gaps.hpc.HpcJobManager

class HpcJobManager(user=None, queue_dict=None)

Bases: ABC

Abstract HPC job manager framework

Parameters:

• user (str | None, optional) – HPC username. None will get your username using getpass.getuser(). By default, None.

• queue_dict (dict | None, optional) – Parsed HPC queue dictionary from parse_queue_str(). None will get the queue info from the hardware. By default, None.

Methods

cancel(arg)	Cancel a job.
check_status_using_job_id(job_id)	Check the status of a job using the HPC queue and job ID.
check_status_using_job_name(job_name)	Check the status of a job using the HPC queue and job name.
make_script_str(name, **kwargs)	Generate the submission script.
parse_queue_str(queue_str)	Parse the hardware queue string into a nested dictionary.
query_queue()	Run the HPC queue command and return the raw stdout string.
reset_query_cache()	Reset the query dict cache so that hardware is queried again.
submit(name[, keep_sh])	Submit a job on the HPC.

Attributes

COLUMN_HEADERS	Column header names.
COMMANDS	Command names.
MAX_NAME_LEN
Q_SUBMITTED_STATUS	String representing the submitted status for this manager.
SHELL_FILENAME_FMT
USER
queue	HPC queue keyed by job ids with values as job properties.

classmethod parse_queue_str(queue_str)

Parse the hardware queue string into a nested dictionary.

This function parses the queue output string into a dictionary keyed by integer job ids with values as dictionaries of job properties (queue printout columns).

Parameters:

queue_str (str) – HPC queue output string. Typically a space-delimited string with line breaks.

Returns:

queue_dict (dict) – HPC queue parsed into dictionary format keyed by integer job ids with values as dictionaries of job properties (queue printout columns).

property queue

HPC queue keyed by job ids with values as job properties.

Type:

dict

reset_query_cache()

Reset the query dict cache so that hardware is queried again.

check_status_using_job_id(job_id)

Check the status of a job using the HPC queue and job ID.

Parameters:

job_id (int) – Job integer ID number.

Returns:

status (str | None) – Queue job status string or None if not found.

check_status_using_job_name(job_name)

Check the status of a job using the HPC queue and job name.

Parameters:

job_name (str) – Job name string.

Returns:

status (str | None) – Queue job status string or None if not found.

cancel(arg)

Cancel a job.

Parameters:

arg (int | list | str) – Integer job id(s) to cancel. Can be a list of integer job ids, ‘all’ to cancel all jobs, or a feature (-p short) to cancel all jobs with a given feature

submit(name, keep_sh=False, **kwargs)

Submit a job on the HPC.

Parameters:

• name (str) – HPC job name.

• keep_sh (bool, optional) – Option to keep the submission script on disk. By default, False.

• **kwargs – Extra keyword-argument pairs to be passed to make_script_str().

Returns:

• out (str) – Standard output from submission. If submitted successfully, this is the Job ID.

• err (str) – Standard error. This is an empty string if the job was submitted successfully.

abstract query_queue()

Run the HPC queue command and return the raw stdout string.

Returns:

stdout (list) – HPC queue output string that can be split into a list on line breaks.

abstract make_script_str(name, **kwargs)

Generate the submission script.

abstract property COLUMN_HEADERS

Column header names.

Type:

namedtuple

abstract property COMMANDS

Command names.

Type:

namedtuple

abstract property Q_SUBMITTED_STATUS

String representing the submitted status for this manager.

Type:

str