lemmings.chain package

Submodules

lemmings.chain.chain module

class lemmings.chain.chain.Lemmings(lemmings_job: LemmingJobBase)

Bases: object

Lemmings assemble - lemmingJob : the workflow to do - machine : the machine file, specialized on the queues - database : a persistent information for lemmings

next()

Execute all necessary functions depending on its status There are 2 kind of function: - Functions that check some conditions - Functions that pass from a status to another

         - - - > spawn_job - - - -
        |             ^             |
        |             |             V
      start             - - - - - post_job
<check_condition>             <check_condition>
        |                           |
        |                           |
        |                           |
         - - - - - > Exit < - - - -

run(): Submit the chain of computations.

lemmings.chain.common module

lemmings.chain.database module

lemmings.chain.farming module

Farming functionalities for the lemmings object

lemmings.chain.farming.continue_farming(lemmings_job: LemmingJobBase)

Function that monitors the number of parallel jobs currently run

Idea is to update the job status in the main database.json. Status should be either: start, hold, end So: 1) I must be able to access the main database.json 2) I must be able to submit new lemming jobs 3) I must update database status

lemmings.chain.farming.run_farming(lemmings_job: LemmingJobBase)

Method that controls the generation of multiple workflows for parallel mode and submits them according to user settings

Split in smaller functions 1) check if all is well activated in workflow.yml -> _check_correct_parallel_settings() 2) perform the workflow copies -> _generate_workflow_replicates 3) launch workflows 3.1) check if max parallel workflow specied 3.2) launch workflows 3.3) update the database -> _launch_workflows_parallel 4) raise LemmingsStop as we did what we had to do at this point

lemmings.chain.lemmingjob_base module

The lemmings base class containing the methods which get inherited by LemmingJob

class lemmings.chain.lemmingjob_base.LemmingJobBase(workflow: str, machine: Machine, database: Database, path_yml: str, job_prefix: str, user: NamedTuple, loop_count: int, status='start', base_dir=None)

Bases: object

Class containing the different lemmings methods and allowing access to database and other information

abort_on_start(): What lemmings does if the criterion is reached in the first loop.

after_end_job(): Actions just before lemmings ends.

check_on_end()

Verifications after each job loop

The function check_on_end needs to return a boolean (default True) with three options:

False: we continue lemmings

True: target reached, we stop lemmings (default setting)

None: crash, we stop lemmings

Default verification by lemmings:

is the cpu condition (.yml file) reached?

check_on_start()

Verify if the condition is already satisfied before launching a lemmings chain.

Function returns a boolean which starts the chain run. Default set to True.

A minimum required action is to set the ‘start_cpu_time’ so that lemmings can check if the max cpu condition is reached.

exit(msg): Stop the workflow

prepare_run(): Prepare the run before submission.

prior_to_job(): Function that prepares the run when the user launches the Lemmings command.

prior_to_new_iteration(): Prepare the new loop specific actions if criterion is not reached.

set_progress_name(name): Update progress of the workflow with a custom value

set_progress_var(value): Update progress of the workflow with a custom value

lemmings.chain.lemmingjob_base.read_farming_params(): Load locally the farming_params, returns None if no file is available