Closed rbdavid closed 2 years ago
Moved codes into respective python files; dask pipeline stuff to the dask_tskmgr.py, science functions to minimization.py, and basic logging functions to logging_functions.py. Calling the minimization.py code within a subprocess call in the dask_tskmgr.py.
Debugged numerous issues. Now, the only bug is the implementation of the dask wokerplugin to log times and process ids to individual files for each worker.
Dask workerplugin has been deemed unnecessary. More testing may get it to work. Cleaned code. Created a README file. Updated the submit script.
Mark suggests that the minimization pipeline code should be moved to a separate python script that is then called within a subprocess function w/in the tskmgr script. This will ensure desired memory clean up between tasks. This separation will also improve the organization of the current code by separating the science from the pipeline.