init datamate

runtime/ops/mapper/traditional_chinese/__init__.py

@@ -0,0 +1,6 @@
+# -*- coding: utf-8 -*-
+
+from datamate.core.base_op import OPERATORS
+
+OPERATORS.register_module(module_name='TraditionalChineseCleaner',
+                          module_path="ops.mapper.traditional_chinese.process")

runtime/ops/mapper/traditional_chinese/metadata.yml

@@ -0,0 +1,16 @@
+name: '繁体转简体'
+name_en: 'Traditional-Simplified Chinese Conversion'
+description: '将繁体转换为简体。'
+description_en: 'Converts traditional Chinese characters to simplified Chinese characters.'
+language: 'python'
+vendor: 'huawei'
+raw_id: 'TraditionalChineseCleaner'
+version: '1.0.0'
+types:
+  - 'cleanse'
+modal: 'text'
+effect:
+  before: '華為的業務涵蓋了從通信設備到智能手機、企業網絡解決方案以及雲計算等多個領域。'
+  after: '华为的业务涵盖了从通信设备到智能手机、企业网络解决方案以及云计算等多个领域。'
+inputs: 'text'
+outputs: 'text'

runtime/ops/mapper/traditional_chinese/process.py

@@ -0,0 +1,33 @@
+#!/user/bin/python
+# -*- coding: utf-8 -*-
+
+"""
+Description: 繁体转简体
+Create: 2025/01/15
+"""
+import time
+from typing import Dict, Any
+
+from zhconv import convert
+from loguru import logger
+
+from datamate.core.base_op import Mapper
+
+
+class TraditionalChineseCleaner(Mapper):
+    """繁体转简体过滤插件"""
+
+    @staticmethod
+    def _traditional_chinese_filter(input_data: str):
+        """ 繁体转简体"""
+        res = []
+        for input_str in input_data.split('\n'):
+            res.append(convert(input_str, 'zh-hans'))
+        return '\n'.join(res)
+
+    def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
+        start = time.time()
+        sample[self.text_key] = self._traditional_chinese_filter(sample[self.text_key])
+        logger.info(
+            f"fileName: {sample[self.filename_key]}, method: TraditionalChinese costs {time.time() - start:6f} s")
+        return sample