算子市场

from typing import Dict, Any

class MyTextCleaner:
    """自定义文本清洗算子"""

    def __init__(self, config: Dict[str, Any]):
        """
        初始化算子

        Args:
            config: 算子配置
                - remove_special_chars: 是否去除特殊字符
                - to_lowercase: 是否转为小写
        """
        self.remove_special_chars = config.get('remove_special_chars', True)
        self.to_lowercase = config.get('to_lowercase', True)

    def process(self, data: Any) -> Any:
        """
        处理数据

        Args:
            data: 输入数据

        Returns:
            处理后的数据
        """
        if isinstance(data, str):
            result = data
            if self.to_lowercase:
                result = result.lower()
            if self.remove_special_chars:
                # 去除特殊字符
                import re
                result = re.sub(r'[^\w\s]', '', result)
            return result
        return data

    def get_config_schema(self) -> Dict[str, Any]:
        """
        获取配置模式

        Returns:
            配置模式定义
        """
        return {
            "type": "object",
            "properties": {
                "remove_special_chars": {
                    "type": "boolean",
                    "title": "去除特殊字符",
                    "default": True
                },
                "to_lowercase": {
                    "type": "boolean",
                    "title": "转为小写",
                    "default": True
                }
            }
        }

import unittest

class TestMyTextCleaner(unittest.TestCase):
    def setUp(self):
        config = {
            'remove_special_chars': True,
            'to_lowercase': True
        }
        self.cleaner = MyTextCleaner(config)

    def test_lowercase(self):
        input_text = "Hello World"
        output = self.cleaner.process(input_text)
        self.assertEqual(output, "hello world")

    def test_remove_special_chars(self):
        input_text = "Hello!!! World..."
        output = self.cleaner.process(input_text)
        self.assertEqual(output, "hello world")