feat(customs): 新增在UI日志界面上显示OCR结果。

[gitlihang]

Summary by Sourcery

添加自定义动作和日志工具，用于在 UI 日志中展示 OCR 识别结果，并可根据这些结果选择性触发后续操作。

新功能：

• 引入 returnOCR 自定义动作，用于获取已配置识别任务的 OCR 结果，并将其记录到 UI。
• 支持基于 OCR 结果驱动的可选点击操作，可以点击在配置的坐标位置，或点击在识别文本区域的中心。
• 提供日志工具，为本模块中的自定义脚本统一规范控制台和文件的日志输出。

文档：

• 新增 README 文档，说明如何配置任务，将 OCR 结果输出到 UI 日志，以及如何配置是否启用后续动作。

Original summary in English

Summary by Sourcery

Add a custom action and logging utility to display OCR recognition results in the UI log and optionally trigger follow-up actions based on those results.

New Features:

• Introduce a returnOCR custom action that retrieves OCR results for a configured recognition task and logs them to the UI.
• Support optional click actions driven by OCR results, either at a configured coordinate or at the recognized text region center.
• Provide a logging utility that standardizes console and file logging output for custom scripts in this module.

Documentation:

• Add README documentation describing how to configure tasks to output OCR results to the UI log with or without follow-up actions.

[sourcery-ai]

你好，我发现了 3 个问题，并给出了一些整体性的反馈：

• returnOCR.py 使用了 from utils import logger，但之后通过 logger.info/debug/warning 调用它，好像 logger 是一个日志实例；由于 utils/__init__.py 是空的，这里的 logger 实际上会是 utils.logger 模块，所以你可能需要 from utils.logger import logger（或者在 utils/__init__.py 中重新导出它）。
• 使用 click_target 时假定它是一个 4 元素列表（[x, y, w, h]），但在索引之前没有对其长度或类型做校验，因此最好对输入做断言或防护，避免因为不符合预期的输入导致 IndexError 或类型问题。
• 当 action_key 既不是 "Click" 也不是空值时，代码目前会静默地继续执行；建议对未知的 action_key 值进行日志记录或处理，这样更容易排查配置错误。

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- `returnOCR.py` does `from utils import logger` but then calls `logger.info/debug/warning` as if `logger` were a logger instance; since `utils/__init__.py` is empty this `logger` will be the `utils.logger` module, so you likely want `from utils.logger import logger` (or to re-export it in `utils/__init__.py`).
- `click_target` is used assuming a 4-element list (`[x, y, w, h]`), but there's no validation of its length or type before indexing, so it would be safer to assert or guard against malformed input to avoid `IndexError`/type issues.
- When `action_key` is neither `"Click"` nor empty, the code currently just falls through silently; consider logging or handling unknown `action_key` values so configuration mistakes are easier to diagnose.

## Individual Comments

### Comment 1
<location path="Storage/customs/gitlihang/output-ocr-result-log/action/returnOCR.py" line_range="56-57" />
<code_context>
+                # 点击传入参数中的坐标位置
+                if click_target:
+                    box = click_target
+                    center_x = box[0] + box[2] // 2
+                    center_y = box[1] + box[3] // 2
+                    logger.debug(f"点击位置: ({center_x}, {center_y})")
+                    context.tasker.controller.post_click(center_x, center_y).wait()
</code_context>
<issue_to_address>
**issue (bug_risk):** Center point calculation for `click_target` is likely incorrect for `[x1, y1, x2, y2]` boxes.

With `[x1, y1, x2, y2]` boxes, `center_x = box[0] + box[2] // 2` computes `x1 + x2/2`, not the midpoint between `x1` and `x2`. If you want the rectangle center, it should be something like `(box[0] + box[2]) // 2` and `(box[1] + box[3]) // 2` (or float division for subpixel precision); otherwise the click will be biased toward the top-left.
</issue_to_address>

### Comment 2
<location path="Storage/customs/gitlihang/output-ocr-result-log/action/returnOCR.py" line_range="31" />
<code_context>
+    ) -> CustomAction.RunResult:
+        # logger.info("进入returnOCR")
+        # 解析自定义参数，并判断是否为空
+        argv_dict: dict = json.loads(argv.custom_action_param)
+        if not argv_dict:
+            return CustomAction.RunResult(success=True)
</code_context>
<issue_to_address>
**issue:** Lack of error handling for invalid `custom_action_param` JSON may cause unhandled exceptions.

If `argv.custom_action_param` is missing, empty, or contains invalid JSON, `json.loads` will raise and terminate the action. Consider catching `json.JSONDecodeError` and either defaulting to an empty dict or returning a failed `RunResult` with a clear log message, depending on the desired behavior for bad input.
</issue_to_address>

### Comment 3
<location path="Storage/customs/gitlihang/output-ocr-result-log/README.md" line_range="86" />
<code_context>
+
+## 注意事项
+
+ - 注意的值"custom_action_param"
+
+
</code_context>
<issue_to_address>
**issue (typo):** The phrase "注意的值" is a bit unclear; consider rephrasing for clarity.

