Run the star/gal recipe with the SLAC installation

[drphilmarshall]

Generating a confluence page on ho w to run at SLAC in the process, perhaps?

[djbard]

I've tested it with the slac installation, and it fails due to a problem with the location of the SED files in the phosim installation. I've asked Jim to add a link to the correct location, but I think it should be fine after that.

[jchiang87]

I've symlinked the SEDs folder to the desired locations in the /afs/slac installations.

[djbard]

This runs now, if you make sure to specify the full path for the output dir. Otherwise I presume phosim looks for the path relative to the release dir. Interestingly, that is not the case for the work dir.
Could someone else confirm this works before we close the ticket?

[jchiang87]

The phosim.py script tries to assign directories in a strange, rather dictatorial way. For any command line specification, e.g., , , etc., the script tests if the directory <phosim_install_dir>/<your_specified_dir> exists. If it does, then it uses that. If not, then it tries to use just <your_specified_dir>, which can be relative or absolute. So if you give an absolute path like /home/dbard/output, then it will test for <phosim_install_dir>//home/dbard/output, and since the "//" isn't a valid directory name, it will abandon trying to use a directory in the install dir and use the user-specified path. So if you gave "my_output" instead of just "output" as the output directory, it would have used the relative path to ./my_output (if it exists).

[djbard]

Hi Jim,
some of the formatting didn't come through in your post - let me check I understood. If I use "output_X" as the output dir, and that dir exists, phosim still fails to copy over the output fits files because it can't find the directory. When you say it would have used the relative path to ./output_X, do you mean relative to the install path? I'm curious that this doesn't work for the output dir, but it's fine for the work dir.

[jchiang87]

By "." I mean the current working directory. According to the assignPath and checkPaths functions in phosim.py, the code assigns the variables outputDir and workDir as I described. The reason it fails with a relative path is because phosim.py executes os.chdir(phosimDir) internally and then later writes to outputDir, without resolving it as an absolute path first. All of the directory handling in that script is messed up. For me, using workdir as a relative path (with outputdir as an absolute path) fails at the very end when it tries to clean up the working directory. In order to get everything to work end-to-end, users should specify only full absolute paths on the command line.

BTW, symlinking to the location of SEDs doesn't always work. There are cases where the code looks for a path relative to that SEDs directory, e.g., data/SEDs/../. So the contents of the SEDs directory need to be symlinked in to an actual SEDs subdirectory in the data subdir. :-P

[drphilmarshall]

Can we submit a bug report?

Dr. Phil Marshall

Kavli Institute for Particle Astrophysics and Cosmology,
SLAC National Accelerator Laboratory,
2575 Sand Hill Road, MS29,
Menlo Park, CA 94025, USA

Tel: +1 650 926 4682
http://www.slac.stanford.edu/~pjm
https://twitter.com/drphilmarshall

On Tue, Nov 26, 2013 at 7:31 AM, James Chiang notifications@github.comwrote:

By "." I mean the current working directory. According to the assignPath
and checkPaths functions in phosim.py, the code assigns the variables
outputDir and workDir as I described. The reason it fails with a relative
path is because phosim.py executes os.chdir(phosimDir) internally and then
later writes to outputDir, without resolving it as an absolute path first.
All of the directory handling in that script is messed up.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/5#issuecomment-29301419
.

[sethdigel]

I've got stargal.py running now on SLAC Linux. It is running in my terminal session - I have not tried on the batch farm yet. After some fiddling around, I found that I could make it go if I changed the definition of subcmd to point to the central installation of phosym.py:

subcmd = ["/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", newcatfilename,  "-c", "examples/nobackground", "-s", sensor, "-w", workdir, "-o", outdir]

If the output directories already exist, the script removes them and complains that it cannot find them (example complaint copied below). If they do not exist, then it creates them and runs without any error messages.

