python如何跨文件调用自己定义的函数

        当自己定义函数过多时，只有一个python文件时代码会很长，不易理清代码框架，比如下面这段代码，如何隐藏具体函数细节呢？也就是把def函数放到另外一个python文件里步骤如下：

一个python文件代码篇幅过长

import os
import gzip
import numpy as np

def load_mnist_gz(image_file_path, label_file_path, flatten=True, normalize=False):
    # 定义内部函数来加载标签
    def load_labels(file_path):
        with gzip.open(file_path, 'rb') as f:
            f.read(8)  # 跳过文件头
            labels = np.frombuffer(f.read(), dtype=np.uint8, offset=0)
            return labels
    
    # 定义内部函数来加载图像
    def load_images(file_path):
        with gzip.open(file_path, 'rb') as f:
            f.read(16)  # 跳过文件头
            images = np.frombuffer(f.read(), dtype=np.uint8, offset=0).reshape(-1, 784)
            if normalize:
                images = images / 255.0
            return images
    
    # 使用内部函数加载数据
    t_labels = load_labels(label_file_path)
    x_images = load_images(image_file_path)
    
    return x_images, t_labels

# 设置数据集文件路径
train_image_file_path = r'D:\myproject\PyTorch-MNIST-master\DataSets\MNIST\raw\train-images-idx3-ubyte.gz'
train_label_file_path = r'D:\myproject\PyTorch-MNIST-master\DataSets\MNIST\raw\train-labels-idx1-ubyte.gz'
test_image_file_path = r'D:\myproject\PyTorch-MNIST-master\DataSets\MNIST\raw\t10k-images-idx3-ubyte.gz'
test_label_file_path = r'D:\myproject\PyTorch-MNIST-master\DataSets\MNIST\raw\t10k-labels-idx1-ubyte.gz'

# 使用定义的函数加载数据集
x_train, t_train = load_mnist_gz(train_image_file_path, train_label_file_path, flatten=True, normalize=False)
x_test, t_test = load_mnist_gz(test_image_file_path, test_label_file_path, flatten=True, normalize=False)

# 打印数据集的形状
print(x_train.shape)  # (60000, 784)
print(t_train.shape)  # (60000,)
print(x_test.shape)   # (10000, 784)
print(t_test.shape)   # (10000,)

 调整后易于理解框架：

1.首先要新建文件夹，因为有两个python文件，必须要有文件夹

2.定义两个python文件，需要注意的时封装的函数也要导入相关的库比如

import os
import gzip
import numpy as np
def load_mnist_gz(image_file_path, label_file_path, flatten=True, normalize=False):
    def load_labels(file_path):
        with gzip.open(file_path, 'rb') as f:
            f.read(8)  # 跳过文件头
            labels = np.frombuffer(f.read(), dtype=np.uint8, offset=0)
            return labels
    def load_images(file_path):
        with gzip.open(file_path, 'rb') as f:
            f.read(16)  # 跳过文件头
            images = np.frombuffer(f.read(), dtype=np.uint8, offset=0).reshape(-1, 784)
            if normalize:
                images = images / 255.0
            return images
    t_labels = load_labels(label_file_path)
    x_images = load_images(image_file_path)
    return x_images, t_labels

如果没有上面的三个库也会报错，原因是这个函数需要用到上面的那三个库，而不是简单的把def函数内容复制过来

3.在主函数里面加入要导入的自定义函数，这个例子是导入

from dataset.minst import load_mnist_gz

自定义函数在dataset文件夹下的minst.py中，然后导入自定义的函数名比如load_minst_gz即可，当相对路径导入报错时候就导入绝对路径即可