init datamate

runtime/ops/mapper/invisible_characters_cleaner/__init__.py

@@ -0,0 +1,7 @@
+# -*- coding: utf-8 -*-
+
+from datamate.core.base_op import OPERATORS
+
+OPERATORS.register_module(module_name='InvisibleCharactersCleaner',
+                          module_path="ops.mapper.invisible_characters_cleaner.process")
+

runtime/ops/mapper/invisible_characters_cleaner/metadata.yml

@@ -0,0 +1,16 @@
+name: '不可见字符去除'
+name_en: 'Invisible Character Removal'
+description: '去除文档中的不可见字符，例如 0-31 号字符中的部分字符。'
+description_en: 'Removes invisible characters from documents, for example, removing invisible characters from characters numbered 0 to 31.'
+language: 'python'
+vendor: 'huawei'
+raw_id: 'InvisibleCharactersCleaner'
+version: '1.0.0'
+types:
+  - 'cleanse'
+modal: 'text'
+effect:
+  before: "对“材料”怎样下\x04定义才臻于 严格和科学?"
+  after: '对“材料”怎样下定义才臻于严格和科学?'
+inputs: 'text'
+outputs: 'text'

runtime/ops/mapper/invisible_characters_cleaner/process.py

@@ -0,0 +1,30 @@
+#!/user/bin/python
+# -*- coding: utf-8 -*-
+
+"""
+Description: 不可见字符去除
+Create: 2025/01/13
+"""
+import re
+import time
+from typing import Dict, Any
+
+from loguru import logger
+
+from datamate.core.base_op import Mapper
+
+
+class InvisibleCharactersCleaner(Mapper):
+    @staticmethod
+    def _invisible_characters_filter(input_data: str):
+        # 移除ASCII中不可见字符，包括0-7、14-19 21-31、127-160的字符
+        invisible_char_pattern = '[\x00-\x07|\x0E-\x13|\x15-\x1F|\x7F-\xA0]'
+        invisible_chars_re = re.compile(invisible_char_pattern)
+        return invisible_chars_re.sub('', input_data)
+
+    def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
+        start = time.time()
+        sample[self.text_key] = self._invisible_characters_filter(sample[self.text_key])
+        logger.info(f"fileName: {sample[self.filename_key]}, "
+                    f"method: InvisibleCharactersCleaner costs {time.time() - start:6f} s")
+        return sample