[PYTHON] Hook to the first import of the module and print the module path

Hook to the first import and print the module path

TL;DR

• In Python2.7.14 [^ 1], I wanted a mechanism to hook and execute after ** first Import ** of a module.
• There should be a smarter way in Python 3 series.
• Created a script that displays the file path of the import source when the module is imported for the first time.
• This content has nothing to do with Autodesk Maya, For problems that come with (feel) on tools that use the Python interpreter, such as Maya, Houdini, and Blender. I'm tagging it because it seems to be one answer.

motivation

• For projects that have been in operation for a long time, the number of in-house tools will increase. By checking the usage status of tools, it is possible to delete tools that are no longer needed to some extent, If the library or design is bad, you will not be satisfied with even grasping the dependency of the script file. If you delete it, you will not be able to delete it because you are afraid that the environment will be destroyed [^ 2].
• As the number of unnecessary script files increases, the cost of maintaining and handing over unnecessary files will increase, but the number of manpower will be insufficient.
• By creating a mechanism to hook at the first import and totaling the usage status of each file, I want to reduce unnecessary costs by finding unnecessary script files and making them easy to delete.
• By the way, I also want to know the source of the module.

[^ 1]: Exactly on Maya's Python interpreter. [^ 2]: Censored.

code

firstimport_hook.py

# -*- coding: utf-8 -*-
from __future__ import unicode_literals

import __builtin__
import types
import copy
import sys

import inspect

# reload()For countermeasures__import__Store after confirming that is Builtin Function.
if isinstance(__import__, types.BuiltinFunctionType):
    _original__import__ = copy.deepcopy(__import__)

#For example, if it is imported hooked by another module,
# _original__import__Can not be defined, so throw an error appropriately
elif '_original__import__' not in globals():
    raise ImportError('Definition of _original__import__ failed.')


def __get_modulefile_from_load_module():
    try:
        #Get the module file to read from
        currentframe = inspect.currentframe()
        module_file = currentframe.f_back.f_back.f_code.co_filename
    finally:
        #Del to correct the reference count.
        del currentframe

    return module_file


def print_modulefile_post_first_import(*args, **kwargs):
    """View module files after initial import"""

    #Get module loading status before importing
    loaded_module_names = sys.modules.keys()

    # __import__Run
    module = _original__import__(*args, **kwargs)

    #Do not process modules that have already been loaded
    if module.__name__ in loaded_module_names:
        return module

    # __file__Do not show the path of modules that do not have
    filepath = getattr(module, '__file__', None)
    if filepath:
        path_from_load_module = __get_modulefile_from_load_module()

        print('%s -> %s' % (path_from_load_module, filepath))

    return module


def enable_import_hook():
    """Enable import hooks"""
    __builtin__.__import__ = print_modulefile_post_first_import


def disable_import_hook():
    """Disable import hooks"""
    __builtin__.__import__ = _original__import__


if __name__ == '__main__':
    #Enable import hook
    enable_import_hook()

    import json
    reload(json)

startup.py

import firstimport_hook
firstimport_hook.enable_import_hook()

Finally

Now, when you import a Python file for the first time, you can print () the path of the loaded module. After that, if you send it to the server appropriately, it will be possible to aggregate.

mel? You don't know ...