如何检查导入python模块的原因？

Question

I have a big library of python modules. Sometimes, when importing a module, I see that unexpected modules are being imported. For this, I have been using python -v , to see which modules are being imported. From the manpage:

-v     Print  a  message  each  time a module is initialized, showing the place
       (filename or built-in module) from which it is loaded.  When given twice,
       print a message for each file that is checked for when searching for a
       module.  Also provides information on module cleanup at exit.

Well, this is not true. For example:

import portalmq # directory /home/blahblah/python_modules/portalmq
# /home/blahblah/python_modules/portalmq/__init__.pyc matches /home/blahblah/python_modules/portalmq/__init__.py
import portalmq # precompiled from /home/blahblah/python_modules/portalmq/__init__.pyc

As you can see, the -v flag just gives me information about which modules are imported, but not about which import statement, in which file/line is triggering the import. Using -vv does not change anything (a list of tried modules is shown, but nothing about why the import is tried in the first place)

But I need to know exactly that: which import statement, in which file/line is triggering those imports. How can I get this information?

Answer 1

Import hooks! Just add this code to your main script entry point to track every import after the sys.meta_path.append is executed.

import traceback

class TracingFinder:
    def find_module(self, fullname, path=None):
        print 'loading module', fullname
        traceback.print_stack()

import sys
sys.meta_path.append(TracingFinder())

Test:

def foo():
    import test
    import this

foo()

Output:

loading module test
  File "moo.py", line 15, in <module>
    foo()
  File "moo.py", line 12, in foo
    import test
  File "moo.py", line 6, in find_module
    traceback.print_stack()
loading module this
  File "moo.py", line 15, in <module>
    foo()
  File "moo.py", line 13, in foo
    import this
  File "moo.py", line 6, in find_module
    traceback.print_stack()