如何修改已保存Pickle文件中的导入库以适配更名后的依赖？

我之前把一个类实例保存成了Pickle文件，显示为<library.module.Class at 0x1c926b2e520>。现在这个类依赖的两个库已经更名了，想问能不能直接编辑Pickle文件，更新这两个导入语句，不用重新生成Pickle文件？

补充信息 #

加载Pickle的代码 #

import pickle
model_path = os.getenv('MODELS_FOLDER') + 'model_20210130.pkl'
model = pickle.load(open(model, 'rb'))

Pickle中类的原内容（需更新的导入已标注） #

# 需要更新的导入
import socceraction.spadl.config as spadlconfig
from socceraction.spadl.base import SPADLSchema

class ExpectedThreat:
    """An implementation of the model."""

    def __init__(self):
        ...
     

    def __solve(self) -> None:
        ...

    def fit(self, actions: DataFrame[SPADLSchema]) -> 'ExpectedThreat':
        """Fits the xT model with the given actions."""
        ...

    def predict(
        self, actions: DataFrame[SPADLSchema], use_interpolation: bool = False
    ) -> np.ndarray:
        """Predicts the model values for the given actions."""
        ...

不建议直接编辑Pickle文件来更新导入语句，核心原因如下：

• Pickle是二进制序列化格式，并非明文Python代码，文件里存储的是类的完整引用路径（比如socceraction.spadl.base.SPADLSchema），而非原始导入语句。直接修改二进制内容极易破坏文件结构，导致后续完全无法加载。
• 即便你能定位到对应的字符串并修改，类的引用路径在序列化时已与类定义强绑定，修改后大概率会触发AttributeError或ModuleNotFoundError——Python会尝试从新路径查找类，但新旧类的内部结构可能存在隐性差异，无法兼容旧序列化数据。

可行的替代方案 #

1. 模块路径映射
   在加载Pickle前，通过修改sys.modules将旧模块名指向新模块，用Python的模块机制兼容旧引用路径：

import sys
# 假设新库名为newsocceraction
import newsocceraction.spadl.config as new_spadlconfig
import newsocceraction.spadl.base as new_spadl_base

# 绑定旧模块名到新模块
sys.modules['socceraction.spadl.config'] = new_spadlconfig
sys.modules['socceraction.spadl.base'] = new_spadl_base

# 正常加载Pickle
import pickle
import os
model_path = os.getenv('MODELS_FOLDER') + 'model_20210130.pkl'
with open(model_path, 'rb') as f:
    model = pickle.load(f)

2. 自定义Unpickler
   如果模块映射不够灵活，可继承pickle.Unpickler重写find_class方法，在加载时动态替换旧类路径：

import pickle
import os
from newsocceraction.spadl.config import * as spadlconfig
from newsocceraction.spadl.base import SPADLSchema

class CustomUnpickler(pickle.Unpickler):
    def find_class(self, module, name):
        # 替换旧模块路径
        if module == 'socceraction.spadl.config':
            module = 'newsocceraction.spadl.config'
        elif module == 'socceraction.spadl.base':
            module = 'newsocceraction.spadl.base'
        return super().find_class(module, name)

model_path = os.getenv('MODELS_FOLDER') + 'model_20210130.pkl'
with open(model_path, 'rb') as f:
    model = CustomUnpickler(f).load()

3. 重新生成Pickle文件
   如果条件允许，用更新后的库重新初始化或训练类实例，再保存为新的Pickle文件——这是最稳妥的方案，能彻底避免后续兼容性问题。

内容的提问来源于stack exchange，提问作者Verance