Merge:'feat/Agent-RedDot'| 专用红点识别器替代模板#294
Merged
Merged
Conversation
(cherry picked from commit df96a84)
初步测试较为优秀。 HIGHLIGHT:红点特征识别，稳健性极高，且高度复用。
1280/1281:隐式Rec=SubName已被废止,声明子识别会被忽略。
Contributor
📝 WalkthroughWalkthrough新增模块级 HSV 工具与连通域标注；在 agent/recognition/binarymatch.py 中实现 RedDotDetector 自定义识别器；在多个 pipeline JSON 中将模板/颜色识别替换为 RedDotDetector 或复合识别，并新增预设与全局识别节点。 走查PR 实现了基于 HSV 的红点检测自定义识别器 RedDotDetector，通过模块级工具函数替换原有的 HSVShapeMatching 类方法，并将游戏全流程的红点识别方式从模板匹配统一迁移到该新识别器，同时大幅重构流程节点的识别条件与控制逻辑，采用复合识别（And/Or）和节点引用模式以降低配置重复性。 变更红点检测器核心实现
预设配置与全局识别节点
游戏流水线识别迁移（示例）
评估🎯 4 (Complex) | ⏱️ ~60 分钟 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 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 |
Reviewer's Guide基于 HSV 颜色 + 拓扑分析，引入独立的 RedDotDetector 自定义识别器，将通用 HSV 工具函数从 HSVShapeMatching 中抽取为模块级辅助方法，并更新 pipeline JSON 定义以支持新的 OCR/场景识别以及红色检测。 RedDotDetector analyze 预设模式 vs 独立流水线的流程图flowchart TD
A[analyze] --> B{params contains preset}
B -- yes --> C[_run_preset]
C --> D[context.run_recognition preset_node]
D --> E{detail and detail.hit}
E -- yes --> F[adjust box with roi offset]
F --> G[return AnalyzeResult]
E -- no --> H[return None]
B -- no --> I[_run_standalone]
I --> J[read hsv_ranges, red_area, inner_v_min, inner_s_max, gap_ratio]
J --> K[apply roi to argv.image]
K --> L[convert work_bgr to HSV]
L --> M[_compute_hsv_mask]
M --> N[_label_blobs on red_mask]
N --> O{for each blob in area range}
O --> P[compute blob bbox]
P --> Q[extract box_red, box_hsv]
Q --> R[_label_blobs on non_red_crop]
R --> S[compute enclosed & inner_bright]
S --> T{any inner_bright pixels}
T -- no --> O
T -- yes --> U[_has_exclamation]
U --> V{has exclamation}
V -- no --> O
V -- yes --> W[compose result_box with roi offset]
W --> X[return AnalyzeResult]
O -->|no more blobs| Y[return None]
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your Experience访问你的 dashboard 以：
Getting HelpOriginal review guide in EnglishReviewer's GuideIntroduces a dedicated RedDotDetector custom recognizer based on HSV color + topology analysis, factors common HSV utilities out of HSVShapeMatching into module‑level helpers, and updates pipeline JSON definitions to support new OCR/scene recognition and red-color checks. Flow diagram for RedDotDetector analyze preset vs standalone pipelineflowchart TD
A[analyze] --> B{params contains preset}
B -- yes --> C[_run_preset]
C --> D[context.run_recognition preset_node]
D --> E{detail and detail.hit}
E -- yes --> F[adjust box with roi offset]
F --> G[return AnalyzeResult]
E -- no --> H[return None]
B -- no --> I[_run_standalone]
I --> J[read hsv_ranges, red_area, inner_v_min, inner_s_max, gap_ratio]
J --> K[apply roi to argv.image]
K --> L[convert work_bgr to HSV]
L --> M[_compute_hsv_mask]
M --> N[_label_blobs on red_mask]
N --> O{for each blob in area range}
O --> P[compute blob bbox]
P --> Q[extract box_red, box_hsv]
Q --> R[_label_blobs on non_red_crop]
R --> S[compute enclosed & inner_bright]
S --> T{any inner_bright pixels}
T -- no --> O
T -- yes --> U[_has_exclamation]
U --> V{has exclamation}
V -- no --> O
V -- yes --> W[compose result_box with roi offset]
W --> X[return AnalyzeResult]
O -->|no more blobs| Y[return None]
File-Level Changes
Tips and commandsInteracting with Sourcery
Customizing Your ExperienceAccess your dashboard to:
Getting Help
|
![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey - 我发现了两个问题，并给出了一些整体反馈：
- 在
_label_blobs中使用了deque，但在当前 diff 中没有看到相应的导入；请确认在模块级别添加from collections import deque（或等效导入），以避免运行时出现 NameError。 - 在
RedDotDetector._run_preset和_run_standalone中，你使用了裸print来输出命中日志；为了与模块其他部分保持一致并更好地控制日志输出，建议改用mfaalog.debug/info等日志函数。
给 AI Agent 的提示
Please address the comments from this code review:
## Overall Comments
- In `_label_blobs` you use `deque` but there is no corresponding import in the shown diff; ensure `from collections import deque` (or equivalent) is present at module level to avoid a runtime NameError.
