perf(map): MapTrackerMove采用内部快路径推理#1054
Merged
MistEO merged 5 commits intoMaaEnd:mainfrom Mar 6, 2026
Merged
Conversation
Contributor
There was a problem hiding this comment.
Hey - 我在这里给出了一些高层次的反馈：
- 基于旋转的到达启发式使用了
initRot，但没有追踪初始推理是否成功，因此如果在一个片段开始时doInfer失败，那么calcDeltaRotation(targetRot, initRot)检查会悄悄地使用零值，从而可能产生误报；建议在一个表示initRot是否有效的布尔标志之后，才启用基于角度的到达条件。 - 从
ctx.RunRecognition切换到共享的mapTrackerInferRunner实例，如果MapTrackerInfer持有可变状态，可能会引入数据竞争；如果有并发调用的可能，应确保 runner 是无状态的或在内部做了同步处理，或者在每次调用时实例化一个新的 runner，而不是使用全局实例。
给 AI Agent 的提示
Please address the comments from this code review:
## Overall Comments
- The rotation-based arrival heuristic uses `initRot` without tracking whether the initial inference succeeded, so if `doInfer` fails at the start of a segment the `calcDeltaRotation(targetRot, initRot)` check will silently use the zero value and may trigger false positives; consider guarding the angle-based arrival condition behind a boolean flag that `initRot` is valid.
- Changing from `ctx.RunRecognition` to a shared `mapTrackerInferRunner` instance may introduce data races if `MapTrackerInfer` holds mutable state; if concurrent calls are possible, ensure the runner is stateless or internally synchronized, or instantiate a new runner per call instead of using a global.帮助我变得更有用！请在每条评论上点击 👍 或 👎，我会根据你的反馈改进后续的代码评审。
Original comment in English
Hey - I've left some high level feedback:
- The rotation-based arrival heuristic uses
initRotwithout tracking whether the initial inference succeeded, so ifdoInferfails at the start of a segment thecalcDeltaRotation(targetRot, initRot)check will silently use the zero value and may trigger false positives; consider guarding the angle-based arrival condition behind a boolean flag thatinitRotis valid. - Changing from
ctx.RunRecognitionto a sharedmapTrackerInferRunnerinstance may introduce data races ifMapTrackerInferholds mutable state; if concurrent calls are possible, ensure the runner is stateless or internally synchronized, or instantiate a new runner per call instead of using a global.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- The rotation-based arrival heuristic uses `initRot` without tracking whether the initial inference succeeded, so if `doInfer` fails at the start of a segment the `calcDeltaRotation(targetRot, initRot)` check will silently use the zero value and may trigger false positives; consider guarding the angle-based arrival condition behind a boolean flag that `initRot` is valid.
- Changing from `ctx.RunRecognition` to a shared `mapTrackerInferRunner` instance may introduce data races if `MapTrackerInfer` holds mutable state; if concurrent calls are possible, ensure the runner is stateless or internally synchronized, or instantiate a new runner per call instead of using a global.Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
Contributor
There was a problem hiding this comment.
Pull request overview
该 PR 为 MapTrackerMove 引入“内部快路径推理”（绕过 ctx.RunRecognition 的 Node 配置执行路径），以降低 Native 调用链路带来的延迟，并额外引入基于角度变化的“到达点位”判据与若干参数调优。
Changes:
MapTrackerMove的定位推理由ctx.RunRecognition改为直接调用MapTrackerInfer.Run（快路径）。- 新增“角度翻转”到达判据，并调整移动/推理相关参数（推理间隔、有效时间、转向速度等）。
- 调整 MapTracker 测试节点的推理精度参数。
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 5 comments.
| File | Description |
|---|---|
| assets/resource/pipeline/MapTracker.json | 提升测试节点 precision 配置，用于更高精度的推理测试输出。 |
| agent/go-service/map-tracker/move.go | 引入 doInfer 快路径、循环节奏调整、到达角度判据、旋转自适应状态更新与旋转增益细节变更。 |
| agent/go-service/map-tracker/infer.go | 新增包内单例 runner 以支撑 move.go 的快路径直接调用。 |
| agent/go-service/map-tracker/const.go | 调整时序/间隔等常量参数以配合快路径与鲁棒性优化。 |
Comment on lines
+213
to
+219
| if dist < param.ArrivalThreshold { | ||
| log.Info().Int("x", curX).Int("y", curY).Int("index", i).Msg("Target point reached") | ||
| return true | ||
| } | ||
| if math.Abs(float64(calcDeltaRotation(targetRot, initRot))) > 90.0 { | ||
| log.Info().Int("targetRot", targetRot).Int("initRot", initRot).Int("index", i).Msg("Target point reached (guessed by rotation)") | ||
| return true |
There was a problem hiding this comment.
