Python抽象工厂模式多文件循环导入ImportError问题修复求助

问题场景 #

项目采用多文件实现抽象工厂模式，核心文件结构如下：

Directory/
|-main.py
|-scrape/
  |--ScrapeFactory.py
  |--ScrapeFromAmazon.py

• ScrapeFactory.py 定义Scrape抽象类与ScrapeFactory抽象工厂类，同时导入ScrapeFromAmazon类
• ScrapeFromAmazon.py 导入ScrapeFactory类，实现ScrapeFromAmazon子类
• 运行Python 3.12.2代码时触发错误：

ImportError: cannot import name 'ScrapeFactory' from partially initialized module 'scrape.ScrapeFactory' (most likely due to a circular import)

• 已添加__init__.py但问题未解决

解决方法 #

1. 延迟导入（局部导入） #

将导入语句从模块顶部移到实际使用的函数/方法内部，避开模块初始化阶段的相互依赖：

# ScrapeFactory.py
from abc import ABC, abstractmethod

class Scrape(ABC):
    @abstractmethod
    def scrape_data(self):
        pass

class ScrapeFactory(ABC):
    @abstractmethod
    def create_scraper(self):
        pass

    # 在需要创建Amazon scraper的方法内延迟导入
    def create_amazon_scraper(self):
        from scrape.ScrapeFromAmazon import ScrapeFromAmazon
        return ScrapeFromAmazon()

ScrapeFromAmazon.py保持正常导入：

# ScrapeFromAmazon.py
from scrape.ScrapeFactory import Scrape, ScrapeFactory

class ScrapeFromAmazon(Scrape):
    def scrape_data(self):
        # 实现逻辑
        pass

2. 重构拆分抽象类与工厂类 #

把抽象类和工厂类拆分到独立文件，彻底打破循环依赖链，调整后文件结构：

Directory/
|-main.py
|-scrape/
  |--__init__.py
  |--base_scrape.py
  |--scrape_factory.py
  |--scrape_from_amazon.py

• base_scrape.py：仅定义Scrape抽象类

from abc import ABC, abstractmethod

class Scrape(ABC):
    @abstractmethod
    def scrape_data(self):
        pass

• scrape_factory.py：定义ScrapeFactory抽象工厂类，依赖base_scrape

from abc import ABC, abstractmethod
from .base_scrape import Scrape

class ScrapeFactory(ABC):
    @abstractmethod
    def create_scraper(self) -> Scrape:
        pass

• scrape_from_amazon.py：实现具体类，依赖抽象类和工厂类

from .base_scrape import Scrape
from .scrape_factory import ScrapeFactory

class ScrapeFromAmazon(Scrape):
    def scrape_data(self):
        # 实现逻辑
        pass

class AmazonScrapeFactory(ScrapeFactory):
    def create_scraper(self):
        return ScrapeFromAmazon()

3. 通过__init__.py中转导入 #

在scrape/__init__.py中统一导出所需类，让模块通过包间接导入：

# scrape/__init__.py
# 先导入抽象类与工厂类
from .ScrapeFactory import Scrape, ScrapeFactory
# 再导入具体实现类
from .ScrapeFromAmazon import ScrapeFromAmazon

修改ScrapeFromAmazon.py的导入方式：

# ScrapeFromAmazon.py
from scrape import Scrape, ScrapeFactory

class ScrapeFromAmazon(Scrape):
    # 实现逻辑
    pass

同时ScrapeFactory.py中如果不需要在模块初始化时使用ScrapeFromAmazon，可移除顶部导入，改用延迟导入或通过包导入。

核心原因 #

循环导入的本质是两个模块在初始化阶段互相尝试引用对方未完成初始化的内容，导致模块加载流程中断。上述方法均通过打破初始化时的直接依赖链解决问题。

内容的提问来源于stack exchange，提问作者Göktürk Sığırtmaç