- In `RedDotDetector._run_preset` and `_run_standalone` you use bare `print` for hit logs; for consistency with the rest of the module and better log control, consider replacing these with `mfaalog.debug/info` calls.
## Individual Comments
### Comment 1
<location path="agent/recognition/binarymatch.py" line_range="620-622" />
<code_context>
+ print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")
</code_context>
<issue_to_address>
**suggestion:** 避免使用裸 `print` 进行日志输出，请改用已有的日志设施。
使用 `print` 会创建一个未配置的日志通道，绕过我们的日志配置，并可能在生产环境中过度输出到 stdout。请通过 `mfaalog`（例如 `mfaalog.info`/`debug`）来记录日志，以保持一致性并实现集中控制。
```suggestion
adjusted = (bx + rx, by + ry, bw, bh)
mfaalog.debug(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")
return CustomRecognition.AnalyzeResult(box=adjusted, detail={"preset": preset_node})
```
</issue_to_address>
### Comment 2
<location path="agent/recognition/binarymatch.py" line_range="606-611" />
<code_context>
+ """
+ preset_node = params["preset"]
+ roi = argv.roi
+ if roi is not None:
+ rx, ry, rw, rh = roi.x, roi.y, roi.w, roi.h
+ else:
+ rx = ry = rw = rh = 0
+
+ if rw > 0 and rh > 0:
+ cropped = argv.image[ry:ry + rh, rx:rx + rw]
+ else:
</code_context>
<issue_to_address>
**issue:** 建议将 ROI 坐标限制在图像边界内，以避免出现负数或越界索引。
由于 `_run_preset` 和 `_run_standalone` 使用 `ry:ry+rh` / `rx:rx+rw` 对 `argv.image` 进行切片，如果从 JSON 中加载的 ROI 含有负值或超出尺寸的数值，将会依赖 NumPy 的索引环绕/截断行为，这可能导致隐蔽的误检测。将 `rx, ry, rw, rh` 约束在 `[0, w/h]` 范围后再进行切片（并在裁剪结果为空时及早返回）会让这种行为更加明确且更安全。
</issue_to_address>帮我变得更有用！请在每条评论上点 👍 或 👎，我会根据这些反馈改进之后的 Review。
Original comment in English
Hey - I've found 2 issues, and left some high level feedback:
- In
_label_blobsyou usedequebut there is no corresponding import in the shown diff; ensurefrom collections import deque(or equivalent) is present at module level to avoid a runtime NameError. - In
RedDotDetector._run_presetand_run_standaloneyou use bareprintfor hit logs; for consistency with the rest of the module and better log control, consider replacing these withmfaalog.debug/infocalls.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- In `_label_blobs` you use `deque` but there is no corresponding import in the shown diff; ensure `from collections import deque` (or equivalent) is present at module level to avoid a runtime NameError.
- In `RedDotDetector._run_preset` and `_run_standalone` you use bare `print` for hit logs; for consistency with the rest of the module and better log control, consider replacing these with `mfaalog.debug/info` calls.
