jade.jobs.job_submitter.JobSubmitter

class jade.jobs.job_submitter.JobSubmitter(config_file, output, is_new)

Bases: JobManagerBase

Submits jobs for execution locally or on an HPC.

Internal constructor. Callers should use create() or load().

Methods

cancel_jobs(cluster)	Cancel running and pending jobs.
create(config[, output])	Creates a new instance.
find_error_log_messages(directory)	Parse output log files for error messages
generate_reports(directory, ...)	Create reports summarizing the output results of a set of jobs.
get_batch_post_process_config()	Get the batch post-process config data
get_completed_results()	Get results of completed jobs.
get_num_jobs()	Get the number of jobs to execute.
get_results_summmary()	Get a job results.
get_results_summmary_report()	Get a summary of job results in text form.
load(output)	Loads an instance from an existing directory.
run_checks()	Checks the configuration for errors.
run_submit_jobs(config, output[, local, ...])	Allows submission from an existing Python process.
submit_jobs(cluster[, force_local])	Submit simulations.
write_results_summary(filename, missing_jobs)	Write the results to filename in the output directory.

Attributes

config

classmethod create(config, output='output')

Creates a new instance.

Parameters:

• config (JobConfiguration)

• output (str) – Output directory

classmethod load(output)

Loads an instance from an existing directory.

cancel_jobs(cluster)

Cancel running and pending jobs.

submit_jobs(cluster, force_local=False)

Submit simulations. Auto-detect whether the current system is an HPC and submit to its queue. Otherwise, run locally.

Parameters:

• cluster (Cluster)

• force_local (bool) – If on HPC, run jobs through subprocess as if local.

Return type:

Status

write_results_summary(filename, missing_jobs)

Write the results to filename in the output directory.

static find_error_log_messages(directory)

Parse output log files for error messages

Parameters:

directory (str) – output directory

static generate_reports(directory, resource_monitor_type)

Create reports summarizing the output results of a set of jobs.

Parameters:

• directory (str) – output directory

• resource_monitor_type (ResourceMonitorType)

run_checks()

Checks the configuration for errors. May mutate the config.

get_batch_post_process_config()

Get the batch post-process config data

get_completed_results()

Get results of completed jobs.

Return type:

list of Result

get_num_jobs()

Get the number of jobs to execute.

Return type:

int

get_results_summmary()

Get a job results.

Return type:

dict

get_results_summmary_report()

Get a summary of job results in text form.

Return type:

str

static run_submit_jobs(config, output, local=False, dry_run=False, pipeline_stage_num=None)

Allows submission from an existing Python process.