En Mon, 05 Nov 2007 10:34:26 -0300, Frank Aune <[EMAIL PROTECTED]>
escribió:
> I have a python library package 'Foo', which contains alot of submodules:
>
> Foo/:
> __init__.py
> module1.py:
> class Bar()
> class Hmm()
> module2.py
> class Bee()
> class Wax()
> module3.py
> etc etc
>
> To prevent namespace pollution, I want to import and use this library in
> the
> following way:
>
> import Foo
> (...)
> t = Foo.module2.Bee()
>
> To accomplish this, I put explicit imports in __init__.py:
>
> import module1
> import module2
> import module3
>
> what Im wondering about, is if its a more refined way of doing this, as
> the
> explicit imports now need to be manually maintained if the library grows.
> I've tried to use __all__, but this only seems to work with "from Foo
> import
> *" and it causes modules to be imported directly into the namespace of
> course.
If I understand your question right, you want some way to automatically
enumerate and import all *.py files inside your package. Try this inside
Foo/__init__.py:
<code>
def import_all_modules():
"Import all modules in this directory"
import os.path
pkgdir = os.path.dirname(__file__)
for filename in os.listdir(pkgdir):
modname, ext = os.path.splitext(filename)
if ext=='.py' and modname!='__init__':
__import__(modname, globals())
import_all_modules()
del import_all_modules
</code>
--
Gabriel Genellina
--
http://mail.python.org/mailman/listinfo/python-list
