Add MiniMax M3 provider metadata#276
Open
octo-patch wants to merge 1 commit into
Open
Conversation
📝 WalkthroughWalkthroughMiniMax 集成将默认模型从 M2.7 切换为 M3，扩展模型元数据和区域端点文档，更新 API/OAuth 入站配置，并新增 M3 thinking 流载荷兼容处理及测试。 ChangesMiniMax M3 支持
Estimated code review effort: 3 (Moderate) | ~20 minutes Sequence Diagram(s)sequenceDiagram
participant EmbeddedRunner
participant M3ThinkingWrapper
participant MiniMaxAnthropicStream
EmbeddedRunner->>M3ThinkingWrapper: 包装 MiniMax M3 的 streamFn
M3ThinkingWrapper->>MiniMaxAnthropicStream: 转发流请求
MiniMaxAnthropicStream-->>M3ThinkingWrapper: 返回 thinking.type=enabled
M3ThinkingWrapper-->>EmbeddedRunner: 返回 thinking.type=adaptive
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (3)
src/agents/pi-embedded-runner/extra-params.ts (1)
371-378: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value建议添加 debug 日志以保持与其他条件包装器的一致性。
同函数中大多数条件包装器（如 SiliconFlow L331-333、Bedrock L361-363、Anthropic fast mode L382-384、MiniMax fast mode L389-391）都包含
log.debug调用。M3 thinking 包装器是条件性的且与 provider/modelId 相关，添加日志有助于排查 M3 兼容包装是否被正确应用。♻️ 建议添加 debug 日志
if ( shouldApplyMinimaxM3ThinkingCompat({ provider: ctx.provider, modelId: ctx.modelId, }) ) { + log.debug( + `applying MiniMax M3 thinking compat wrapper for ${ctx.provider}/${ctx.modelId}`, + ); ctx.agent.streamFn = createMinimaxM3ThinkingWrapper(ctx.agent.streamFn); }🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/agents/pi-embedded-runner/extra-params.ts` around lines 371 - 378, 在应用 createMinimaxM3ThinkingWrapper 的条件分支中添加与其他条件包装器一致的 log.debug 调用，记录 provider、modelId 及 M3 thinking 兼容包装已应用；保持现有 shouldApplyMinimaxM3ThinkingCompat 判断和包装逻辑不变。src/agents/pi-embedded-runner/minimax-stream-wrappers.ts (1)
29-39: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value建议添加简要注释说明 thinking 归一化逻辑。
{ type: "enabled", budget_tokens }被替换为{ type: "adaptive" }时会丢弃budget_tokens等额外字段。虽然测试确认了该行为，但此处属于不太直观的兼容性改写，添加一行注释有助于后续维护者理解为何要丢弃budget_tokens。As per coding guidelines:
**/*.{ts,tsx}: Add brief code comments for tricky or non-obvious logic♻️ 建议添加注释
return streamWithPayloadPatch(underlying, model, context, options, (payloadObj) => { const thinking = payloadObj.thinking; if ( thinking && typeof thinking === "object" && !Array.isArray(thinking) && (thinking as Record<string, unknown>).type === "enabled" ) { + // MiniMax M3 uses "adaptive" thinking; drop budget_tokens since adaptive mode manages it automatically. payloadObj.thinking = { type: "adaptive" }; } });🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/agents/pi-embedded-runner/minimax-stream-wrappers.ts` around lines 29 - 39, 在 streamWithPayloadPatch 的 payloadObj 处理回调中，为将 thinking 的 enabled 形式归一化为 adaptive 的赋值逻辑添加一行简短注释，明确说明该兼容性改写会丢弃 budget_tokens 等额外字段；不要修改现有转换行为。Source: Coding guidelines
src/agents/pi-embedded-runner/minimax-stream-wrappers.test.ts (1)
25-60: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win建议补充关键边界测试用例。
当前测试覆盖了主要路径，但缺少以下边界场景：
- 非 anthropic-messages API 的 M3 模型：
createMinimaxM3ThinkingWrapper在 L23 有model.api !== "anthropic-messages"守卫，但无测试覆盖。如果该守卫被意外移除，测试无法捕获。- 非 minimax provider 的 M3 modelId：
shouldApplyMinimaxM3ThinkingCompat的 provider 守卫缺少反向测试（如provider: "openai"+modelId: "MiniMax-M3"应返回false）。- 已是 adaptive 的 thinking：测试 idempotency，确保
{ type: "adaptive" }不会被重复改写。♻️ 建议补充的测试用例
it("does not patch non-anthropic-messages API payloads", () => { expect( captureThinkingPayload({ api: "openai", provider: "minimax", id: "MiniMax-M3", } as Parameters<StreamFn>[0]), ).toEqual({ thinking: { type: "enabled", budget_tokens: 1024 } }); }); it("rejects M3 modelId on non-minimax providers", () => { expect( shouldApplyMinimaxM3ThinkingCompat({ provider: "openai", modelId: "MiniMax-M3" }), ).toBe(false); }); it("leaves already-adaptive thinking unchanged", () => { let captured: unknown; const underlying: StreamFn = (payloadModel, _context, options) => { const payload = { thinking: { type: "adaptive" } }; options?.onPayload?.(payload, payloadModel); captured = payload; return {} as ReturnType<StreamFn>; }; void createMinimaxM3ThinkingWrapper(underlying)( { api: "anthropic-messages", provider: "minimax", id: "MiniMax-M3" } as Parameters<StreamFn>[0], { messages: [] } as Parameters<StreamFn>[1], {} as Parameters<StreamFn>[2], ); expect(captured).toEqual({ thinking: { type: "adaptive" } }); });🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@src/agents/pi-embedded-runner/minimax-stream-wrappers.test.ts` around lines 25 - 60, Extend the minimax stream wrapper tests with boundary coverage: verify captureThinkingPayload leaves a MiniMax-M3 payload unchanged for non-anthropic-messages APIs, verify shouldApplyMinimaxM3ThinkingCompat returns false for MiniMax-M3 on non-minimax providers, and verify createMinimaxM3ThinkingWrapper preserves an already-adaptive thinking payload without rewriting it.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@extensions/minimax/model-definitions.ts`:
