RunImageJ on cluster

Hi
We would like to be able to run pipelines that use the module RunImageJ on a cluster. As a first attempt i downloaded the latest source code from github.com/CellProfiler/CellProfiler/ and installed it on my local Linux machine to try it out. CellProfiler 2.0.0 (rev 40aad83) seems to be working, however when I tried to add the RunImageJ module i couldn’t find it in the list. On my Mac, if I list all, it appears between Resize and SaveImages. However on the linux machine it is missing. If I look in the folder CellProfiler/cellprofiler/modules I find the file run_imagej.py and the compiled run_imagej.pyc. So why can’t I find it in the list?

Martin

I am still very confused. I tried installing an older version r10996, however the RunImageJ module still doesn’t show up in the list. When I check the modules folder the run_imagej.py file is present and it is also listed in the init.py in the same folder. Nevertheless, no RunImageJ module to use within the software.

What am I missing? Do you have to change any settings or edit some file to make it work? Would very much appreciate some guidance.

It should be working for any recent checkout from the master branch. We use RunImageJ all the time on our Linux cluster. I’m wondering if there’s some problem with CellProfiler finding the JDK in order to compile the Python / Java bridge which uses JNI.

Do you see any console output that might indicate a problem? Also, you might try starting Python with your PYTHONPATH pointing to CellProfiler and try to start the JVM from within Python. For instance:

python

import cellprofiler.uitilities.jutil as J
J.start_vm(])
J.to_string(2+2)
‘4’

The last possibility is that some of the .jar files are missing. You should see the following files (at least) in the imagej subdirectory:
ij.jar
imglib.jar
javacl-1.0-beta-4-shaded.jar
junit-4.5.jar
precompiled_headless.jar
imagej-2.0-SNAPSHOT-all.jar

–Lee

Thanks, that sounds reasonable. I will look into that.