feat: enable thinking mode on ali thinking model

[feitianbubu]

渠道测试支持qwen3-thinking模型
docs: https://bailian.console.aliyun.com/?tab=api#/api/?type=model&url=2712216
qwen3-thinking需要EnableThinking, 并且支持流模式
否则报错: InternalError.Algo.InvalidParameter: The value of the enable_thinking parameter is restricted to True.

修复前:

修复后:

Summary by CodeRabbit

• Bug Fixes

  • Ensured the streaming status is accurately reflected in channel test logs.
  • Fixed potential issues with unset start times by ensuring a valid timestamp is always set.

• New Features

  • Automatically enables "thinking" mode and streaming for models containing "thinking" in their name, complying with API requirements.

[coderabbitai]

Walkthrough

The changes introduce dynamic handling for the IsStream field in channel testing, enforce specific flags for models containing "thinking" in the Ali channel adaptor, and ensure a valid startTime is always set in relay info generation. No public interfaces or exported entity signatures are altered.

Changes

Cohort / File(s)	Change Summary
Channel Test Stream Handling controller/channel-test.go	Sets IsStream in RecordConsumeLog dynamically using info.IsStream instead of hardcoding to false.
Ali Channel "Thinking" Model Logic relay/channel/ali/adaptor.go	Adds logic to detect "thinking" in model names; sets EnableThinking and streaming flags accordingly; adjusts imports.
Relay Info Start Time Enforcement relay/common/relay_info.go	Ensures startTime is set to current time if unset during relay info generation.

Sequence Diagram(s)

sequenceDiagram
    participant Client
    participant AliAdaptor
    participant RelayInfo

    Client->>AliAdaptor: ConvertOpenAIRequest(request)
    AliAdaptor->>AliAdaptor: Check if model contains "thinking"
    alt Model contains "thinking"
        AliAdaptor->>AliAdaptor: Set EnableThinking = true
        AliAdaptor->>AliAdaptor: Set Stream = true, info.IsStream = true
    else Model does not contain "thinking"
        AliAdaptor->>AliAdaptor: Normal logic
    end
    AliAdaptor->>RelayInfo: GenRelayInfo(ctx)
    RelayInfo->>RelayInfo: If startTime == 0, set to time.Now()

Loading

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~7 minutes

Poem

A rabbit peers at code anew,
Where streams now flow as they ought to do.
"Thinking" models hop in line,
With flags and times all set just fine.
Each field now true, each time now set—
Bugs and errors? None to fret!
🐇✨

Note

⚡️ Unit Test Generation is now available in beta!

Learn more here, or try it out under "Finishing Touches" below.

---

📜 Recent review details

Configuration used: CodeRabbit UI
Review profile: CHILL
Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 0c0caad and 38067f1.

📒 Files selected for processing (3)

• controller/channel-test.go (1 hunks)
• relay/channel/ali/adaptor.go (3 hunks)
• relay/common/relay_info.go (1 hunks)

🧰 Additional context used

🧠 Learnings (4)

📓 Common learnings

Learnt from: 9Ninety
PR: QuantumNous/new-api#1273
File: relay/channel/gemini/relay-gemini.go:97-116
Timestamp: 2025-06-21T03:37:41.726Z
Learning: In relay/channel/gemini/relay-gemini.go, the thinking budget calculation logic (including the MaxOutputTokens multiplication) was introduced in PR #1247. PR #1273 focused specifically on decoupling the thoughts summary feature from thinking budget settings and did not modify the existing thinking budget behavior.

📚 Learning: claude models handle "-thinking" variants differently from gemini models. for claude models, only th...

Learnt from: neotf
PR: QuantumNous/new-api#1511
File: setting/ratio_setting/model_ratio.go:118-123
Timestamp: 2025-08-05T17:14:17.246Z
Learning: Claude models handle "-thinking" variants differently from Gemini models. For Claude models, only the base model (without "-thinking") gets an entry in defaultModelRatio map. The "-thinking" variants rely on the Claude relay handler stripping the suffix using strings.TrimSuffix(textRequest.Model, "-thinking") before looking up the ratio, so they automatically use the base model's ratio.

Applied to files:

• relay/channel/ali/adaptor.go

📚 Learning: in relay/channel/gemini/relay-gemini.go, the thinking budget calculation logic (including the maxout...

Learnt from: 9Ninety
PR: QuantumNous/new-api#1273
File: relay/channel/gemini/relay-gemini.go:97-116
Timestamp: 2025-06-21T03:37:41.726Z
Learning: In relay/channel/gemini/relay-gemini.go, the thinking budget calculation logic (including the MaxOutputTokens multiplication) was introduced in PR #1247. PR #1273 focused specifically on decoupling the thoughts summary feature from thinking budget settings and did not modify the existing thinking budget behavior.