## Individual Comments
### Comment 1
<location path="agent/recognition/binarymatch.py" line_range="620-622" />
<code_context>
+ print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")
</code_context>
<issue_to_address>
**suggestion:** Avoid using bare `print` for logging and use the existing logging facility instead.
Using `print` creates an unconfigured logging path that bypasses our logging setup and can spam stdout in production. Please route this through `mfaalog` (e.g. `mfaalog.info`/`debug`) for consistency and centralized control.
```suggestion
adjusted = (bx + rx, by + ry, bw, bh)
mfaalog.debug(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")
return CustomRecognition.AnalyzeResult(box=adjusted, detail={"preset": preset_node})
```
</issue_to_address>
### Comment 2
<location path="agent/recognition/binarymatch.py" line_range="606-611" />
<code_context>
+ """
+ preset_node = params["preset"]
+ roi = argv.roi
+ if roi is not None:
+ rx, ry, rw, rh = roi.x, roi.y, roi.w, roi.h
+ else:
+ rx = ry = rw = rh = 0
+
+ if rw > 0 and rh > 0:
+ cropped = argv.image[ry:ry + rh, rx:rx + rw]
+ else:
</code_context>
<issue_to_address>
**issue:** Consider clamping ROI coordinates to image bounds to avoid negative or out-of-range indexing.
Since `_run_preset` and `_run_standalone` slice `argv.image` with `ry:ry+rh` / `rx:rx+rw`, an ROI loaded from JSON with negative or oversized values will rely on NumPy’s index wrapping/truncation, which can cause subtle mis-detections. Clamping `rx, ry, rw, rh` to `[0, w/h]` before slicing (and early-returning if the clamped crop is empty) would make this behavior explicit and safer.
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Actionable comments posted: 5
🧹 Nitpick comments (4)
assets/resource/base/pipeline/Global.json (1)
727-744: 💤 Low value红色色相仅覆盖单段 0–10，可能漏检环绕段红色。
HSV 色相空间中红色横跨 0 与 180 两端。本节点
lower:[0,150,230]/upper:[10,255,255]只覆盖低色相段，而同 PR 的RedDot_Preset红色阈值是 0–12 与 165–180 双段。若目标红色文字/背景偏向高色相段，会被漏检。如确认目标红仅落在 0–10，可忽略；否则建议补一段高色相判定（ColorMatch 单节点无法同时覆盖两段，可拆为Or复合两段色核）。🤖 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 `@assets/resource/base/pipeline/Global.json` around lines 727 - 744, The Sub_Ocr_Red_Clr ColorMatch node only covers red hue 0–10 and may miss red hues near 165–180; update the pipeline so red is matched in both hue ranges by either splitting Sub_Ocr_Red_Clr into two ColorMatch nodes (one for 0–10 and one for 165–180) and combine them with an Or node, or mirror the approach used by RedDot_Preset (dual-range thresholds); ensure the new nodes keep the same roi "OCR", method, connected/count settings and that the combined result replaces the original Sub_Ocr_Red_Clr output.assets/resource/base/pipeline/Daily.json (1)
113-141: ⚡ Quick win建议澄清/清理
Daily_UnionBack的target（不要直接补Click）
Daily_UnionBack未配置action，默认按 DoNothing 执行，因此target: [120, 35]不会触发点击；其返回主页实际依赖next跳转到Global_ToHomePage_Enter/Global_ToHomePage，而Global.json中Global_ToHomePage_HomeButton、Global_ToHomePage_Reset才包含action: "Click"来完成返回。建议删除该无用target（或补注释说明由 Global 承担），避免误加action: "Click"导致重复/误触。🤖 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 `@assets/resource/base/pipeline/Daily.json` around lines 113 - 141, The Daily_UnionBack node declares a target ([120,35]) but has no action so it defaults to DoNothing and that target is misleading; either remove the target entry or add a comment clarifying that navigation is handled by next -> Global_ToHomePage_Enter/Global_ToHomePage and that the actual Click actions live in Global_ToHomePage_HomeButton and Global_ToHomePage_Reset; do not add action:"Click" here (would duplicate/trigger unintended clicks).agent/recognition/binarymatch.py (2)
621-621: ⚡ Quick win命中日志建议统一走
mfaalog，而非本文件其余位置（含异常、调试图保存）均使用
mfaalog，而预设模式（Line 621）与独立模式（Line 716）的“命中”成功日志却用了♻️ 统一日志
- print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}") + mfaalog.info(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")- print(f"[RedDotDetector] 命中: box={result_box}, red_area={area}") + mfaalog.info(f"[RedDotDetector] 命中: box={result_box}, red_area={area}")Also applies to: 716-716
🤖 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 `@agent/recognition/binarymatch.py` at line 621, Replace the direct print calls that log hits with the project logger: change the print in the RedDotDetector hit path (the statement printing f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}") and the similar print in the independent mode branch to use mfaalog (e.g. mfaalog.info) with the same message/variables so the hit messages go through the standard logging pipeline and retain consistent formatting and levels; keep the message text unchanged and reuse existing mfaalog import/usages in this module for consistency.
