init datamate

runtime/ops/mapper/emoji_cleaner/__init__.py

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

runtime/ops/mapper/emoji_cleaner/metadata.yml

@@ -0,0 +1,16 @@
+name: '文档表情去除'
+name_en: 'Emoticon Removal'
+description: '去除文档中表情字符或者emoji符号。'
+description_en: 'Removes emoticons or emojis from documents.'
+language: 'python'
+vendor: 'huawei'
+raw_id: 'EmojiCleaner'
+version: '1.0.0'
+types:
+  - 'cleanse'
+modal: 'text'
+effect:
+  before: '使用方式很简单，只需要将代码放入Markdown文本中即可，富文本格式可直接复制表情😀使用。'
+  after: '使用方式很简单，只需要将代码放入Markdown文本中即可，富文本格式可直接复制表情使用。'
+inputs: 'text'
+outputs: 'text'

runtime/ops/mapper/emoji_cleaner/process.py

@@ -0,0 +1,27 @@
+
+"""
+Description: 文档表情去除
+Create: 2023/12/7 15:43
+"""
+import time
+from typing import Dict, Any
+
+import emoji
+from loguru import logger
+
+from datamate.core.base_op import Mapper
+
+
+class EmojiCleaner(Mapper):
+    @staticmethod
+    def _emoji_filter(input_data: str):
+        res = []
+        for input_s in input_data.split('\n'):
+            res.append(emoji.replace_emoji(input_s, replace=''))
+        return '\n'.join(res)
+
+    def execute(self, sample: Dict[str, Any]) -> Dict[str, Any]:
+        start = time.time()
+        sample[self.text_key] = self._emoji_filter(sample[self.text_key])
+        logger.info(f"fileName: {sample[self.filename_key]}, method: EmojiCleaner costs {time.time() - start:6f} s")
+        return sample