是否可以在Python中的方法链中创建组/层？

Question

我知道我可以通过简单地让方法返回来使用方法链

self

，例如

object.routine1().routine2().routine3()

但是在应用方法链时是否可以将方法组织成层或组？例如

object.Layer1.routine1().routine2().Layer2.routine3()

上下文是我正在尝试构建一个文本分析管道，不同的层将对应于文本级别、句子级别和标记级别预处理步骤：

text = "This is an example foo text with some special characters!!!! And some sentences"

pr = TextPreprocessor(text)
processed_text = (
    pr.text_level.lower_case()
      .sentence_level.split_sentences().remove_special_characters()
      .token_level.tokenize()
      .text
)

这是几乎（！）使文本处理示例正常工作的代码：

import re

class TextLevelPreprocessor:
    def __init__(self, parent):
        self.parent = parent

    def lower_case(self):
        self.parent.text = self.parent.text.lower()
        return self.parent

class SentenceLevelPreprocessor:
    def __init__(self, parent):
        self.parent = parent

    def split_sentences(self):
        self.parent.text = self.parent.text.split('. ')
        return self.parent
    
    def remove_special_characters(self):
        self.parent.text = [re.sub('[!@#$]', '', s) for s in self.parent.text]
        return self.parent

class TokenLevelPreprocessor:
    def __init__(self, parent):
        self.parent = parent

    def tokenize(self):
        self.parent.text = [t.split() for t in self.parent.text]
        return self.parent

class TextPreprocessor:
    def __init__(self, text):
        self.text = text
        self.text_level = TextLevelPreprocessor(self)
        self.sentence_level = SentenceLevelPreprocessor(self)
        self.token_level = TokenLevelPreprocessor(self)

但是这里只有这个语法才有效

pr = TextPreprocessor(text)
processed_text = (
    pr.text_level.lower_case()
      .sentence_level.split_sentences().
      .sentence_level.remove_special_characters()
      .token_level.tokenize()
      .text
)

这意味着每次使用方法时都必须添加“层”或“组”，这看起来很冗长。

Answer 1

为了按类别组织方法，我会将它们放在单独的 mixin 类中，并使主类继承自 mixin 类。由于所有方法都在同一个实例上运行，因此不需要“父”对象。相反，mixin 类可以从初始化实例所有属性的基类继承：

class TextPreprocessorBase:
    def __init__(self, text):
        self.text = text

class TextLevelPreprocessor(TextPreprocessorBase):
    def lower_case(self):
        self.text = self.text.lower()
        return self

class SentenceLevelPreprocessor(TextPreprocessorBase):
    def split_sentences(self):
        self.text = self.text.split('. ')
        return self
    
    def remove_special_characters(self):
        self.text = [re.sub('[!@#$]', '', s) for s in self.text]
        return self

class TokenLevelPreprocessor(TextPreprocessorBase):
    def tokenize(self):
        self.text = [t.split() for t in self.parent.text]
        return self

class TextPreprocessor(TextLevelPreprocessor, SentenceLevelPreprocessor, TokenLevelPreprocessor):
    pass