feat: 新增ocr參考庫 #221
Open
Abino01 wants to merge 8 commits into
Open
Conversation
Contributor
Reviewer's Guide引入一个集中式多语言 OCR 校验库和辅助函数，通过将任务名称映射到预期的本地化字符串，并对 OCR 输出执行精确/正则匹配和模糊匹配，从而将文本匹配规则与 JSON 流水线配置解耦。 使用 verify_single_task_ocr 的集中式多语言 OCR 校验流程图flowchart TD
A["Frontend OCR node<br/>(task_name, roi_text)"] --> B[Call verify_single_task_ocr]
B --> C{"task_name in<br/>MULTILANG_OCR_LIBRARY?"}
C -->|No| Z["Return False<br/>(skip action)"]
C -->|Yes| D["Load task_data<br/>from MULTILANG_OCR_LIBRARY"]
D --> E["cleaned_text =<br/>frontend_ocr_text.strip()"]
E --> F{"Exact/regex match<br/>re.search over expected"}
F -->|Matched| X["Return True<br/>(hit on exact/regex)"]
F -->|No match| G{"Fuzzy match<br/>SequenceMatcher<br/>similarity >= threshold"}
G -->|Matched| Y["Return True<br/>(hit on fuzzy)"]
G -->|No match| Z
File-Level Changes
提示与命令与 Sourcery 交互
自定义使用体验打开你的 dashboard：
获取帮助Original review guide in EnglishReviewer's GuideIntroduce a centralized multi-language OCR verification library and helper function that decouples text-matching rules from JSON pipeline configs by mapping task names to expected localized strings and performing exact/regex and fuzzy matching on OCR outputs. Flow diagram for centralized multi-language OCR verification using verify_single_task_ocrflowchart TD
A["Frontend OCR node<br/>(task_name, roi_text)"] --> B[Call verify_single_task_ocr]
B --> C{"task_name in<br/>MULTILANG_OCR_LIBRARY?"}
C -->|No| Z["Return False<br/>(skip action)"]
C -->|Yes| D["Load task_data<br/>from MULTILANG_OCR_LIBRARY"]
D --> E["cleaned_text =<br/>frontend_ocr_text.strip()"]
E --> F{"Exact/regex match<br/>re.search over expected"}
F -->|Matched| X["Return True<br/>(hit on exact/regex)"]
F -->|No match| G{"Fuzzy match<br/>SequenceMatcher<br/>similarity >= threshold"}
G -->|Matched| Y["Return True<br/>(hit on fuzzy)"]
G -->|No match| Z
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}
Contributor
There was a problem hiding this comment.
Hey - 我发现了两个问题，并给出了一些整体性的反馈：
- 在若干个
expected列表中，因为缺少逗号，你不小心把相邻的字符串字面量拼接在一起了（例如：InCharactersMenu中有r"(?i)Profile" r"(?i)INFO"，以及Ptory中"入る"后面紧跟着r"(?i)ENTER"），这会改变原本预期的匹配模式；请在这些字面量之间补上逗号。 - 建议在
verify_single_task_ocr中对frontend_ocr_text为非字符串或None的情况做防御性处理（例如：如果它为假值或不是str，就尽早返回False），以避免在调用.strip()或正则时出现意外异常。 - 由于
MULTILANG_OCR_LIBRARY体积较大且大量依赖正则，你可能会希望预先编译一次这些正则模式（例如将编译后的 pattern 存起来而不是原始字符串），从而避免在每次调用verify_single_task_ocr时重复解析，降低单次调用开销。
给 AI 代理的提示词
Please address the comments from this code review:
## Overall Comments
- In several `expected` lists you are accidentally concatenating adjacent string literals due to missing commas (e.g. `InCharactersMenu` has `r"(?i)Profile" r"(?i)INFO"`, and `Ptory` has `"入る"` followed directly by `r"(?i)ENTER"`), which will change the intended patterns; please add commas between these literals.
- Consider defensively handling non-string or `None` values for `frontend_ocr_text` in `verify_single_task_ocr` (e.g. early return `False` if it's falsy or not a `str`) to avoid unexpected exceptions from `.strip()` or the regex calls.
- Since `MULTILANG_OCR_LIBRARY` is large and heavily regex-based, you may want to precompile regex patterns once (e.g. store compiled patterns instead of raw strings) to avoid re-parsing them on every `verify_single_task_ocr` call and reduce per-call overhead.
## Individual Comments
### Comment 1
<location path="agent/utils/ocr_i18n_lib.py" line_range="334-335" />
<code_context>
+ "InCharactersMenu": {
+ "desc": "在角色菜单内",
+ "expected": [
+ "个人信息", "個人資訊", "信息", "資訊",
+ r"(?i)PROFILE", r"(?i)Profile" r"(?i)INFO", r"(?i)Info",
+ "メッセージ", "情報", "정보"
+ ]
</code_context>
<issue_to_address>
**issue (bug_risk):** Missing comma between the `Profile` and `INFO` patterns causes an unintended concatenated regex.
Python concatenates adjacent string literals, so `r"(?i)Profile" r"(?i)INFO"` becomes `r"(?i)Profile(?i)INFO"`, which won’t behave as two separate patterns. Add a comma after `r"(?i)Profile"` to make them distinct entries in the list.