- Line 43: 更新 buildMinimaxModelDefinition 中 input 的回退逻辑：未传入 params.input
时，优先使用对应 catalog 项的 input，只有 catalog?.input 也不存在时才回退为 ["text"]；保留显式 params.input
的现有行为。
---
Nitpick comments:
In `@src/agents/pi-embedded-runner/extra-params.ts`:
- Around line 371-378: 在应用 createMinimaxM3ThinkingWrapper 的条件分支中添加与其他条件包装器一致的
log.debug 调用，记录 provider、modelId 及 M3 thinking 兼容包装已应用；保持现有
shouldApplyMinimaxM3ThinkingCompat 判断和包装逻辑不变。
In `@src/agents/pi-embedded-runner/minimax-stream-wrappers.test.ts`:
- Around line 25-60: Extend the minimax stream wrapper tests with boundary
coverage: verify captureThinkingPayload leaves a MiniMax-M3 payload unchanged
for non-anthropic-messages APIs, verify shouldApplyMinimaxM3ThinkingCompat
returns false for MiniMax-M3 on non-minimax providers, and verify
createMinimaxM3ThinkingWrapper preserves an already-adaptive thinking payload
without rewriting it.
In `@src/agents/pi-embedded-runner/minimax-stream-wrappers.ts`:
- Around line 29-39: 在 streamWithPayloadPatch 的 payloadObj 处理回调中，为将 thinking 的
enabled 形式归一化为 adaptive 的赋值逻辑添加一行简短注释，明确说明该兼容性改写会丢弃 budget_tokens
等额外字段；不要修改现有转换行为。
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: d40cdd2d-0082-4336-b392-1b0da346f079
📒 Files selected for processing (14)
docs/providers/minimax.mdextensions/minimax/index.tsextensions/minimax/model-definitions.test.tsextensions/minimax/model-definitions.tsextensions/minimax/onboard.test.tsextensions/minimax/onboard.tsextensions/minimax/openclaw.plugin.jsonextensions/minimax/provider-catalog.tsextensions/minimax/provider-models.tssrc/agents/models-config.providers.minimax.test.tssrc/agents/pi-embedded-runner/extra-params.tssrc/agents/pi-embedded-runner/minimax-stream-wrappers.test.tssrc/agents/pi-embedded-runner/minimax-stream-wrappers.tssrc/plugins/bundled-plugin-metadata.generated.ts
| reasoning: params.reasoning ?? catalog?.reasoning ?? false, | ||
| input: ["text"], | ||
| cost: params.cost, | ||
| input: params.input ? [...params.input] : ["text"], |
There was a problem hiding this comment.
🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win
input 回退逻辑与 name/reasoning 不一致。
buildMinimaxModelDefinition 中，name 和 reasoning 在未传入时会回退到 catalog 值，但 input 直接默认为 ["text"]，未检查 catalog?.input。若直接调用此函数并传入 catalog 中已有的模型 ID（如 "MiniMax-M3"）但未传 input，将得到 ["text"] 而非 ["text", "image"]。
🔧 建议修复：使 `input` 回退到 catalog
- input: params.input ? [...params.input] : ["text"],
+ input: [...(params.input ?? catalog?.input ?? ["text"])],📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| input: params.input ? [...params.input] : ["text"], | |
| input: [...(params.input ?? catalog?.input ?? ["text"])], |
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@extensions/minimax/model-definitions.ts` at line 43, 更新
buildMinimaxModelDefinition 中 input 的回退逻辑：未传入 params.input 时，优先使用对应 catalog 项的
input，只有 catalog?.input 也不存在时才回退为 ["text"]；保留显式 params.input 的现有行为。
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
1 participant
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.
Summary
Change Type (select all)
Scope (select all touched areas)
Linked Issue/PR
User-visible / Behavior Changes
MiniMax M3 becomes the recommended default for API-key and Token Plan onboarding. MiniMax M2.7 and MiniMax M2.7 Highspeed remain available. The provider guide now lists global and CN endpoints for both Anthropic-compatible and OpenAI-compatible clients.
Security Impact (required)
No)No)No)No)No)Yes, explain risk + mitigation: None.Repro + Verification
Environment
Steps
extensions/minimax/openclaw.plugin.json.Expected
adaptive, and generated metadata matches the canonical manifest.Actual
Evidence
Verification results:
extensions/minimax/model-definitions.test.tsandextensions/minimax/onboard.test.ts: 17 tests passed.src/agents/pi-embedded-runner/minimax-stream-wrappers.test.ts: 3 tests passed.Human Verification (required)
Review Conversations
Compatibility / Migration
Yes)No)No)Failure Recovery (if this breaks)
minimax/MiniMax-M2.7explicitly or revert the commit.Risks and Mitigations