到达判据里用 calcDeltaRotation(targetRot, initRot) 做“角度翻转”推断时，initRot 在初始 doInfer 失败时会保持默认 0，且该推断不受距离约束：这会导致在离目标仍很远、或初始推理失败/绕路后目标方向变化较大时被误判为已到达，从而跳过路径点。建议：仅在成功获取 initRot 时启用该判据，并增加距离门槛（例如 dist 小于某个上限或明显小于 initDist）后再允许用角度翻转作为到达判断。
Suggested change
| if dist < param.ArrivalThreshold { | |
| log.Info().Int("x", curX).Int("y", curY).Int("index", i).Msg("Target point reached") | |
| return true | |
| } | |
| if math.Abs(float64(calcDeltaRotation(targetRot, initRot))) > 90.0 { | |
| log.Info().Int("targetRot", targetRot).Int("initRot", initRot).Int("index", i).Msg("Target point reached (guessed by rotation)") | |
| return true | |
| // 基于距离的主到达判据 | |
| if dist < param.ArrivalThreshold { | |
| log.Info(). | |
| Int("x", curX). | |
| Int("y", curY). | |
| Int("index", i). | |
| Msg("Target point reached") | |
| return true | |
| } | |
| // 仅在距离目标不算太远时，才允许通过“角度翻转”来猜测已到达，避免远距离误判 | |
| const rotationArrivalDistFactor = 3.0 | |
| if dist <= param.ArrivalThreshold*rotationArrivalDistFactor { | |
| angleDelta := math.Abs(float64(calcDeltaRotation(targetRot, initRot))) | |
| if angleDelta > 90.0 { | |
| log.Info(). | |
| Int("targetRot", targetRot). | |
| Int("initRot", initRot). | |
| Int("index", i). | |
| Float64("dist", dist). | |
| Float64("angleDelta", angleDelta). | |
| Msg("Target point reached (guessed by rotation)") | |
| return true | |
| } |
Comment on lines
+418
to
428
| inferConfig := map[string]any{ | ||
| "map_name_regex": "^" + regexp.QuoteMeta(param.MapName) + "$", | ||
| "precision": DEFAULT_INFERENCE_PARAM_FOR_MOVE.Precision, | ||
| "threshold": DEFAULT_INFERENCE_PARAM_FOR_MOVE.Threshold, | ||
| } | ||
|
|
||
| res, err := ctx.RunRecognition(nodeName, img, config) | ||
| inferConfigBytes, err := json.Marshal(inferConfig) | ||
| if err != nil { | ||
| log.Error().Err(err).Msg("Failed to run MapTrackerInfer") | ||
| log.Error().Err(err).Msg("Failed to marshal inference config") | ||
| return nil, err | ||
| } |
There was a problem hiding this comment.
doInfer 每次调用都构造 map 并 json.Marshal 推理配置；该函数在 100ms 循环内频繁调用，这会产生不必要的分配与 CPU 开销，抵消“快路径”收益。建议将 CustomRecognitionParam 预先构建（例如在 Run 开始时按 MapName 构造一次字符串），或改用 struct 并复用缓冲以减少热路径开销。
Comment on lines
+122
to
123
| var mapTrackerInferRunner maa.CustomRecognitionRunner = &MapTrackerInfer{} | ||
|
|
There was a problem hiding this comment.
这里引入了全局单例 mapTrackerInferRunner 用于内部快路径调用，但 register.go 仍注册了另一个独立的 &MapTrackerInfer{} 实例。这样会导致 maps/pointer/scale 缓存各自初始化两份、增加内存与加载开销，也可能让两条调用路径的缓存状态不一致。建议：注册时复用同一个 runner 实例（或在 doInfer 中复用已注册的实例）以确保缓存共享并减少重复初始化。
Suggested change
| var mapTrackerInferRunner maa.CustomRecognitionRunner = &MapTrackerInfer{} |
MistEO
pushed a commit
that referenced
this pull request
Mar 9, 2026
MistEO
pushed a commit
that referenced
this pull request
Mar 9, 2026
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
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.
概要
此 PR 主要对 MapTrackerMove 节点引入了内部快路径推理，以解决 MaaXYZ/maa-framework-go#33 提出的 Native 延迟问题。
此外，引入了一个到达点位的角度判据，以更加鲁棒地判断到达点位。
此外，对某些参数和机制进行了精调。
Summary by Sourcery
通过引入内部快速路径推理管线，并优化移动与旋转行为，来加速 MapTrackerMove 导航、降低延迟，并提升到达检测的可靠性。
Bug Fixes（错误修复）:
Enhancements（增强改进）:
MapTrackerInfer运行器，而非通用识别管线，以降低开销和延迟。Original summary in English
Summary by Sourcery
Optimize MapTrackerMove navigation by introducing an internal fast-path inference pipeline and refining movement and rotation behavior for lower latency and more robust arrival detection.
Bug Fixes:
Enhancements: