How to know who is importing me in python?

问题:

How can I find out what file is importing a particular file in python?

Consider the following example:

#a.py
import cmn
....

#b.py
import cmn
...

#cmn.py
#Here, I want to know which file (a.py or b.py)
#is importing this one.
#Is it possible to do this?
...

All the files a.py, b.py and cmn.py are in the same directory.

Why do I want to do this?
In C/C++, they have include feature. What i want to do can illuminate by the C/C++ code.

//a.cpp
....
#define SOME_STUFF   ....
#include "cmn.h"

//b.cpp
...
#define SOME_STUFF   ....

#include "cmn.h"

//cmn.h
//Here, I'll define some functions/classes that will use the symbol define
//in the a.cpp or b.cpp
...
....code refer to the SOME_STUFF.....

In C/C++, we can use this method to reuse sourecode.

Now return to my python code.
When a.py import cmn.py, i hope to run cmn.py and the cmn.py will refer to the symbol defined in the a.py.
When b.py import cmn.py, i hope to run cmn.py and the cmn.py will refer to the symbol defined in the b.py.

回答1:

The namedtuple code in the collections module has an example of how (and when) to do this:

#cmn.py
import sys
print 'I am being imported by', sys._getframe(1).f_globals.get('__name__')

One limitation of this approach is that the outermost module is always named __main__. If that is the case, the name of the outermost module can be determined from sys.argv[0].

A second limitation is that if the code using sys._getframe is in the module scope it is only executed on the first import of cmn.py. You'd need to call a function of some sort after imports if you want to monitor all imports of the module.

回答2:

Well, this is a kind of bizarre thing to do. You haven't explained why you want to know what is importing your module, so I can't actually help you solve your problem. You also haven't explained how or when you want to know the importing module.

def who_imports(studied_module):
    for loaded_module in sys.modules.values():
        for module_attribute in dir(loaded_module):
            if getattr(loaded_module, module_attribute) is studied_module:
                yield loaded_module

This will give you an iterator over all the modules which use your module as a top-level object. It won't find modules that do from cmn import *, and the list will change over time.

>>> import os
>>> for m in who_imports(os):
...     print m.__name__
... 
site
__main__
posixpath
genericpath
posixpath
linecache

回答3:

You'd need to install an import hook that tracks all imports. See PEP 302 and http://docs.python.org/dev/py3k/library/importlib.html. However, as the comments above point out, there is probably a better way to structure your code.