</issue_to_address>
### Comment 2
<location path="agent/utils/ocr_i18n_lib.py" line_range="561-123" />
<code_context>
+ "Ptory": {
+ "desc": "在钮",
+ "expected": [
+ "进入", "進入", "入る"
+ r"(?i)ENTER", r"(?i)Enter",
+ "들"
+ ]
</code_context>
<issue_to_address>
**issue (bug_risk):** Missing comma after the last CJK string leads to invalid syntax or unintended concatenation.
Add a trailing comma after "入る" so this remains a separate list element and doesn’t get concatenated with the following `r"(?i)ENTER"` string literal.
</issue_to_address>帮我变得更有用！请在每条评论上点 👍 或 👎，我会根据你的反馈改进后续的代码审查。
Original comment in English
Hey - I've found 2 issues, and left some high level feedback:
- In several
expectedlists you are accidentally concatenating adjacent string literals due to missing commas (e.g.InCharactersMenuhasr"(?i)Profile" r"(?i)INFO", andPtoryhas"入る"followed directly byr"(?i)ENTER"), which will change the intended patterns; please add commas between these literals. - Consider defensively handling non-string or
Nonevalues forfrontend_ocr_textinverify_single_task_ocr(e.g. early returnFalseif it's falsy or not astr) to avoid unexpected exceptions from.strip()or the regex calls. - Since
MULTILANG_OCR_LIBRARYis large and heavily regex-based, you may want to precompile regex patterns once (e.g. store compiled patterns instead of raw strings) to avoid re-parsing them on everyverify_single_task_ocrcall and reduce per-call overhead.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- In several `expected` lists you are accidentally concatenating adjacent string literals due to missing commas (e.g. `InCharactersMenu` has `r"(?i)Profile" r"(?i)INFO"`, and `Ptory` has `"入る"` followed directly by `r"(?i)ENTER"`), which will change the intended patterns; please add commas between these literals.
- Consider defensively handling non-string or `None` values for `frontend_ocr_text` in `verify_single_task_ocr` (e.g. early return `False` if it's falsy or not a `str`) to avoid unexpected exceptions from `.strip()` or the regex calls.
- Since `MULTILANG_OCR_LIBRARY` is large and heavily regex-based, you may want to precompile regex patterns once (e.g. store compiled patterns instead of raw strings) to avoid re-parsing them on every `verify_single_task_ocr` call and reduce per-call overhead.
## Individual Comments
### Comment 1
<location path="agent/utils/ocr_i18n_lib.py" line_range="334-335" />
<code_context>
+ "InCharactersMenu": {
+ "desc": "在角色菜单内",
+ "expected": [
+ "个人信息", "個人資訊", "信息", "資訊",
+ r"(?i)PROFILE", r"(?i)Profile" r"(?i)INFO", r"(?i)Info",
+ "メッセージ", "情報", "정보"
+ ]
</code_context>
<issue_to_address>
**issue (bug_risk):** Missing comma between the `Profile` and `INFO` patterns causes an unintended concatenated regex.
Python concatenates adjacent string literals, so `r"(?i)Profile" r"(?i)INFO"` becomes `r"(?i)Profile(?i)INFO"`, which won’t behave as two separate patterns. Add a comma after `r"(?i)Profile"` to make them distinct entries in the list.
</issue_to_address>
### Comment 2
<location path="agent/utils/ocr_i18n_lib.py" line_range="561-123" />
<code_context>
+ "Ptory": {
+ "desc": "在钮",
+ "expected": [
+ "进入", "進入", "入る"
+ r"(?i)ENTER", r"(?i)Enter",
+ "들"
+ ]
</code_context>
<issue_to_address>
**issue (bug_risk):** Missing comma after the last CJK string leads to invalid syntax or unintended concatenation.
Add a trailing comma after "入る" so this remains a separate list element and doesn’t get concatenated with the following `r"(?i)ENTER"` string literal.
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
| "desc": "在详情页点进入按钮", | ||
| "expected": [ | ||
| "进入", "進入", "入る", | ||
| r"(?i)ENTER", r"(?i)Enter", |
Contributor
There was a problem hiding this comment.
issue (bug_risk): 最后一个 CJK 字符串后缺少逗号，会导致语法错误或意外的字符串拼接。
请在 "入る" 之后补一个逗号，这样它会保持为单独的列表元素，而不会与后面的 r"(?i)ENTER" 字符串字面量拼接在一起。
Original comment in English
issue (bug_risk): Missing comma after the last CJK string leads to invalid syntax or unintended concatenation.
Add a trailing comma after "入る" so this remains a separate list element and doesn’t get concatenated with the following r"(?i)ENTER" string literal.
Contributor
Author
|
@sourcery-ai guide |
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.
新增ocr_i18n_lib.py做為未來的ocr文字判定對應庫 不再透過assets/resource/base/pipeline/內的.json內容作判定 .json之後只作為roi區域讀取跟判定結果的操作 不參與文字檢驗跟判定
舉例來說
更動.json前
更動.json後
可以大幅降低未來維護的時間
路徑 agent/utils/ocr_i18n_lib.py
更新 move 0526*0951
現在在思考怎麼改 在agent/custom/action/裡的.py
Summary by Sourcery
引入一个集中管理的多语言 OCR 匹配库和验证辅助工具，将文本识别规则与 JSON 流水线配置解耦。
新功能：
MULTILANG_OCR_LIBRARY，用于在多种语言中将任务名称映射到本地化的 OCR 匹配模式。verify_single_task_ocr辅助函数，通过精确匹配和模糊匹配来验证指定任务的 OCR 输出。Original summary in English
Summary by Sourcery
Introduce a centralized multi-language OCR matching library and verification helper to decouple text recognition rules from JSON pipeline configurations.
New Features: