Skip to content

OpenQuake Engine 2.9.0
Compare
Choose a tag to compare
@nastasi-oq nastasi-oq released this 26 Feb 14:47
· 141218 commits to master since this release
v2.9.0
86f1f94

[Michele Simionato (@micheles)]

  • Deprecated the NRML format for the GMFs

[Matteo Nastasi (@nastasi-oq)]

  • Debian package moved to Python 3.5

[Graeme Weatherill (@g-weatherill)]

  • Small bug fix for Derras et al (2014) GMPE when Rjb = 0.0

[Michele Simionato (@micheles)]

  • Improved the .rst reports for classical calculations with tiling
  • Reduced the data transfer in the event based risk calculator by
    reading the event IDs directly from the workers
  • Integrated the gmf_ebrisk calculator inside the event based calculator
  • Improved the weighting algorithm for the sources in the event based
    rupture calculator
  • Improved error message for source model files declared as nrml/0.5 when
    they actually are nrml/0.4
  • Optimized the classical_bcr calculator for the case of many realizations
  • Extended the exposure CSV importer to manage the retrofitted field

[Marco Pagani (@mmpagani), Changlong Li]

  • Adds the Yu et al. (2013) GMPEs

[Michele Simionato (@micheles)]

  • Fixed a bug in the hazard outputs: they were displayed in the WebUI even
    if they were missing
  • Implemented splitting of nonparametric sources

[Marco Pagani (@mmpagani)]

  • Fixed the 'get_closest_points' method for the
    openquake.hazardlib.geo.surface.gridded.GriddedSurface class

[Michele Simionato (@micheles)]

  • Now the source model paths are relative to the source model logic tree file
  • Fixed an international date line bug when using rtree for prefiltering
  • Deprecated nrml.parse, it is now called nrml.to_python
  • Improved the task distribution by splitting the AreaSources upfront
    and by improving the weighting algorithm

[Daniele Viganò (@daniviga)]

  • TMPDIR can be customized via the openquake.cfg file
  • Updated dependencies compatibility in setup.py

[Michele Simionato (@micheles)]

  • If the hazard is precomputed, setting the site_model_file,
    gsim_logic_tree_file or source_model_logic_tree_file gives a warning
  • Removed the obsolete API /extract/qgis- and added extract/hcurves,
    extract/hmaps, extract/uhs for use with the QGIS plugin
  • Removed the deprecated GeoJSON exporters
  • Fixed a bug with oq engine --run job.ini --exports npz
  • Fixed the ordering of the IMTs in hazardlib
  • oq engine --delete-calculation now aborts the calculation first
  • Added some documentation about how to access the datastore
  • Introduced a minimum_distance for the GSIMs
  • Fixed several small issues with the UCERF calculators; now ucerf_hazard
    can be used as a precalculator of gmf_ebrisk
  • Initial support for disaggregation by source
  • Added the ability to import large exposures as CSV (experimental)
  • Changed the source weights to be proportional to the number of GSIMs
    relevant for the tectonic region type, thus improving the task distribution

[Daniele Viganò (@daniviga)]

  • The RPM python3-oq-engine package replaced python-oq-engine
  • RPM packages moved to Python 3.5

[Michele Simionato (@micheles)]

  • Added the ability to dump a specific calculation
  • Changed the signature of the extract command to oq extract what calc_id,
    where what is a path info plus query string;

[Graeme Weatherill (@g-weatherill)]

  • Implements significant duration GMPEs of Bommer et al. (2009) and Afshari &
    Stewart (2016)
  • Adds significant duration IMT definitions to support IMTs

[Michele Simionato (@micheles)]

  • Run the DbServer as a detached process
  • Improved the test coverage for event based with GMF correlation
  • Optimized the event based risk calculator from ruptures: now the ruptures
    are instantiated in the workers and not in the controller if possible
  • Exported the parameter ses_per_logic_tree_path in the ruptures.csv file
  • Improved the display names for the risk outputs
  • Added a /v1/:calc_id/abort route to the engine server and Abort buttons
    to the WebUI
  • Fixed the seeds properly in the case of vulnerability functions with PMFs:
    now even if the ground motion fields are all equal, the risk numbers
    will be different since there is a different seed per each field
  • Stored a rupture loss table in event based risk calculations
  • Made sure that the number of effective ruptures is stored in csm_info
  • Fixed the HMTK tests to work with numpy from 1.11 to 1.14
  • Added a command oq shell to open an embedded (I)Python shell
  • Turned the 'realizations' output into a dynamic output

[Matteo Nastasi (@nastasi-oq)]

  • Split package from python-oq-engine to python-oq-engine,
    python-oq-engine-master and python-oq-engine-worker
  • Implemented an API /v1/on_same_fs to check filesystem accessibility
    between engine and a client application

[Michele Simionato (@micheles)]

  • Reduced the data transfer when computing the hazard curves in postprocessing
  • Removed the FilteredSiteCollection class

[Nick Ackerley (@nackerley)]

  • Some improvements to the plotting routines of the HMTK

[Michele Simionato (@micheles)]

  • Removed the ability to run oq engine --run job_h.ini,job_r.ini
  • Forbidden the site model in gmf_ebrisk calculations
  • When the option --hc is given the ruptures can now be read directly
    from the workers, thus saving some startup time
  • Optimized the storage of the ruptures: the site IDs are not saved anymore
  • Added a check for missing risk_investigation_time
  • Reduced the data transfer in the gmf_ebrisk calculator
  • Now the gmf_ebrisk calculator builds the aggregate loss curves too
  • Extended the gmf_ebrisk calculator to use the GMFs produced by an event
    based hazard calculation, both via CSV and via the --hc option
  • Fixed a performance bug in the computations of the aggregate loss curves
    by reading the full event loss table at once
  • Fixed oq zip to work with gmf_ebrisk calculations
  • Fixed a serious bug in the gmf_ebrisk calculator: the results were in most
    cases wrong and dependent on the number of spawned tasks
  • Now the master_seed controls the generation of the epsilons in all
    situations (before in event_based_risk without asset_correlation it was
    managed by random_seed)
  • Changed the management of the epsilons in the gmf_ebrisk calculator to
    be the same as in the event_based_risk calculator
  • Added a check on the input files: the listed paths must be relative paths
  • Fixed a bug when storing the disagg-bins in the case the lenghts of the
    arrays are not the same for all sites
  • In the case of a single tile the prefiltering was applied twice: fixed the
    problem and generally improved the filtering/weighting of the sources
  • Fixed the CSV exporter for disaggregation outputs when iml_disagg is set

[Graeme Weatherill (@g-weatherill)]

  • Fixed ASK14 GMPE behaviour to remove ValueError
  • Implements comprehensive suite of NGA East ground motion models for
    central and eastern United States
  • Minor modification of check_gsim functions to permit instantiated classes
    to be passed
Assets 2
Loading