This is a continuation of the problems I had building and testing nupic. I
tried building nupic with both vagrant and the virtualenv.
1. Virtualenv
I could build nupic but $NTA not set error continues when running
testeverything. This is a set of env variables produced by "source
bin/activate" :
NuPIC environment
=================
BUILDDIR=/tmp/ntabuild
DYLD_LIBRARY_PATH=/home/kevin/Projects/nupic-env/nupic-linux64/lib64:/home/kevin/Projects/nupic-env/nupic-linux64/lib:/home/kevin/Projects/nupic-env/nupic-linux64/lib64:/home/kevin/Projects/nupic-env/nupic-linux64/lib:
LD_LIBRARY_PATH=/home/kevin/Projects/nupic-env/nupic-linux64/lib64:/home/kevin/Projects/nupic-env/nupic-linux64/lib:
MK_JOBS=
NUPIC=
NTA=/home/kevin/Projects/nupic-env/nupic-linux64/nta/eng
PATH=/home/kevin/Projects/nupic-env/nupic-linux64/bin:/home/kevin/local/bin:/usr/lib/x86_64-linux-gnu/qt4/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games
PYTHONPATH=/home/kevin/Projects/nupic-env/nupic-linux64/lib64/python2.7/site-packages:/home/kevin/Projects/nupic-env/nupic-linux64/lib/python2.7/site-packages:
As can be clearly seen, NTA is set. Another curious thing is that
PYTHONPATH points to python2.7 directories while there are only python2.6
directories in the folders lib and lib64! Is this a mistake? And yes, I
tried building nupic after renaming the directories but testeverything
still fails.
I do not really know what virtualenv is supposed to do but I believe it is
setting up a python environment to build nupic. import numpy inside the
python interactive shell produces an error even after activating the
virtualenv. Is this normal?
2. Vagrant
I installed all the dependencies using pip and my
nupic codebase was on /vagrant. When I ran the build script using sudo
./build.sh, the build hangs when it reaches here :
[vagrant@nupic-vagrant:/vagrant]$ sudo ./build.sh
In file included from
/usr/lib64/python2.6/site-packages/numpy/core/include/numpy/ndarraytypes.h:1728,
from
/usr/lib64/python2.6/site-packages/numpy/core/include/numpy/ndarrayobject.h:17,
from
/usr/lib64/python2.6/site-packages/numpy/core/include/numpy/arrayobject.h:15,
from math_py.cpp:2749:
/usr/lib64/python2.6/site-packages/numpy/core/include/numpy/npy_deprecated_api.h:11:2:
warning: #warning "Using deprecated NumPy API, disable it by #defining
NPY_NO_DEPRECATED_API NPY_1_7_API_VERSION"
In file included from /vagrant/nta/utils/Log.hpp:31,
from /vagrant/lang/py/support/PyHelpers.hpp:34,
from math_py.cpp:2707:
/vagrant/nta/utils/LoggingException.hpp: In member function
‘nta::LoggingException&
nta::LoggingException::operator<<(std::vector<unsigned int,
std::allocator<unsigned int> >)’:
/vagrant/nta/utils/LoggingException.hpp:66: warning: comparison between
signed and unsigned integer expressions
In file included from math_py.cpp:2735:
/vagrant/nta/math/utils.hpp: At global scope:
/vagrant/nta/math/utils.hpp:78: warning: expected [error|warning|ignored]
after ‘#pragma GCC diagnostic’
/vagrant/nta/math/utils.hpp:82: warning: expected [error|warning|ignored]
after ‘#pragma GCC diagnostic’
Another curious thing that I observed is that no matter which method i use
- normal, vagrant or virtualenv, the 'nta' folder goes to my home. Even if
$NTA is set to "/nupic-linux64/nta/eng" it goes to $HOME.
_______________________________________________
nupic mailing list
[email protected]
http://lists.numenta.org/mailman/listinfo/nupic_lists.numenta.org