`注意的值

```suggestion
- 请注意配置项 `custom_action_param` 的取值
```
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

Original comment in English

Hey - I've found 3 issues, and left some high level feedback:

• returnOCR.py does from utils import logger but then calls logger.info/debug/warning as if logger were a logger instance; since utils/__init__.py is empty this logger will be the utils.logger module, so you likely want from utils.logger import logger (or to re-export it in utils/__init__.py).
• click_target is used assuming a 4-element list ([x, y, w, h]), but there's no validation of its length or type before indexing, so it would be safer to assert or guard against malformed input to avoid IndexError/type issues.
• When action_key is neither "Click" nor empty, the code currently just falls through silently; consider logging or handling unknown action_key values so configuration mistakes are easier to diagnose.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- `returnOCR.py` does `from utils import logger` but then calls `logger.info/debug/warning` as if `logger` were a logger instance; since `utils/__init__.py` is empty this `logger` will be the `utils.logger` module, so you likely want `from utils.logger import logger` (or to re-export it in `utils/__init__.py`).
- `click_target` is used assuming a 4-element list (`[x, y, w, h]`), but there's no validation of its length or type before indexing, so it would be safer to assert or guard against malformed input to avoid `IndexError`/type issues.
- When `action_key` is neither `"Click"` nor empty, the code currently just falls through silently; consider logging or handling unknown `action_key` values so configuration mistakes are easier to diagnose.

## Individual Comments

### Comment 1
<location path="Storage/customs/gitlihang/output-ocr-result-log/action/returnOCR.py" line_range="56-57" />
<code_context>
+                # 点击传入参数中的坐标位置
+                if click_target:
+                    box = click_target
+                    center_x = box[0] + box[2] // 2
+                    center_y = box[1] + box[3] // 2
+                    logger.debug(f"点击位置: ({center_x}, {center_y})")
+                    context.tasker.controller.post_click(center_x, center_y).wait()
</code_context>
<issue_to_address>
**issue (bug_risk):** Center point calculation for `click_target` is likely incorrect for `[x1, y1, x2, y2]` boxes.

With `[x1, y1, x2, y2]` boxes, `center_x = box[0] + box[2] // 2` computes `x1 + x2/2`, not the midpoint between `x1` and `x2`. If you want the rectangle center, it should be something like `(box[0] + box[2]) // 2` and `(box[1] + box[3]) // 2` (or float division for subpixel precision); otherwise the click will be biased toward the top-left.
</issue_to_address>

### Comment 2
<location path="Storage/customs/gitlihang/output-ocr-result-log/action/returnOCR.py" line_range="31" />
<code_context>
+    ) -> CustomAction.RunResult:
+        # logger.info("进入returnOCR")
+        # 解析自定义参数，并判断是否为空
+        argv_dict: dict = json.loads(argv.custom_action_param)
+        if not argv_dict:
+            return CustomAction.RunResult(success=True)
</code_context>
<issue_to_address>
**issue:** Lack of error handling for invalid `custom_action_param` JSON may cause unhandled exceptions.

If `argv.custom_action_param` is missing, empty, or contains invalid JSON, `json.loads` will raise and terminate the action. Consider catching `json.JSONDecodeError` and either defaulting to an empty dict or returning a failed `RunResult` with a clear log message, depending on the desired behavior for bad input.
</issue_to_address>

### Comment 3
<location path="Storage/customs/gitlihang/output-ocr-result-log/README.md" line_range="86" />
<code_context>
+
+## 注意事项
+
+ - 注意的值"custom_action_param"
+
+
</code_context>
<issue_to_address>
**issue (typo):** The phrase "注意的值" is a bit unclear; consider rephrasing for clarity.

`注意的值

```suggestion
- 请注意配置项 `custom_action_param` 的取值
```
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue (bug_risk): click_target 的中心点计算对于 [x1, y1, x2, y2] 格式的框来说很可能不正确。

在使用 [x1, y1, x2, y2] 的情况下，center_x = box[0] + box[2] // 2 实际计算的是 x1 + x2/2，而不是 x1 与 x2 的中点。如果你想要的是矩形中心，更合适的写法是 (box[0] + box[2]) // 2 和 (box[1] + box[3]) // 2（或者使用浮点除法以获得子像素精度）；否则点击位置会偏向左上角。

Original comment in English

issue (bug_risk): Center point calculation for click_target is likely incorrect for [x1, y1, x2, y2] boxes.

With [x1, y1, x2, y2] boxes, center_x = box[0] + box[2] // 2 computes x1 + x2/2, not the midpoint between x1 and x2. If you want the rectangle center, it should be something like (box[0] + box[2]) // 2 and (box[1] + box[3]) // 2 (or float division for subpixel precision); otherwise the click will be biased toward the top-left.

[sourcery-ai]

issue: 对无效的 custom_action_param JSON 缺少错误处理，可能会导致未捕获的异常。

如果 argv.custom_action_param 缺失、为空或包含无效 JSON，json.loads 会抛出异常并中断该动作。建议捕获 json.JSONDecodeError，并根据对错误输入的期望行为，选择要么将其默认为一个空字典，要么返回一个失败的 RunResult 并记录清晰的日志信息。

Original comment in English

issue: Lack of error handling for invalid custom_action_param JSON may cause unhandled exceptions.

If argv.custom_action_param is missing, empty, or contains invalid JSON, json.loads will raise and terminate the action. Consider catching json.JSONDecodeError and either defaulting to an empty dict or returning a failed RunResult with a clear log message, depending on the desired behavior for bad input.

[sourcery-ai]

issue (typo): "注意的值" 这一表述有些不清晰，建议重新表述以提高可读性。

`注意的值

Suggested change

	- 注意的值"custom_action_param"
	- 请注意配置项 `custom_action_param` 的取值

Original comment in English

issue (typo): The phrase "注意的值" is a bit unclear; consider rephrasing for clarity.

`注意的值

Suggested change

	- 注意的值"custom_action_param"
	- 请注意配置项 `custom_action_param` 的取值