[digel@noric45 test]$ python stargal.py
running phosim over msstars for atmosphere realisation 1
Traceback (most recent call last):
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 743, in <module>
    main()
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 723, in main
    outputDir, workDir, binDir, dataDir, extraCommands, instrDir = checkPaths(opt, phosimDir)
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 89, in checkPaths
    'outputDir workDir binDir dataDir'.split()]    #see use of path[3] below
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 84, in assignPath
    raise RuntimeError('%s does not exist.' % path)
RuntimeError: output_msstars_R22_S21_atm1 does not exist.

[sethdigel]

I wrote too soon. stargal.py makes the output directory but I guess that phosim is not finding it. I'll try with absolute paths. There may be some more going on.

Reading out chip R22_S21 with amplifier chain R22_S21_C12
Reading out chip R22_S21 with amplifier chain R22_S21_C13
Reading out chip R22_S21 with amplifier chain R22_S21_C14
Reading out chip R22_S21 with amplifier chain R22_S21_C15
Reading out chip R22_S21 with amplifier chain R22_S21_C16
Reading out chip R22_S21 with amplifier chain R22_S21_C17
Process Process-2:
Traceback (most recent call last):
  File "/afs/slac/g/glast/ground/GLAST_EXT/redhat5-x86_64-64bit-gcc41/python/2.7.2/lib/python2.7/multiprocessing/process.py", line 258, in _bootstrap
    self.run()
  File "/afs/slac/g/glast/ground/GLAST_EXT/redhat5-x86_64-64bit-gcc41/python/2.7.2/lib/python2.7/multiprocessing/process.py", line 114, in run
    self._target(*self._args, **self._kwargs)
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 54, in jobChip
    shutil.move(rawImage+'.gz',rawImageRename)
  File "/afs/slac/g/glast/ground/GLAST_EXT/redhat5-x86_64-64bit-gcc41/python/2.7.2/lib/python2.7/shutil.py", line 298, in move
    copy2(src, real_dst)
  File "/afs/slac/g/glast/ground/GLAST_EXT/redhat5-x86_64-64bit-gcc41/python/2.7.2/lib/python2.7/shutil.py", line 127, in copy2
    copyfile(src, dst)
  File "/afs/slac/g/glast/ground/GLAST_EXT/redhat5-x86_64-64bit-gcc41/python/2.7.2/lib/python2.7/shutil.py", line 82, in copyfile
    with open(dst, 'wb') as fdst:
IOError: [Errno 2] No such file or directory: 'output_msstars_R22_S21_atm1/lsst_a_87393588_f2_R22_S21_C00_E001.fits.gz'
Traceback (most recent call last):
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 743, in <module>
    main()
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 740, in main
    focalplane.cleanup(bool(opt.keepscreens))
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 593, in cleanup
    os.chdir(self.workDir)
OSError: [Errno 2] No such file or directory: 'work_msstars_R22_S21_atm1'
running phosim over bdgals for atmosphere realisation 1
Traceback (most recent call last):
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 743, in <module>
    main()
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 723, in main
    outputDir, workDir, binDir, dataDir, extraCommands, instrDir = checkPaths(opt, phosimDir)
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 89, in checkPaths
    'outputDir workDir binDir dataDir'.split()]    #see use of path[3] below
  File "/afs/slac/g/lsst/software/redhat6-x86_64-64bit-gcc44/phoSim/phosim-3.3.2/phosim.py", line 84, in assignPath
    raise RuntimeError('%s does not exist.' % path)
RuntimeError: output_bdgals_R22_S21_atm1 does not exist.

[jchiang87]

Might as well reopen this since it seems that there are still lingering issues.

[jchiang87]

Bug report submitted:

https://www.lsstcorp.org:8480/browse/PHOSIM-2

[drphilmarshall]

Nice!

On Tuesday, November 26, 2013, James Chiang wrote:

Bug report submitted:

https://www.lsstcorp.org:8480/browse/PHOSIM-2

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/5#issuecomment-29331497
.

Dr. Phil Marshall

Kavli Institute for Particle Astrophysics and Cosmology,
SLAC National Accelerator Laboratory,
2575 Sand Hill Road, MS29,
Menlo Park, CA 94025, USA

Tel: +1 650 926 4682
http://www.slac.stanford.edu/~pjm
https://twitter.com/drphilmarshall