When you name a "script" with the same name as a dependency, you have intercepted the lookup for that name.
You could modify your PYTHONPATH i.e. sys.path
, putting your dependencies in front of your "scripts" but then you've intercepted your scripts - this seems like a very bad idea.
>>> import sys
>>> import pprint
>>> pprint.pprint(sys.path)
['', # Current Working Directory
'~\\Anaconda3\\python36.zip',
'~\\Anaconda3\\DLLs',
'~\\Anaconda3\\lib',
'~\\Anaconda3',
'~\\Anaconda3\\lib\\site-packages',
'~\\Anaconda3\\lib\\site-packages\\Sphinx-1.5.6-py3.6.egg',
'~\\Anaconda3\\lib\\site-packages\\win32',
'~\\Anaconda3\\lib\\site-packages\\win32\\lib',
'~\\Anaconda3\\lib\\site-packages\\Pythonwin',
'~\\Anaconda3\\lib\\site-packages\\setuptools-27.2.0-py3.6.egg']
Don't modify the lookup order in your sys.path
.
I would not name any script with the same name as a dependency, with the sole exception that I was intending to intercept the name with something that is semantically the same thing.
Namespaces are one honking great idea -- let's do more of those!
You need your own namespace. Create a directory, for example a_namespace
, put your scripts in that directory, and work from the parent directory. For example:
>>> from pathlib import Path
>>> Path('a_namespace').mkdir()
>>> (Path('a_namespace') / 'sys.py').write_text('print("importing my sys")')
25
>>> import a_namespace.sys
importing my sys
And you can still get to your dependencies:
>>> import sys
>>> sys.version
'3.6.1 |Anaconda 4.4.0 (64-bit)| (default, May 11 2017, 13:25:24) [MSC v.1900 64 bit (AMD64)]'