617-622: 调整：detail.box的解包大概率是可行的，不应按“重大静默失效”处理公开的 MaaFramework Python 绑定里
Rect表示[x, y, w, h]，并且仓库内也已有对box[0]、box[2]这类下标访问的实际用法（如agent/fishing_agent.py），因此把bx, by, bw, bh = detail.box直接认定为Rect不可解包从而导致预设模式“恒返回 None”的判断依据不足。可保留当前写法；仅建议为一致性把红点检测处的 box 访问风格与roi.x/roi.y/roi.w/roi.h（或其它 box 用法）统一。🤖 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 `@agent/recognition/binarymatch.py` around lines 617 - 622, The current change treats unpacking detail.box as unsafe and causes presets to always return None; instead keep treating detail.box as a Rect and make the access style consistent with the rest of the code: in the RedDotDetector matching block (where detail = context.run_recognition(preset_node, cropped) and you currently use bx, by, bw, bh = detail.box and return CustomRecognition.AnalyzeResult), either retain the tuple unpacking or switch to explicit index access (detail.box[0], detail.box[1], detail.box[2], detail.box[3]) so it matches other usages like roi.x/roi.y/roi.w/roi.h and agent/fishing_agent.py; ensure the adjusted box is constructed the same way (e.g., (bx + rx, by + ry, bw, bh)) and return the AnalyzeResult as before.
🤖 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 `@agent/recognition/binarymatch.py`:
- Around line 339-343: The comment's documented default for red_area is
inconsistent with the actual default used in _run_standalone; update the comment
to match the code by changing the red_area default from [50, 1200] to [30,
1200], and optionally mention that _run_standalone reads params.get("red_area",
[30, 1200]) so callers not passing red_area will get 30 as the min; keep
references to hsv_ranges and red_area in the same comment block so readers see
both parameter purposes and the corrected default.
In `@assets/interface.json`:
- Around line 411-436: The "风" and "光" entries have an incorrect nested
pipeline_override: they wrap QuickHunt_StoneSelect_* entries inside
QuickHunt_FastBattleTouch_StoneSelect, which treats them as fields rather than
node overrides; remove the QuickHunt_FastBattleTouch_StoneSelect wrapper so
pipeline_override is a flat mapping keyed directly by node names
(QuickHunt_StoneSelect_Fire, QuickHunt_StoneSelect_Wind for 风 and
QuickHunt_StoneSelect_Fire, QuickHunt_StoneSelect_Light for 光) and set the
correct enabled values (Fire: false, Wind/Light: true) to match the working
pattern used by the "水" and "暗" entries.
In `@assets/resource/base/pipeline/Battle.json`:
- Around line 792-808: The QuickHunt_StoneSelect_Wind node is missing the
"action":"Click" field so it never triggers a click like its siblings
QuickHunt_StoneSelect_Fire/Water/Light/Duck; update the
QuickHunt_StoneSelect_Wind JSON object to include "action": "Click" (placed
after "expected") so that when "风之洞穴" is recognized the automation performs the
click and proceeds, preserving the same behavior described by the node's desc
("圣石选属性.Loc分离").
