[Tutor] loading modules dynamically

[Danny Yoo]

On Mon, 10 Sep 2001, Girish Gajwani wrote:

> Hi all,
> 
> I have been trying to load modules dynamically using the
> __import__ hook , as follows:
> 
> #######################
> #! /usr/bin/env  python
> 
> def my_import(name):
>   mod = __import__(name)
>   print mod
>   print dir(mod)
>   return mod
> 
> mod = my_import("os")
> print mod.listdir(".")
> #######################
> 
> My interpretation of the __import__ hook is that it loads
> the module mentioned & also, aliases it to mod. Is this
> correct? I could not understand the documentation mentioned
> in the library reference wrt __import__, hence this
> question.

Yes, __import__ does load the module and return it:

###
>>> m = __import__('os')
>>> m
<module 'os' from '/home/dyoo/local/lib/python2.1/os.pyc'>
###

Note, though, that importing modules this way doesn't automatically allow
us to reference the os module with the name 'os':

###
>>> os
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
NameError: name 'os' is not defined
###




> Also, in what case can the above scene fail? What error
> checking would I need to add here?

When Python imports a module, and runs into an error, it will raise an
"ImportError".  For example:

###
>>> m = __import__('skljfflskj')
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
ImportError: No module named skljfflskj
###

What we can do is write a small exception handler that checks for this
situation.  Here's one version of my_import() that checks for
ImportErrors:


def my_import(name):
    try:
        mod = __import__(name)
    except ImportError:
        print "Whoops, I don't know about", name
        return None
    print mod
    print dir(mod)
    return mod


HOpe this helps!