Python path: why can't pip see latest version of numpy

Question

Working inside a virtualenv, I'm trying to install scikit-image, but it fails, telling me that I need a newer (or any) version of numpy.

$ sudo pip install -U scikit-image
...
ImportError: You need `numpy` version 1.6 or later.
----------------------------------------
Command python setup.py egg_info failed with error code 1 in /Users/aps/Dropbox/experiments/build/scikit-image

I don't understand this, because from Python, numpy 1.7 is already on my path.

$ python 
Python 2.7.1 (r271:86832, Jul 31 2011, 19:30:53) 
[GCC 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2335.15.00)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> numpy.__version__
'1.7.1'

What is going on - why can't the system installing scikit-image see the newest version of numpy? And how can I fix this?

The path to numpy already seems to be in my ~/.bash_profile :

>>> numpy.__path__
['/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/numpy']
>>> exit()
$ vi ~/.bash_profile
...
export PYTHONPATH=/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages:$PYTHONPATH

UPDATE:

numpy version 1.7 is also in the su path:

$ sudo python 
Password:
Python 2.7.5 (default, Aug  1 2013, 00:59:40) 
[GCC 4.2.1 Compatible Apple Clang 4.1 ((tags/Apple/clang-421.11.66))] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> numpy.__version__
'1.7.1'

Answer 1

You need to make sure that all those packages are available within the virtualenv itself. Did you try running pip freeze after sourcing bin/activate ? Using the virtualenv package, the only library that comes "preinstalled" by default is distribute , and all other dependencies must be manually added. With venv (built into py3k) all of your system libraries should be available directly. Which are you using?