def _set_up_outputs(self):
# Set up test Output records
self.uhs_output = models.Output(owner=self.uhs_job.owner,
oq_job=self.uhs_job,
db_backed=True,
output_type='uh_spectra')
self.uhs_output.save()
self.cpsha_hc_output = models.Output(owner=self.cpsha_job_fail.owner,
oq_job=self.cpsha_job_fail,
db_backed=True,
output_type='hazard_curve')
self.cpsha_hc_output.save()
self.cpsha_mean_hc_output = models.Output(
owner=self.cpsha_job_fail.owner,
oq_job=self.cpsha_job_fail,
db_backed=True,
output_type='hazard_curve')
self.cpsha_mean_hc_output.save()
self.cpsha_lc_output = models.Output(owner=self.cpsha_job_fail.owner,
oq_job=self.cpsha_job_fail,
db_backed=True,
output_type='loss_curve')
self.cpsha_lc_output.save()
def setUpClass(cls):
default_user = helpers.default_user()
cls.job = models.OqJob(owner=default_user)
cls.job.save()
# dmg dist per asset
cls.ddpa_output = models.Output(
owner=default_user, oq_job=cls.job,
display_name='Test dmg dist per asset',
output_type='dmg_dist_per_asset',
db_backed=True)
cls.ddpa_output.save()
cls.ddpa = models.DmgDistPerAsset(
output=cls.ddpa_output, dmg_states=cls.DMG_STATES)
cls.ddpa.save()
# We also need some sample exposure data records (to satisfy the dmg
# dist per asset FK).
test_input = models.Input(
owner=default_user, input_type='exposure', path='fake', size=0)
test_input.save()
i2j = models.Input2job(input=test_input, oq_job=cls.job)
i2j.save()
exp_model = models.ExposureModel(
owner=default_user, input=test_input, name='test-exp-model',
category='economic loss', stco_type='per_asset', stco_unit='CHF')
exp_model.save()
test_site = shapes.Site(3.14, 2.17)
cls.exp_data = models.ExposureData( # Asset
exposure_model=exp_model, asset_ref=helpers.random_string(),
taxonomy=helpers.random_string(), number_of_units=37,
site=test_site.point.to_wkt(), stco=1234.56)
cls.exp_data.save()
# dmg dist per taxonomy
cls.ddpt_output = models.Output(
owner=default_user, oq_job=cls.job,
display_name='Test dmg dist per taxonomy',
output_type='dmg_dist_per_taxonomy',
db_backed=True)
cls.ddpt_output.save()
cls.ddpt = models.DmgDistPerTaxonomy(
output=cls.ddpt_output, dmg_states=cls.DMG_STATES)
cls.ddpt.save()
# total dmg dist
cls.ddt_output = models.Output(
owner=default_user, oq_job=cls.job,
display_name='Test dmg dist total',
output_type='dmg_dist_total',
db_backed=True)
cls.ddt_output.save()
cls.ddt = models.DmgDistTotal(
output=cls.ddt_output, dmg_states=cls.DMG_STATES)
cls.ddt.save()
def insert_output(self, output_type):
"""Insert an `uiapi.output` record for the job at hand."""
assert self.output is None
job = models.OqJob.objects.get(id=self.oq_job_id)
# al-maisan, Fri, 29 Jun 2012 17:36:35 +0200
# https://bugs.launchpad.net/openquake/+bug/1019317
# figure out why using a single output record for gmf_data breaks the
# probablistic event-based risk calculator.
if output_type == "gmf":
one_or_none = []
else:
one_or_none = models.Output.objects.filter(oq_job=job,
display_name=basename(
self.nrml_path),
output_type=output_type)
if len(one_or_none) == 1:
self.output = one_or_none[0]
LOGGER.info("using output = '%s'", self.output)
else:
self.output = models.Output(owner=job.owner,
oq_job=job,
db_backed=True,
display_name=basename(self.nrml_path),
output_type=output_type)
self.output.save()
LOGGER.info("creating output = '%s'", self.output)
def make_dist(self):
output = models.Output(
owner=self.job.owner,
oq_job=self.job,
display_name="",
db_backed=True,
output_type="dmg_dist_total")
output.save()
self.ddt = models.DmgDistTotal(
output=output,
dmg_states=self.damage_states)
self.ddt.save()
def make_map(self):
output = models.Output(
owner=self.job.owner,
oq_job=self.job,
display_name="",
db_backed=True,
output_type="collapse_map")
output.save()
self.cm = models.CollapseMap(
output=output,
exposure_model=self.em)
self.cm.save()
def make_dist(self):
output = models.Output(
owner=self.job.owner,
oq_job=self.job,
display_name="",
db_backed=True,
output_type="dmg_dist_per_asset")
output.save()
self.dda = models.DmgDistPerAsset(
output=output,
dmg_states=self.damage_states)
self.dda.save()
def post_execute(self):
"""Perform the following post-execution actions:
* Write loss curves to XML
* Save the aggregate loss curve to the database
* Write BCR output (NOTE: If BCR mode, none of the other artifacts will
be written.
Not all of these actions will be executed; this depends on the
configuration of the job.
"""
if self.is_benefit_cost_ratio_mode():
self.write_output_bcr()
return
self.write_output()
# Save the aggregate loss curve to the database:
job = self.job_ctxt.oq_job
agg_lc_display_name = ('Aggregate Loss Curve for calculation %s' %
job.id)
output = models.Output(oq_job=job,
owner=job.owner,
display_name=agg_lc_display_name,
db_backed=True,
output_type='agg_loss_curve')
output.save()
loss_curve = models.LossCurve(output=output, aggregate=True)
loss_curve.save()
agg_lc_data = models.AggregateLossCurveData(
loss_curve=loss_curve,
losses=self.agg_curve.x_values,
poes=self.agg_curve.y_values)
agg_lc_data.save()