Applied to files:

• relay/channel/ali/adaptor.go

📚 Learning: gin.context implements context.context interface since gin v1.8.0, providing the methods deadline(),...

Learnt from: feitianbubu
PR: QuantumNous/new-api#1228
File: router/main.go:28-36
Timestamp: 2025-06-15T12:38:11.806Z
Learning: gin.Context implements context.Context interface since Gin v1.8.0, providing the methods Deadline(), Done(), Err(), and Value(). When using Gin v1.8.0 or later, gin.Context can be passed directly to functions expecting context.Context without needing to extract c.Request.Context().

Applied to files:

• relay/channel/ali/adaptor.go
• relay/common/relay_info.go

🔇 Additional comments (4)

relay/common/relay_info.go (1)

228-230: LGTM! Good defensive programming practice.

This change ensures that startTime is never zero by setting it to the current time if unset, providing reliable timing information for downstream processing. This aligns well with the overall PR changes that require consistent timing data for streaming operations.

relay/channel/ali/adaptor.go (2)

68-74: LGTM! Core functionality correctly implements the thinking model requirements.

This change properly addresses the qwen3-thinking model requirements by:

1. Enabling EnableThinking parameter when model name contains "thinking"
2. Forcing streaming mode as required by the API
3. Updating the relay info to reflect streaming state

The implementation follows the documented API requirements and resolves the specific error mentioned in the PR objectives.

---

6-6: Import organization looks good.

Added strings package import needed for the new thinking model detection logic, and reorganized imports appropriately.

Also applies to: 15-15

controller/channel-test.go (1)

278-278: Excellent consistency improvement!

Changing from hardcoded false to info.IsStream ensures that the consumption log accurately reflects the actual streaming state, especially important now that streaming can be dynamically enabled for thinking models in the Ali adaptor.

✨ Finishing Touches

• 📝 Generate Docstrings

🧪 Generate unit tests

• Create PR with unit tests
• Post copyable unit tests in a comment

---

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

🪧 Tips

Chat

There are 3 ways to chat with CodeRabbit:

• Review comments: Directly reply to a review comment made by CodeRabbit. Example:
  • I pushed a fix in commit <commit_id>, please review it.
  • Explain this complex logic.
  • Open a follow-up GitHub issue for this discussion.
• Files and specific lines of code (under the "Files changed" tab): Tag @coderabbitai in a new review comment at the desired location with your query. Examples:
  • @coderabbitai explain this code block.
• PR comments: Tag @coderabbitai in a new PR comment to ask questions about the PR branch. For the best results, please provide a very specific query, as very limited context is provided in this mode. Examples:
  • @coderabbitai gather interesting stats about this repository and render them as a table. Additionally, render a pie chart showing the language distribution in the codebase.
  • @coderabbitai read src/utils.ts and explain its main purpose.
  • @coderabbitai read the files in the src/scheduler package and generate a class diagram using mermaid and a README in the markdown format.

Support

Need help? Create a ticket on our support page for assistance with any issues or questions.

CodeRabbit Commands (Invoked using PR comments)

• @coderabbitai pause to pause the reviews on a PR.
• @coderabbitai resume to resume the paused reviews.
• @coderabbitai review to trigger an incremental review. This is useful when automatic reviews are disabled for the repository.
• @coderabbitai full review to do a full review from scratch and review all the files again.
• @coderabbitai summary to regenerate the summary of the PR.
• @coderabbitai generate docstrings to generate docstrings for this PR.
• @coderabbitai generate sequence diagram to generate a sequence diagram of the changes in this PR.
• @coderabbitai generate unit tests to generate unit tests for this PR.
• @coderabbitai resolve resolve all the CodeRabbit review comments.
• @coderabbitai configuration to show the current CodeRabbit configuration for the repository.
• @coderabbitai help to get help.

Other keywords and placeholders

• Add @coderabbitai ignore anywhere in the PR description to prevent this PR from being reviewed.
• Add @coderabbitai summary to generate the high-level summary at a specific location in the PR description.
• Add @coderabbitai anywhere in the PR title to generate the title automatically.

CodeRabbit Configuration File (.coderabbit.yaml)

• You can programmatically configure CodeRabbit by adding a .coderabbit.yaml file to the root of your repository.
• Please see the configuration documentation for more information.
• If your editor has YAML language server enabled, you can add the path at the top of this file to enable auto-completion and validation: # yaml-language-server: $schema=https://coderabbit.ai/integrations/schema.v2.json

Documentation and Community

• Visit our Documentation for detailed information on how to use CodeRabbit.
• Join our Discord Community to get help, request features, and share feedback.
• Follow us on X/Twitter for updates and announcements.

[JrantKing]

希望能支持魔搭渠道的qwen3-thinking模型，魔搭渠道的设置和阿里是一样的，但是不能用阿里的供应商设置魔搭的地址，所以有一样的问题