- Around line 1069-1078: The replace mapping in
QuickHunt_CollectMap_Initialize_Out_Ty2 currently contains a no-op entry
["跡","跡"] which prevents OCR normalization to simplified Chinese; update the
second tuple in the "replace" array for the handler
QuickHunt_CollectMap_Initialize_Out_Ty2 to map "跡" to "迹" (i.e., ["跡","迹"]) so
OCR outputs like "哥布林遺跡" are normalized to "哥布林遗迹" and downstream
match/next/all_of checks (e.g., QuickHunt_CollectMap_Initialize_Out_Ty1,
QuickHunt_CollectMapAdventureRouteDoubleCk2No_ToCentral_CK,
QuickHunt_CrystalCave_End, QuickHunt_CollectMap_SwipLeft,
QuickHunt_CollectMap_SwipNoth) will work correctly.
In `@assets/resource/base/pipeline/Global.json`:
- Around line 877-889: The regex in the Rec_Reward_Ocr entry uses character
classes like [点击画面] which match individual characters rather than the full
phrases; replace those with phrase lookaheads so OCR must contain the full
substrings—for example change the "expected" value to use
(?=.*点击画面)(?=.*即可)(?=.*返回).{4,12}$ (ensure any JSON escaping needed for
backslashes is applied in the file) so the key "Rec_Reward_Ocr" and its
"expected" field require the whole phrases instead of single characters.
---
Nitpick comments:
In `@agent/recognition/binarymatch.py`:
- Line 621: Replace the direct print calls that log hits with the project
logger: change the print in the RedDotDetector hit path (the statement printing
f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}") and the similar
print in the independent mode branch to use mfaalog (e.g. mfaalog.info) with the
same message/variables so the hit messages go through the standard logging
pipeline and retain consistent formatting and levels; keep the message text
unchanged and reuse existing mfaalog import/usages in this module for
consistency.
- Around line 617-622: The current change treats unpacking detail.box as unsafe
and causes presets to always return None; instead keep treating detail.box as a
Rect and make the access style consistent with the rest of the code: in the
RedDotDetector matching block (where detail =
context.run_recognition(preset_node, cropped) and you currently use bx, by, bw,
bh = detail.box and return CustomRecognition.AnalyzeResult), either retain the
tuple unpacking or switch to explicit index access (detail.box[0],
detail.box[1], detail.box[2], detail.box[3]) so it matches other usages like
roi.x/roi.y/roi.w/roi.h and agent/fishing_agent.py; ensure the adjusted box is
constructed the same way (e.g., (bx + rx, by + ry, bw, bh)) and return the
AnalyzeResult as before.
In `@assets/resource/base/pipeline/Daily.json`:
- Around line 113-141: The Daily_UnionBack node declares a target ([120,35]) but
has no action so it defaults to DoNothing and that target is misleading; either
remove the target entry or add a comment clarifying that navigation is handled
by next -> Global_ToHomePage_Enter/Global_ToHomePage and that the actual Click
actions live in Global_ToHomePage_HomeButton and Global_ToHomePage_Reset; do not
add action:"Click" here (would duplicate/trigger unintended clicks).
In `@assets/resource/base/pipeline/Global.json`:
- Around line 727-744: The Sub_Ocr_Red_Clr ColorMatch node only covers red hue
0–10 and may miss red hues near 165–180; update the pipeline so red is matched
in both hue ranges by either splitting Sub_Ocr_Red_Clr into two ColorMatch nodes
(one for 0–10 and one for 165–180) and combine them with an Or node, or mirror
the approach used by RedDot_Preset (dual-range thresholds); ensure the new nodes
keep the same roi "OCR", method, connected/count settings and that the combined
result replaces the original Sub_Ocr_Red_Clr output.
