fix(convert): forward Claude output_config.effort as reasoning_effort to OpenAI-format upstreams

[fengshao1227]

⚠️ 提交说明 / PR Notice

Important

• 本描述为人工整理的简洁摘要，未直接粘贴未经整理的 AI 输出。
• AI 辅助声明：本 PR 的代码与描述由 AI（Claude Code）辅助生成，提交者 git 身份不属于仓库历史核心开发者。

📝 变更描述 / Description

修复 #5922：Claude 格式客户端设置 reasoning effort（请求体中的 output_config.effort，如 high）后，经 OpenAI 类型渠道转发时该参数被静默丢弃，上游收不到 reasoning_effort。

根因：service/convert.go 的 ClaudeToOpenAIRequest 中，claudeRequest.GetEfforts()（解析 output_config.effort）只在 OpenRouter 分支被消费；非 OpenRouter 的通用分支（即 OpenAI 类型渠道走的路径）完全没有读取 effort，转换后的 GeneralOpenAIRequest.ReasoningEffort 恒为空。

修复：在通用分支中把 effort 赋给 openAIRequest.ReasoningEffort（一行改动）。之所以只需要改这一处：OpenAI 渠道 adaptor 的 ConvertClaudeRequest 会把转换结果继续送入 ConvertOpenAIRequest（relay/channel/openai/adaptor.go），后者对 request.ReasoningEffort 已有完整的下游处理链（o 系列 / gpt-5 的模型后缀与参数适配、OpenRouter 的 reasoning JSON 映射、info.ReasoningEffort 日志记录），本次填上字段后即可复用全部既有逻辑。OpenRouter 分支的 effort 映射行为保持不变。

🚀 变更类型 / Type of change

• 🐛 Bug 修复 (Bug fix) - 请关联对应 Issue，避免将设计取舍、理解偏差或预期不一致直接归类为 bug
• ✨ 新功能 (New feature) - 重大特性建议先通过 Issue 沟通
• ⚡ 性能优化 / 重构 (Refactor)
• 📝 文档更新 (Documentation)

🔗 关联任务 / Related Issue

• Closes [Bug] reasoning_effort=high from Claude client is not forwarded to upstream OpenAI channel #5922

✅ 提交前检查项 / Checklist

• 人工确认: 我已亲自整理并撰写此描述，没有直接粘贴未经处理的 AI 输出。
• 非重复提交: 我已搜索现有的 Issues 与 PRs，确认不是重复提交。
• Bug fix 说明: 若此 PR 标记为 Bug fix，我已提交或关联对应 Issue，且不会将设计取舍、预期不一致或理解偏差直接归类为 bug。
• 变更理解: 我已理解这些更改的工作原理及可能影响。
• 范围聚焦: 本 PR 未包含任何与当前任务无关的代码改动。
• 本地验证: 已在本地运行并通过测试或手动验证，维护者可以据此复核结果。
• 安全合规: 代码中无敏感凭据，且符合项目代码规范。

📸 运行证明 / Proof of Work

新增回归测试 service/convert_effort_test.go（testify，三个场景）：

=== RUN   TestClaudeToOpenAIRequestForwardsEffortAsReasoningEffort
--- PASS: TestClaudeToOpenAIRequestForwardsEffortAsReasoningEffort (0.00s)
=== RUN   TestClaudeToOpenAIRequestWithoutEffortLeavesReasoningEffortEmpty
--- PASS: TestClaudeToOpenAIRequestWithoutEffortLeavesReasoningEffortEmpty (0.00s)
=== RUN   TestClaudeToOpenAIRequestOpenRouterEffortMappingUnchanged
--- PASS: TestClaudeToOpenAIRequestOpenRouterEffortMappingUnchanged (0.00s)
PASS
ok      github.com/QuantumNous/new-api/service  0.025s

• OpenAI 渠道 + output_config: {"effort": "high"} → 转换结果含 "reasoning_effort":"high"（对应 issue 复现场景）。
• 无 output_config → reasoning_effort 不出现（omitempty，不污染无推理需求的请求）。
• OpenRouter 渠道 → effort 仍走原有 Verbosity 映射，不会重复产生 reasoning_effort 字段。

全量回归：go test ./service/... ./relay/... 全部通过，go vet ./service/ 无告警。

Summary by CodeRabbit

• Bug Fixes

  • Request conversions now carry effort settings through correctly for standard channels, so upstream requests can include the expected reasoning level.
  • Requests without an effort setting continue to omit that field.
  • OpenRouter behavior remains unchanged, preserving the existing effort-to-verbosity handling.

• Tests

  • Added regression coverage for effort forwarding, missing-effort behavior, and OpenRouter compatibility.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 336dc72d-ccda-46d0-9a76-a4ab72df07dd

📥 Commits

Reviewing files that changed from the base of the PR and between 8bc4bf1 and be30703.

📒 Files selected for processing (2)

• service/convert.go
• service/convert_effort_test.go

---

Walkthrough

ClaudeToOpenAIRequest now maps non-empty Claude effort values to ReasoningEffort for non-OpenRouter channels, fixing a gap where this mapping previously only applied to OpenRouter via Verbosity. Three new tests validate forwarding, absence handling, and unchanged OpenRouter behavior.

Changes

Effort Mapping Fix and Tests

Layer / File(s)	Summary
Map Claude effort to ReasoningEffort service/convert.go	Non-OpenRouter branch of ClaudeToOpenAIRequest now sets openAIRequest.ReasoningEffort from claudeRequest.GetEfforts() when non-empty.
Regression tests for effort forwarding service/convert_effort_test.go	Adds tests verifying effort is forwarded as reasoning_effort for OpenAI channels, remains absent when unset, and OpenRouter's Verbosity mapping stays unchanged.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related PRs

• QuantumNous/new-api#1577: Modifies how ReasoningEffort is handled for OpenRouter vs non-OpenRouter requests, directly related to this mapping fix.
• QuantumNous/new-api#3120: Introduces/uses ClaudeRequest.GetEfforts() and updates service/convert.go to map Claude effort into OpenAI reasoning fields.
• QuantumNous/new-api#2863: Adds logic to derive/set effort in Claude requests via OutputConfig, which feeds into the effort forwarded by this PR.

Suggested reviewers: seefs001

Poem

A hop, a skip, an effort set free,
"high" now travels from Claude to me,
ReasoningEffort finally in place,
No more lost fields in outer space,
Tests confirm it — carrot in hand! 🥕

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly states the main fix: forwarding Claude effort to OpenAI-format upstreams.
Linked Issues check	✅ Passed	The change forwards Claude reasoning effort to non-OpenRouter OpenAI requests and preserves OpenRouter behavior, matching #5922.
Out of Scope Changes check	✅ Passed	The added test coverage and code changes stay within the effort-forwarding bug fix and don't introduce unrelated behavior.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

Warning

Tools execution failed with the following error:

Failed to run tools: 13 INTERNAL: Received RST_STREAM with code 2 (Internal server error)

---

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.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}