def make_curve_params(self, oqparam):
# the CurveParams are used only in classical_risk, classical_bcr
# NB: populate the inner lists .loss_types too
cps = []
for l, loss_type in enumerate(self.loss_types):
if oqparam.calculation_mode in ('classical', 'classical_risk'):
curve_resolutions = set()
lines = []
allratios = []
for taxo in sorted(self):
rm = self[taxo]
rf = rm.risk_functions.get((loss_type, 'vulnerability'))
if rf and loss_type in rm.loss_ratios:
ratios = rm.loss_ratios[loss_type]
allratios.append(ratios)
curve_resolutions.add(len(ratios))
lines.append('%s %d' % (rf, len(ratios)))
if len(curve_resolutions) > 1:
# number of loss ratios is not the same for all taxonomies:
# then use the longest array; see classical_risk case_5
allratios.sort(key=len)
for rm in self.values():
if rm.loss_ratios[loss_type] != allratios[-1]:
rm.loss_ratios[loss_type] = allratios[-1]
# logging.debug(f'Redefining loss ratios for {rm}')
cp = scientific.CurveParams(
l, loss_type, max(curve_resolutions), allratios[-1], True
) if curve_resolutions else scientific.CurveParams(
l, loss_type, 0, [], False)
else: # used only to store the association l -> loss_type
cp = scientific.CurveParams(l, loss_type, 0, [], False)
cps.append(cp)
self.lti[loss_type] = l
return cps
def make_curve_params(self, oqparam):
# the CurveParams are used only in classical_risk, classical_bcr
# NB: populate the inner lists .loss_types too
cps = []
loss_types = self._get_loss_types()
for l, loss_type in enumerate(loss_types):
if oqparam.calculation_mode in ('classical', 'classical_risk'):
curve_resolutions = set()
lines = []
for key in sorted(self):
rm = self[key]
if loss_type in rm.loss_ratios:
ratios = rm.loss_ratios[loss_type]
curve_resolutions.add(len(ratios))
lines.append(
'%s %d' %
(rm.risk_functions[loss_type], len(ratios)))
if len(curve_resolutions) > 1: # example in test_case_5
logging.info('Different num_loss_ratios:\n%s',
'\n'.join(lines))
cp = scientific.CurveParams(l, loss_type,
max(curve_resolutions), ratios,
True)
else: # used only to store the association l -> loss_type
cp = scientific.CurveParams(l, loss_type, 0, [], False)
cps.append(cp)
self.lti[loss_type] = l
return cps