🪄 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: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: 8eb0b41e-3329-4bb3-8a5f-d384fafbd0bb
⛔ Files ignored due to path filters (18)
.gitignoreis excluded by none and included by noneassets/resource/PC/model/.gitignoreis excluded by none and included by noneassets/resource/base/image/ActiveGe.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/ActivePeerGe.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/FreeGacha.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_Free.ReservedRights_Red_GE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_ListFull2_GE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_ListFull3.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_ListFull_GE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_PoolSort.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GACHA_Special.ReservedRights_Red_GE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/GaChaEquRDGE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/Gacha-homepage.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/MailAtMainRedGe.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/MailRedPr.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/MissionsGE.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/Pass.pngis excluded by!**/*.png,!**/*.pngand included by noneassets/resource/base/image/pass1.pngis excluded by!**/*.png,!**/*.pngand included by none
📒 Files selected for processing (14)
agent/recognition/binarymatch.pyassets/interface.jsonassets/resource/base/pipeline/Activities.jsonassets/resource/base/pipeline/Battle.jsonassets/resource/base/pipeline/Daily.jsonassets/resource/base/pipeline/Dummy.jsonassets/resource/base/pipeline/GACHA.jsonassets/resource/base/pipeline/GACHA_ADV.jsonassets/resource/base/pipeline/Global.jsonassets/resource/base/pipeline/Mail.jsonassets/resource/base/pipeline/Pass.jsonassets/resource/base/pipeline/QuestList.jsonassets/resource/base/pipeline/StartGame.jsonassets/resource/base/pipeline/Test.json
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (2)
agent/recognition/binarymatch.py (2)
566-623: 💤 Low value建议统一使用
mfaalog替代print()Line 621 使用
print()输出命中信息，与文件其他位置使用mfaalog.info()/mfaalog.debug()风格不一致。生产环境下print()输出可能不会被日志系统捕获。♻️ 建议修改
- print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}") + mfaalog.info(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")🤖 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 `@agent/recognition/binarymatch.py` around lines 566 - 623, The preset branch in RedDotDetector._run_preset currently logs hits via print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}") which is inconsistent with the module logging; replace that print call with an mfaalog logging call (e.g., mfaalog.info or mfaalog.debug) and include the same message and variables (preset_node and adjusted) so hits are captured by the logging system.
629-722: 💤 Low value同上，建议统一日志输出方式
Line 716 同样使用
print()而非mfaalog。♻️ 建议修改
- print(f"[RedDotDetector] 命中: box={result_box}, red_area={area}") + mfaalog.info(f"[RedDotDetector] 命中: box={result_box}, red_area={area}")🤖 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 `@agent/recognition/binarymatch.py` around lines 629 - 722, The final hit logging in _run_standalone uses print(...) instead of the project's logger; replace the print call inside _run_standalone that outputs "[RedDotDetector] 命中: box=..." with a mfaalog logging call (e.g., mfaalog.info or mfaalog.debug to match surrounding messages) and include the same context (result_box and red area) in the formatted log string so it follows the same logging conventions as the other messages in this function.
🤖 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.
Nitpick comments:
In `@agent/recognition/binarymatch.py`:
- Around line 566-623: The preset branch in RedDotDetector._run_preset currently
logs hits via print(f"[RedDotDetector] [preset:{preset_node}] 命中 → {adjusted}")
which is inconsistent with the module logging; replace that print call with an
mfaalog logging call (e.g., mfaalog.info or mfaalog.debug) and include the same
message and variables (preset_node and adjusted) so hits are captured by the
logging system.
- Around line 629-722: The final hit logging in _run_standalone uses print(...)
instead of the project's logger; replace the print call inside _run_standalone
that outputs "[RedDotDetector] 命中: box=..." with a mfaalog logging call (e.g.,
mfaalog.info or mfaalog.debug to match surrounding messages) and include the
same context (result_box and red area) in the formatted log string so it follows
the same logging conventions as the other messages in this function.
ℹ️ Review info
⚙️ Run configuration
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
Run ID: d7663e83-3263-4a4f-b22d-9c73e1b38b8a
📒 Files selected for processing (1)
agent/recognition/binarymatch.py
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 by Sourcery
引入专用的红点通知检测器并共享 HSV 工具，同时扩展用于新 UI 状态的 OCR/颜色匹配流水线节点。
新特性：
RedDotDetector自定义识别模块，使用 HSV 与形状/拓扑分析检测带有内部感叹号的红色通知小圆点。增强：
HSVShapeMatching中重构为模块级辅助函数，以便其他检测器复用。Original summary in English
Summary by Sourcery
Introduce a dedicated red-dot notification detector and share HSV utilities while expanding OCR/color-match pipeline nodes for new UI states.
New Features:
Enhancements:
Summary by CodeRabbit
发布说明
新功能
功能优化
功能调整