def queue_new_job(self, job_type, data, event, configuration=None):
"""Creates a new job for the given type and data. The new job is immediately placed on the queue. The new job,
job_exe, and queue models are saved in the database in an atomic transaction.
:param job_type: The type of the new job to create and queue
:type job_type: :class:`job.models.JobType`
:param data: The job data to run on
:type data: :class:`job.configuration.data.job_data.JobData`
:param event: The event that triggered the creation of this job
:type event: :class:`trigger.models.TriggerEvent`
:param configuration: The optional initial job configuration
:type configuration: :class:`job.configuration.configuration.job_configuration.JobConfiguration`
:returns: The new queued job
:rtype: :class:`job.models.Job`
:raises job.configuration.data.exceptions.InvalidData: If the job data is invalid
"""
job = Job.objects.create_job(job_type, event)
if not configuration:
configuration = JobConfiguration()
job.configuration = configuration.get_dict()
job.save()
# No lock needed for this job since it doesn't exist outside this transaction yet
Job.objects.populate_job_data(job, data)
self._queue_jobs([job])
return job
def queue_new_job(self, job_type, data, event, configuration=None):
"""Creates a new job for the given type and data. The new job is immediately placed on the queue. The new job,
job_exe, and queue models are saved in the database in an atomic transaction.
:param job_type: The type of the new job to create and queue
:type job_type: :class:`job.models.JobType`
:param data: The job data to run on
:type data: :class:`job.configuration.data.job_data.JobData`
:param event: The event that triggered the creation of this job
:type event: :class:`trigger.models.TriggerEvent`
:param configuration: The optional initial job configuration
:type configuration: :class:`job.configuration.configuration.job_configuration.JobConfiguration`
:returns: The new queued job
:rtype: :class:`job.models.Job`
:raises job.configuration.data.exceptions.InvalidData: If the job data is invalid
"""
job = Job.objects.create_job(job_type, event)
if not configuration:
configuration = JobConfiguration()
job.configuration = configuration.get_dict()
job.save()
# No lock needed for this job since it doesn't exist outside this transaction yet
Job.objects.populate_job_data(job, data)
self._queue_jobs([job])
return job
def test_populate_default_job_settings(self):
"""Tests the addition of default settings to the configuration."""
job_config = JobConfiguration()
config_dict = {
'version': '1.0',
'default_settings': {
'setting_name': 'some_val',
'setting2': 'other_val'
}
}
job_exe = MagicMock()
job_exe.get_job_type_configuration.return_value = JobTypeConfiguration(config_dict)
job_config.populate_default_job_settings(job_exe)
populated_config = job_config.get_dict()
populated_settings = populated_config['job_task']['settings']
populated_setting_values = [x.values() for x in populated_settings]
results_dict = {x[0]: x[1] for x in populated_setting_values}
self.assertTrue(results_dict == config_dict['default_settings'])
def populate_job_configuration(apps, schema_editor):
from job.configuration.configuration.job_configuration import JobConfiguration, MODE_RO, MODE_RW
from job.configuration.data.job_data import JobData
# Go through all of the job models that have job data and populate their configuration
Job = apps.get_model("job", "Job")
ScaleFile = apps.get_model("storage", "ScaleFile")
Workspace = apps.get_model("storage", "Workspace")
total_count = Job.objects.all().count()
workspaces = {}
for workspace in Workspace.objects.all().iterator():
workspaces[workspace.id] = workspace
print "Populating new configuration field for %s jobs" % str(total_count)
done_count = 0
batch_size = 1000
while done_count < total_count:
percent = (float(done_count) / float(total_count)) * 100.00
print "Completed %s of %s jobs (%f%%)" % (done_count, total_count, percent)
batch_end = done_count + batch_size
for job in Job.objects.select_related("job_type").order_by("id")[done_count:batch_end]:
# Ignore jobs that don't have their job data populated yet
if not job.data:
continue
data = JobData(job.data)
input_file_ids = data.get_input_file_ids()
input_files = ScaleFile.objects.filter(id__in=input_file_ids).select_related("workspace").iterator()
input_workspaces = set()
for input_file in input_files:
input_workspaces.add(input_file.workspace.name)
configuration = JobConfiguration()
for name in input_workspaces:
configuration.add_job_task_workspace(name, MODE_RO)
if not job.job_type.is_system:
for name in input_workspaces:
configuration.add_pre_task_workspace(name, MODE_RO)
# We add input workspaces to post task so it can perform a parse results move if requested by the
# job's results manifest
configuration.add_post_task_workspace(name, MODE_RW)
for workspace_id in data.get_output_workspace_ids():
workspace = workspaces[workspace_id]
if workspace.name not in input_workspaces:
configuration.add_post_task_workspace(workspace.name, MODE_RW)
elif job.job_type.name == "scale-ingest":
ingest_id = data.get_property_values(["Ingest ID"])["Ingest ID"]
from ingest.models import Ingest
ingest = Ingest.objects.select_related("workspace").get(id=ingest_id)
configuration.add_job_task_workspace(ingest.workspace.name, MODE_RW)
job.configuration = configuration.get_dict()
job.save()
done_count += batch_size
print "All %s jobs completed" % str(total_count)
def populate_job_configuration(apps, schema_editor):
from job.configuration.configuration.job_configuration import JobConfiguration, MODE_RO, MODE_RW
from job.configuration.data.job_data import JobData
# Go through all of the job models that have job data and populate their configuration
Job = apps.get_model('job', 'Job')
ScaleFile = apps.get_model('storage', 'ScaleFile')
Workspace = apps.get_model('storage', 'Workspace')
total_count = Job.objects.all().count()
workspaces = {}
for workspace in Workspace.objects.all().iterator():
workspaces[workspace.id] = workspace
print 'Populating new configuration field for %s jobs' % str(
total_count)
done_count = 0
batch_size = 1000
while done_count < total_count:
percent = (float(done_count) / float(total_count)) * 100.00
print 'Completed %s of %s jobs (%f%%)' % (done_count, total_count,
percent)
batch_end = done_count + batch_size
for job in Job.objects.select_related('job_type').order_by(
'id')[done_count:batch_end]:
# Ignore jobs that don't have their job data populated yet
if not job.data:
continue
data = JobData(job.data)
input_file_ids = data.get_input_file_ids()
input_files = ScaleFile.objects.filter(
id__in=input_file_ids).select_related(
'workspace').iterator()
input_workspaces = set()
for input_file in input_files:
input_workspaces.add(input_file.workspace.name)
configuration = JobConfiguration()
for name in input_workspaces:
configuration.add_job_task_workspace(name, MODE_RO)
if not job.job_type.is_system:
for name in input_workspaces:
configuration.add_pre_task_workspace(name, MODE_RO)
# We add input workspaces to post task so it can perform a parse results move if requested by the
# job's results manifest
configuration.add_post_task_workspace(name, MODE_RW)
for workspace_id in data.get_output_workspace_ids():
workspace = workspaces[workspace_id]
if workspace.name not in input_workspaces:
configuration.add_post_task_workspace(
workspace.name, MODE_RW)
elif job.job_type.name == 'scale-ingest':
ingest_id = data.get_property_values(['Ingest ID'
])['Ingest ID']
from ingest.models import Ingest
ingest = Ingest.objects.select_related('workspace').get(
id=ingest_id)
configuration.add_job_task_workspace(
ingest.workspace.name, MODE_RW)
job.configuration = configuration.get_dict()
job.save()
done_count += batch_size
print 'All %s jobs completed' % str(total_count)
