从当前目录导入大量含__init__.py的包的最Pythonic方法

Great question! Let's break this down properly—you're right that plain from . import * doesn't work for subpackages by default, but we have two solid Pythonic ways to handle this, depending on your use case.

方法1：显式定义__all__（推荐固定子包场景） #

The most Pythonic approach (aligning with the "explicit is better than implicit" mantra) is to define an __all__ list in your current directory's __init__.py. This tells Python exactly which subpackages to include when someone uses from . import *.

Add this to your root __init__.py:

__all__ = [
    'x1', 'x2', 'x3', 'x4', 'x5',
    'x6', 'x7', 'x8', 'x9', 'x10',
    'x11', 'x12', 'x13', 'x14', 'x15',
    'x16', 'x17', 'x18', 'x19', 'x20'
]

Then, in your code, you can import all 20 subpackages with a single line:

from . import *

This method is clean, readable, and gives you full control over what gets imported—other developers can immediately see which subpackages are exposed, no surprises.

方法2：动态自动导入（适合多变的子包集合） #

If your subpackage list changes often (you add/remove them regularly) and you don't want to manually update __all__ every time, you can use Python's importlib and pkgutil modules to dynamically discover and import all subpackages.

Put this in your root __init__.py:

import importlib
import pkgutil

__all__ = []

# Iterate over all modules/subpackages in the current directory
for _, pkg_name, is_package in pkgutil.iter_modules(__path__):
    if is_package:  # Only process actual subpackages
        # Dynamically import the subpackage
        imported_pkg = importlib.import_module(f'.{pkg_name}', __name__)
        # Add it to __all__ so it's included in wildcard imports
        __all__.append(pkg_name)
        # Optional: Bind the subpackage to the current namespace
        globals()[pkg_name] = imported_pkg

This will automatically import any subpackage in the directory, no manual list required. A few caveats though:

• It's implicit, so you might accidentally import subpackages you didn't intend (like test or utility subpackages) — you'll need to add filtering logic if that's a concern.
• There's a tiny startup overhead, but it's negligible for 20 subpackages.

Which is the Most Pythonic Choice? #

For your case with a fixed set of 20 subpackages, Method 1 is the clear winner. Explicitly listing them in __all__ makes your code transparent and follows Python's core philosophies. Reserve dynamic importing for scenarios where your subpackage list is truly dynamic and changes frequently.

Also, note that using from . import * is generally discouraged in regular modules, but it's widely accepted in a package's __init__.py as a way to create a clean, unified API for users of your package.

内容的提问来源于stack exchange，提问作者Ericson Willians