reasoning: "router-style thought" }, index: 0 }] } | ||
| yield { | ||
| choices: [{ delta: {}, index: 0 }], | ||
| usage: { prompt_tokens: 1, completion_tokens: 1, total_tokens: 2 }, | ||
| } | ||
| }, | ||
| })) | ||
|
|
||
| const chunks: any[] = [] | ||
| for await (const chunk of handler.createMessage(systemPrompt, messages)) { | ||
| chunks.push(chunk) | ||
| } | ||
|
|
||
| expect(chunks).toContainEqual({ type: "reasoning", text: "router-style thought" }) | ||
| }) |
Contributor
There was a problem hiding this comment.
Does the helper priority rule get exercised here? If a delta has both reasoning_content and reasoning populated, the expected winner is reasoning_content — but that case isn't covered at the provider level. The unit tests for extractReasoningFromDelta cover it, but a provider-level check would catch a future regression where the helper call gets bypassed or re-inlined.
Contributor
Author
There was a problem hiding this comment.
Good catch. I added a provider-level priority test to cover the case where both delta.reasoning_content and delta.reasoning are present. The test asserts that only reasoning_content is emitted as the reasoning chunk, so it should catch regressions where the helper is bypassed or the logic is re-inlined incorrectly.
…imo, opencode-go, qwen-code
edelauna
approved these changes
Jun 15, 2026
edelauna
left a comment
edelauna
left a comment
Contributor
There was a problem hiding this comment.
Thank you for taking on this cleanup.
edelauna
pushed a commit
that referenced
this pull request
Jun 17, 2026
#588) * refactor(providers): use extractReasoningFromDelta helper for reasoning streams Replace duplicated reasoning_content extraction logic across 7 providers (deepseek, mimo, openai, opencode-go, qwen-code, requesty, unbound) with the shared extractReasoningFromDelta helper. This also adds the OpenRouter- style 'reasoning' field fallback to each provider for free. * test(providers): cover reasoning_content + reasoning fallback streaming Add streaming tests for openai, requesty, and unbound covering: - delta.reasoning_content yields a reasoning chunk - delta.reasoning fallback yields a reasoning chunk when reasoning_content is absent (OpenRouter-style) This raises line coverage on the changed reasoning paths from 50%. * test(providers): add reasoning_content streaming tests to deepseek, mimo, opencode-go, qwen-code --------- Co-authored-by: Oh Daewoong <dw.oh@samsung.com>
7 tasks
nigeldelviero
pushed a commit
to nigeldelviero/Zoo-Code
that referenced
this pull request
Jun 22, 2026
Zoo-Code-Org#588) * refactor(providers): use extractReasoningFromDelta helper for reasoning streams Replace duplicated reasoning_content extraction logic across 7 providers (deepseek, mimo, openai, opencode-go, qwen-code, requesty, unbound) with the shared extractReasoningFromDelta helper. This also adds the OpenRouter- style 'reasoning' field fallback to each provider for free. * test(providers): cover reasoning_content + reasoning fallback streaming Add streaming tests for openai, requesty, and unbound covering: - delta.reasoning_content yields a reasoning chunk - delta.reasoning fallback yields a reasoning chunk when reasoning_content is absent (OpenRouter-style) This raises line coverage on the changed reasoning paths from 50%. * test(providers): add reasoning_content streaming tests to deepseek, mimo, opencode-go, qwen-code --------- Co-authored-by: Oh Daewoong <dw.oh@samsung.com>
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
Related GitHub Issue
#570
Description
This PR replaces duplicated inline reasoning-delta extraction logic across seven providers with the shared
extractReasoningFromDeltahelper.Affected providers:
deepseekmimoopenaiopencode-goqwen-coderequestyunboundPreviously, each provider had a near-identical streaming check for
delta.reasoning_contentand emitted reasoning chunks inline. Sincesrc/api/providers/utils/extract-reasoning.tsalready provides the canonical extraction behavior and is already used by other OpenAI-compatible providers, this PR brings the remaining duplicated call sites in line with the shared implementation.Test Procedure
Pre-Submission Checklist
Existing tests cover the changed provider behavior, and the shared helper is already test-covered.
Screenshots / Videos
N/A
Documentation Updates
Additional Notes
This is a focused provider refactor. It removes duplicated reasoning extraction logic and routes all seven affected providers through the shared
extractReasoningFromDeltahelper.The intended behavior for existing
reasoning_contentchunks is unchanged for normal string payloads. The practical improvement is that these providers now also handle thereasoningfallback supported by the helper, so future or model-specific deltas using that field will stream reasoning text instead of being ignored.Get in Touch
hehegwk_23849
Summary by CodeRabbit
reasoningchunks using a shared approach (including correct field precedence and fallback).