[Paolo Tormene (@ptormene), Antonio Ettorre (@antonioettorre), , Armando Scarpati (@hascar)]
- Improved installation documentation
from previous OpenQuake Engine 3.20.0 version:
[Paolo Tormene (@ptormene), Michele Simionato (@micheles)]
- Internal: added commands
oq compare assetcol
andoq compare sitecol
[Michele Simionato (@micheles)]
- Internal: added a command
oq reducexp exposure_xml site_model_csv
[Paul Henshaw (@pslh), Antonio Ettorre (@antonioettorre), Michele Simionato (@micheles)]
- Updated the documentation about installing the engine
[Michele Simionato (@micheles)]
- Raised the default
max_potential_paths
to 1,000,000 - Optimized the calculation of mean hazard curves when
use_rates=true
;
now it is possible to compute exactly mean curves even with millions
of realizations
[Paolo Tormene (@ptormene), Michele Simionato (@micheles)]
- Added command
oq plot 'gmf_scenario?imt=PGA&kind=rlz-0'
[Fatemeh Alishahiha (@FatemehAlsh)]
- Added vertical component to
abrahamson_silva_1997
gsim library
[Michele Simionato (@micheles)]
- Internal: removed ProbabilityCurve and rename ProbabilityMap->MapArray
- Optimized building the CompositeSourceModel for complex logic trees
- Optimized "reading rates" in postclassical
- Extended the
delta_loss
warning to scenario_risk calculations
and documented it
[Michele Simionato (@micheles), Marco Pagani (@mmpagani)]
- Fixed the multifault rupture indices and the MultiLine class so
that the New Zealand model can run
[Paolo Tormene (@ptormene), Michele Simionato (@micheles)]
- Added a new parameter
asce_version
for AELO calculations
[Michele Simionato (@micheles)]
- Saving memory and disk space in conditioned GMFs scenarios, improving
also the performance in many cases - Fixed the passing of parameters to the underlying GMPE in NRCan15SiteTerm
- Added another check for missing gsim in scenario calculations
- Using custom
hcurves
anduhs
exporters in AELO mode
[Fatemeh Alishahiha (@FatemehAlsh)]
- Added a new gsim library named
ghasemi_2009
developed for Iran.
[Marco Pagani (@mmpagani)]
- Added a new epistemic uncertainty for modelling relative uncertainty on
MMax
[Michele Simionato (@micheles)]
- Added a check to forbid
individual_rlzs=true
andcollect_rlzs=true
- Fixed a bug in
oq extract "ruptures?rup_id=XXX
- Raising an early error if the user forgets to specify a site_model_file
when needed (i.e. for parameters region and xvf)
[Marco Pagani (@mmpagani), Christopher Brooks (@CB-quakemodel)]
- Refactored Chiou and Youngs (2014) GMM to be adjustable using methodology
of the Boore et al (2022) paper
[Michele Simionato (@micheles)]
- Discarded the stations (if any) from the avg_gmf plot
- Added an early check for homeless consequences
- Improved the check for missing site parameters
- Internal: fixed the serialization of logic trees using
simpleFaultGeometryAbsolute, complexFaultGeometryAbsolute or
characteristicFaultGeometryAbsolute - Added a check when importing hazard_curves.csv files: the values must be
probabilities - Internal: functions must be shorter than 100 lines and have less than
16 arguments to enter in the engine codebase - Fixed scenarios with CanadaSHM6 GMPEs and ModifiableGMPE
- Fixed
avg_losses-stats
exporter in the case of a single realization - Managed duplicated multi fault sources in event based for the New
Zealand model - Forbidden characters .:; in the branchID
[Michal Kolaj (@MichalKolaj)]
- Fixed a couple bugs in the Canadians GMPEs (can_shm6_inslab) causing wrong
means
[Michele Simionato (@micheles), Paolo Tormene (@ptormene)]
- Extended the engine and the WebUI to accept a rupture_file
in the contexts of the Aristotle project - Extended the engine and the WebUI to run scenario_risk calculations
from USGS ruptures in the contexts of the Aristotle project
[Michele Simionato (@micheles)]
- Raised a clear error for GMF-conditioned calculations with too many sites
- Rounding lon, lat with numpy in oqparam.sites
- Reduced by half the data transfer in classical by using 32 bit rates
- Fixed a bug in
oq zip
which was missing the exposure.csv files - Introduced a parameter
config.memory.avg_losses_max
- Changed
import_gmfs_hdf5
to not use ExternalLinks - Added
oq show usgs_rupture:<usgs_id>
printing the rupture parameters
[Matteo Nastasi (@nastasi-oq)]
- Add helper to check consistency between debian/changelog and
CONTRIBUTORS.txt
[Michele Simionato (@micheles)]
- Added an optional flag
config.distribution.compress
to reduce the
data transfer by compressing pickles larger than 1 MB - Optimized postclassical both for regular and tiling calculations
- Saving disk space in classical calculations (~4x) by gzipping the rates
- Removed Python version checks in install.py