[ Python 常見問題 ] How to check if a python module exists without importing it

Source From Here
Question
I need to know if a python module exists, without importing it. Importing something that might not exist (not what I want):

view plaincopy to clipboardprint?

1. try:  
2.     import eggs  
3. except ImportError:  
4.     pass  

How-To
To check if import can find something in python2, using imp:

view plaincopy to clipboardprint?

1. # -*- coding: utf-8 -*-    
2. import imp  
3.   
4. def chkModExist(mn):  
5.     try:  
6.         imp.find_module(mn)  
7.         return True  
8.     except ImportError:  
9.         return False  
10.       
11. for mn in ['egg', 'selenium', 'os']:  
12.     print "Module '{0}' exist? {1}".format(mn, chkModExist(mn))  

To find dotted imports, you need to do more:

view plaincopy to clipboardprint?

1. import imp  
2. try:  
3.     spam_info = imp.find_module('spam')  
4.     spam = imp.load_module('spam', *spam_info)  
5.     imp.find_module('eggs', spam.__path__) # __path__ is already a list  
6.     found = True  
7. except ImportError:  
8.     found = False  

For a full path module (including dot), below is an example for reference:

view plaincopy to clipboardprint?

1. # -*- coding: utf-8 -*-    
2. import imp  
3. import os.path  
4.   
5. def chkModExist(mn, path=None):  
6.     try:  
7.         #print "Search moudle='{0}' from '{1}'...".format(mn, path)  
8.         if path is None:  
9.             m_info=imp.find_module(mn)  
10.         else:  
11.             m_info=imp.find_module(mn, [path])  
12.         return (True, m_info)  
13.     except ImportError:  
14.         return False  
15.   
16. md = {}  
17. print "\t[Info] Loading desired modules..."  
18. for mn in ['egg', 'selenium', 'os', 'demo.faq.CheckModules']:  
19.     mn_dot = mn.split(".")  
20.     lm=None  
21.     for sm in mn_dot:  
22.         if lm:  
23.             if "__init__.py" in lm.__file__:  
24.                 rt=chkModExist(sm, os.path.dirname(lm.__file__))              
25.         else:  
26.             rt=chkModExist(sm)      
27.         if rt:          
28.             m_info = rt[1]  
29.             m=imp.load_module(sm, *m_info)  
30.             print "\t\tLoading module...{0}/{1}".format(m, m.__file__)  
31.             lm=m  
32.         else:  
33.             print "\t\tModule '{0}' doesn't exist ({1})".format(mn, sm)              
34.             lm=None  
35.             break  
36.     if lm:  
37.         md[mn]=lm  
38.           
39. print "\t[Info] Leverage loaded module 'os' to show env:\n%s" % md['os'].environ  

Here "demo.faq.CheckModules" is a local module for testing. A execution sample output will be:

[Info] Loading desired modules...
Loading module.../C:\Python27\lib\site-packages\egg-0.2.0-py2.7.egg\egg\__init__.pyc
Module 'selenium' doesn't exist (selenium)
Loading module.../C:\Python27\lib\os.pyc
Loading module.../C:\John\Eclipse\PyLab\demo\__init__.pyc
Loading module.../C:\John\Eclipse\PyLab\demo\faq\__init__.pyc
Loading module.../C:\John\Eclipse\PyLab\demo\faq\CheckModules.pyc
[Info] Leverage loaded module 'os' to show env:
{'TMP': 'C:\\Users\\\xab\xb6\xbcy\\AppData\\Local\\Temp', ...}

Python3 wants you to use importlib, How I went about doing this was:

view plaincopy to clipboardprint?

1. import importlib  
2. spam_loader = importlib.find_loader('spam')  
3. found = spam_loader is not None  

My expectation being, if you can find a loader for it, then it exists. You can also be a bit more smart about it, like filtering out what loaders you will accept. For example

view plaincopy to clipboardprint?

1. import importlib  
2. spam_loader = importlib.find_loader('spam')  
3. # only accept it as valid if there is a source file for the module - no bytecode only.  
4. found = issubclass(type(spam_loader), importlib.machinery.SourceFileLoader)