![]() |
Rosetta
2019.07
|
Namespaces | |
worker | |
Functions | |
def | _normflags |
def | worker_extra |
def | init_notebook |
|
private |
Normalize tuple/list/str of flags into str.
References fmt.join(), and binder.split().
Referenced by pyrosetta.distributed.dask.init_notebook(), and pyrosetta.distributed.dask.worker_extra().
def pyrosetta.distributed.dask.init_notebook | ( | init_flags = None | ) |
Initialize PyRosetta with command line flags. For example, to execute in a Jupyter Notebook: import pyrosetta.distributed.dask pyrosetta.distributed.dask.init_notebook(init_flags)
References pyrosetta.distributed.dask._normflags(), and pyrosetta.distributed.maybe_init().
Referenced by pyrosetta.tests.distributed.test_dask_worker.TestDaskArgs.test_worker_extra().
def pyrosetta.distributed.dask.worker_extra | ( | init_flags = None , |
|
local_directory = None |
|||
) |
Format flags and local directory for dask worker preload. Convenience function that optionally accepts unformatted PyRosetta command line flags as the `init_flags` option, and a working directory for dask worker processes to output files as the `local_directory` option, which are then formatted into a dask worker process extra commands list. If `init_flags` is not specified, PyRosetta is initialized on the dask worker process without command line flags. If `local_directory` is not specified, the dask worker default directory ./dask-worker-space is used. For example, to execute on a SLURM scheduler via dask_jobqueue: from dask_jobqueue import SLURMCluster import pyrosetta.distributed.dask cluster = SLURMCluster( cores=1, processes=1, memory="4GB", extra=pyrosetta.distributed.dask.worker_extra(init_flags, local_directory) )
References pyrosetta.distributed.dask._normflags(), and pyrosetta.tests.distributed.test_dask.format.
Referenced by pyrosetta.tests.distributed.test_dask_worker.TestDaskArgs.test_worker_extra().