项目中有关于对Python包动态加载的需求，而且要求是不局限于当前目录下的全局导包，而且要能支持包更新，源程序同步更新
这里使用到了importlib这个包,先放出工具类
import importlib
import sys
import os
class NameConflictError(Exception):
def __init__(self, error_msg: str) -> None:
super().__init__(self)
self.error_msg = error_msg
def __str__(self) -> str:
return self.error_msg
class ModuleLoader(object):
@staticmethod
def get_module(module_name: str, package_path=None):
if package_path is None:
package_path = os.getcwd()
# Ensure wanted module import from wanted pkg
if package_path in sys.path:
sys.path.remove(package_path)
sys.path.insert(0, package_path)
mod = importlib.import_module(module_name)
exist_module_pkg_path = os.path.dirname(mod.__file__)
if package_path != exist_module_pkg_path:
mod = importlib.reload(mod)
# Read again
exist_module_pkg_path = os.path.dirname(mod.__file__)
if package_path != exist_module_pkg_path:
raise NameConflictError(
f"The module {package_path} not exist or priority lower than {exist_module_pkg_path}")
else:
return mod
@staticmethod
def get_member(module_name: str, func_name: str, package_path=None):
return getattr(ModuleLoader.get_module(module_name, package_path), func_name)
其实这边有几种导出场景是需要注意的，场景如下
| Pkg路径 | Module路径 | 函数名 | 结果 |
|---|---|---|---|
| ❌ | ❌ | ❌ | A包有则成功,否则报错 |
| ❌ | ❌ | ✅ | A包有则成功,否则报错 |
| ❌ | ✅ | ❌ | 假如A包有F函数则读A包的F函数，否则读B包的F函数,均无则报错 |
| ✅ | ❌ | ❌or ✅ | A包有则成功,否则报错 |
# test.A
def add():
return 100
def del():
return 500
# --------上下不同包--------
# try.A
def add():
return 200