Ctrl+K

playmolecule.apps package#

Module contents#

class playmolecule.apps.ExecutableDirectory(dirname, runsh='run.sh', _execution_resources=None)#

Bases: object

Executable directory class.

All app functions will create a folder and return to you an ExecutableDirectory object. This is a self-contained directory including all app input files which can be executed either locally or on a cluster. If it’s not executed locally make sure the directory can be accessed from all machines in the cluster (i.e. is located on a shared filesystem).

pmws(name=None, group=None, child_of=None, queue_config=None, pm_options=None, description=None, _logger=True)#

Submits the job to the PlayMolecule backend server.

Parameters:

  • child_of (str) – The id of another job. If provided, the new job will be submited as a child of that job.

  • queue_config (dict) – A dictionary containing key-value pairs for defailed configuration for this job on the queueing system. You can specify “cpus”, “memory” and “priority” for the job.

  • _logger (bool) – Set to False to reduce verbosity

Examples

>>> ed = proteinprepare(outdir="test", pdbid="3ptb")
>>> ed.pmws(queue_config={"ncpu": 2, "memory": 4000, "priority": 500})
run(queue=None, **kwargs)#

Execute the directory locally

If no queue is specified it will run the job locally.

Examples

>>> ed = proteinprepare(outdir="test", pdbid="3ptb")
>>> ed.run()

Specifying a queue

>>> ed.run(queue="slurm", partition="normalCPU", ncpu=3, ngpu=1)

Alternative syntax for

>>> ed.slurm(partition="normalCPU", ncpu=3, ngpu=1)
slurm(**kwargs)#

Submit simulations to SLURM cluster

Parameters:

  • partition (str or list of str) – The queue (partition) or list of queues to run on. If list, the one offering earliest initiation will be used.

  • jobname (str) – Job name (identifier)

  • priority (str) – Job priority

  • ncpu (int) – Number of CPUs to use for a single job

  • ngpu (int) – Number of GPUs to use for a single job

  • memory (int) – Amount of memory per job (MiB)

  • gpumemory (int) – Only run on GPUs with at least this much memory. Needs special setup of SLURM. Check how to define gpu_mem on SLURM.

  • walltime (int) – Job timeout (s)

  • mailtype (str) – When to send emails. Separate options with commas like ‘END,FAIL’.

  • mailuser (str) – User email address.

  • outputstream (str) – Output stream.

  • errorstream (str) – Error stream.

  • nodelist (list) – A list of nodes on which to run every job at the same time! Careful! The jobs will be duplicated!

  • exclude (list) – A list of nodes on which not to run the jobs. Use this to select nodes on which to allow the jobs to run on.

  • envvars (str) – Envvars to propagate from submission node to the running node (comma-separated)

  • prerun (list) – Shell commands to execute on the running node before the job (e.g. loading modules)

Examples

>>> ed = proteinprepare(outdir="test", pdbid="3ptb")
>>> ed.slurm(partition="normalCPU", ncpu=1, ngpu=0)
property status#

Returns current status of the ExecutableDirectory

Examples

>>> ed = proteinprepare(outdir="test", pdbid="3ptb")
>>> ed.slurm(ncpu=1, ngpu=0)
>>> print(ed.status)
class playmolecule.apps.JobStatus(value)#

Bases: IntEnum

Job status codes describing the current status of a job

  • WAITING_INFO : Waiting for status from the job. Job has not started yet computation.

  • RUNNING : Job is currently running

  • COMPLETED : Job has successfully completed

  • ERROR : Job has exited with an error

COMPLETED = 2#
ERROR = 3#
RUNNING = 1#
WAITING_INFO = 0#
describe()#
class playmolecule.apps.KWARGS#

Bases: dict

playmolecule.apps.slurm_mps(exec_dirs, **kwargs)#

Submit a list of ExecutableDirectories to SLURM as a single MPS job.

This means that all jobs submitted will be executed on the same GPU

Parameters:

  • exec_dirs (list[ExecutableDirectory]) – An iterable of ExecutableDirectory objects

  • partition (str or list of str) – The queue (partition) or list of queues to run on. If list, the one offering earliest initiation will be used.

  • jobname (str) – Job name (identifier)

  • priority (str) – Job priority

  • ncpu (int) – Number of CPUs to use for a single job

  • ngpu (int) – Number of GPUs to use for a single job

  • memory (int) – Amount of memory per job (MiB)

  • gpumemory (int) – Only run on GPUs with at least this much memory. Needs special setup of SLURM. Check how to define gpu_mem on SLURM.

  • walltime (int) – Job timeout (s)

  • mailtype (str) – When to send emails. Separate options with commas like ‘END,FAIL’.

  • mailuser (str) – User email address.

  • outputstream (str) – Output stream.

  • errorstream (str) – Error stream.

  • nodelist (list) – A list of nodes on which to run every job at the same time! Careful! The jobs will be duplicated!

  • exclude (list) – A list of nodes on which not to run the jobs. Use this to select nodes on which to allow the jobs to run on.

  • envvars (str) – Envvars to propagate from submission node to the running node (comma-separated)

  • prerun (list) – Shell commands to execute on the running node before the job (e.g. loading modules)

Examples

>>> ed1 = kdeep(outdir="test1", pdb=apps.kdeep.files["tests/10gs_protein.pdb"], sdf=apps.kdeep.files["tests/10gs_ligand.sdf"], modelfile=kdeep.datasets.default)
>>> ed2 = kdeep(outdir="test2", dataset=apps.kdeep.files["tests/dataset.zip"], modelfile=kdeep.datasets.default)
>>> slurm_mps([ed1, ed2], partition="normalGPU", ncpu=1, ngpu=1)
On this page