Monkey patching -更改不会反映在其他模块中

Question

我想在不更改源代码的情况下编辑Python类中的方法(包lightgbm)。这个类被其他模块调用，但我所做的更改不会反映在这些模块中。

这是我的代码(我在self.__higher_better_inner_eval下面编辑)：

class Booster_fix(lightgbm.basic.Booster):

    def __get_eval_info(self):
        """Get inner evaluation count and names."""
        if self.__need_reload_eval_info:
            self.__need_reload_eval_info = False
            out_num_eval = ctypes.c_int(0)
            # Get num of inner evals
            _safe_call(_LIB.LGBM_BoosterGetEvalCounts(
                self.handle,
                ctypes.byref(out_num_eval)))
            self.__num_inner_eval = out_num_eval.value
            if self.__num_inner_eval > 0:
                # Get name of evals
                tmp_out_len = ctypes.c_int(0)
                reserved_string_buffer_size = 255
                required_string_buffer_size = ctypes.c_size_t(0)
                string_buffers = [
                    ctypes.create_string_buffer(reserved_string_buffer_size) for i in range_(self.__num_inner_eval)
                ]
                ptr_string_buffers = (ctypes.c_char_p * self.__num_inner_eval)(*map(ctypes.addressof, string_buffers))
                _safe_call(_LIB.LGBM_BoosterGetEvalNames(
                    self.handle,
                    ctypes.c_int(self.__num_inner_eval),
                    ctypes.byref(tmp_out_len),
                    ctypes.c_size_t(reserved_string_buffer_size),
                    ctypes.byref(required_string_buffer_size),
                    ptr_string_buffers))
                if self.__num_inner_eval != tmp_out_len.value:
                    raise ValueError("Length of eval names doesn't equal with num_evals")
                if reserved_string_buffer_size < required_string_buffer_size.value:
                    raise BufferError(
                        "Allocated eval name buffer size ({}) was inferior to the needed size ({})."
                        .format(reserved_string_buffer_size, required_string_buffer_size.value)
                    )
                self.__name_inner_eval = \
                    [string_buffers[i].value.decode('utf-8') for i in range_(self.__num_inner_eval)]
                self.__higher_better_inner_eval = \
                    [name.startswith(('auc', 'ndcg@', 'map@', 'average_precision')) for name in self.__name_inner_eval] # replacement row
                    # [name.startswith(('auc', 'ndcg@', 'map@')) for name in self.__name_inner_eval] # original row

lightgbm.basic.Booster = Booster_fix

在其他一些模块中，有如下调用：

from .basic import Booster

我猜我的猴子补丁不管用了。

基本上，我想实现这个修复：https://github.com/microsoft/LightGBM/pull/3649/commits/f73407f05e389a74e6f44a2cb9c637df6afdb33b，不等待发布的更新，也不编辑源代码。

你能帮我拿一下吗？

编辑：@Maurice，要测试代码，您可以运行：

import pandas as pd
from lightgbm import LGBMClassifier # lgbm version 3.1.1
from sklearn.model_selection import train_test_split

df = pd.read_csv("https://web.stanford.edu/class/archive/cs/cs109/cs109.1166/stuff/titanic.csv")
df['Sex'] = df['Sex'].astype('category')
X, y = df.drop(columns=['Name', 'Survived']), df['Survived']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
model = LGBMClassifier()
model.fit(X_train, y_train, early_stopping_rounds=10, eval_metric='average_precision', eval_set=(X_test, y_test))

Answer 1

您只需为lightgbm.basic.Booster的__get_eval_info打补丁

from lightgbm.basic import Booster


def patched__get_eval_info(self):
    print("patched ...")
    # your code here


Booster.__get_eval_info = patched__get_eval_info
b = Booster(model_file='model.txt')
b.__get_eval_info()

# ... other imports that use lightgbm follow here

输出：

patched ...