Hi, we are trying to install a local version of Galaxy on a computing cluster using sge but there are a few issues with dependencies and setting paths. For instance we can get gsummary.py to run on the command line by manually setting PYTHONPATH to the RPy area, however within Galaxy the following error is returned:
File "/exports/work/biology_cell_galaxy/galaxy/tools/stats/gsummary.py", line 4, in ? from rpy import * ImportError: No module named rpy
Is there a preferred way to set the PYTHONPATH environment variable? I see references to it in setup_paths.sh, but I'm not sure when/if this gets run. It appears that the PYTHONPATH is ignored by setup.sh and by run.sh. run.sh calls python with environment disabled (-E). Even with -E removed the PYTHONPATH appears not to get through.
The error is the same on the command line with PYTHONPATH undefined. It is clear the problem is getting Galaxy to respect the PYTHONPATH.
Any help would be greatly appreciated.
Thanks Shaun Webb