Conda Recipes in support of Psychopy
Most dependencies were created with conda pipbuild depname
, which builds a binstar package that can be installed but doesn't create a recipe. This repo contains recipes that required special attention because they failed with pipbuild
. For example, the LabJack library uses dates for version names, where the dashes break version number parsing. The version can be specified manually, and there are a handful of similar work-arounds.
To install psychopy with conda, try this:
conda config --add channels https://conda.binstar.org/erik
conda config --add channels https://conda.binstar.org/quasiben
conda config --add channels https://conda.binstar.org/pramukta
conda create -n psychopyenv psychopy
source activate psychopyenv
The ones that could be auto-generated with pipbuild were run with:
# OSX Only
for i in pyobjc-core pyobjc pyobjc-framework-qtkit py2app; do
conda pipbuild $i;
done
for i in bdist_mpkg pyserial ioLabs pyxid altgraph macholib PySoundCard PySoundFile PyOpenGl altgraph ; do
conda pipbuild $i;
done
# For testing, you can also use `--use-local`, eg:
conda pipbuild altgraph
conda install altgraph --use-local
# labjackpython has non-standard versioning and can't use pipbuild
conda build labjackpython
Linux:
yum install gtk+-devel gtk+extra-devel gtk2-devel gtk2-engines-devel gtkglext-devel gtkmm24-devel wxGTK-devel wxBase SDL-devel gstreamer-devel gstreamer-plugins-base-devel
pygtk2.x86_64 pygtk2.x86_64 pygtkglext-devel.x86_64 python-gtkextra.x86_64