openquake.calculators package

base module

class openquake.calculators.base.BaseCalculator(oqparam, calc_id=None)[source]

Bases: object

Abstract base class for all calculators.

  • oqparam – OqParam object
  • monitor – monitor object
  • calc_id – numeric calculation ID
accept_precalc = []

Set the attributes nbytes


Defensive programming against users providing an incorrect pre-calculation ID (with --hazard-calculation-id).

Parameters:precalc_mode – calculation_mode of the previous calculation

Core routine running on the workers.


Execution phase. Usually will run in parallel the core function and return a dictionary with the results.


Export all the outputs in the datastore in the given export formats. Individual outputs are not exported if there are multiple realizations.

from_engine = False
is_stochastic = False
monitor(operation='', **kw)[source]
Returns:a new Monitor instance

Post-processing phase of the aggregated output. It must be overridden with the export code. It will return a dictionary of output files.


Initialization phase.

precalc = None
run(pre_execute=True, concurrent_tasks=None, close=True, **kw)[source]

Run the calculation and return the exported outputs.


Update the current calculation parameters and save engine_version

class openquake.calculators.base.HazardCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.BaseCalculator

Base class for hazard calculators based on source models

Returns:the number of stored events
Returns:the total number of sites
Returns:the number of realizations
block_splitter(sources, weight=operator.attrgetter('weight'), key=<function HazardCalculator.<lambda>>)[source]
  • sources – a list of sources
  • weight – a weight function (default .weight)
  • key – None or ‘src_group_id’

an iterator over blocks of sources


Overridden in event based


To be overridden to initialize the datasets needed by the calculation


Read the risk model and set the attribute .riskmodel. The riskmodel can be empty for hazard calculations. Save the loss ratios (if any) in the datastore.


For compatibility with the engine


Check if there is a previous calculation ID. If yes, read the inputs by retrieving the previous calculation; if not, read the inputs directly.


Read the exposure, the riskmodel and update the attributes .sitecol, .assetcol


Read risk data and sources if any

Returns:an RtreeFilter

Save the risk models in the datastore

Returns:a SourceFilter/UcerfFilter

Save info about the composite source model inside the csm_info dataset


Save (weight, num_sites, calc_time) inside the source_info dataset

exception openquake.calculators.base.InvalidCalculationID[source]

Bases: Exception

Raised when running a post-calculation on top of an incompatible pre-calculation

class openquake.calculators.base.RiskCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.HazardCalculator

Base class for all risk calculators. A risk calculator must set the attributes .riskmodel, .sitecol, .assetcol, .riskinputs in the pre_execute phase.

build_riskinputs(kind, eps=None, num_events=0)[source]
  • kind – kind of hazard getter, can be ‘poe’ or ‘gmf’
  • eps – a matrix of epsilons (or None)
  • num_events – how many events there are

a list of RiskInputs objects, sorted by IMT.

combine(acc, res)[source]

Parallelize on the riskinputs and returns a dictionary of results. Require a .core_task to be defined with signature (riskinputs, riskmodel, rlzs_assoc, monitor).

get_getter(kind, sid)[source]
  • kind – ‘poe’ or ‘gmf’
  • sid – a site ID

a PmapGetter or GmfDataGetter

read_shakemap(haz_sitecol, assetcol)[source]

Enabled only if there is a shakemap_id parameter in the job.ini. Download, unzip, parse USGS shakemap files and build a corresponding set of GMFs which are then filtered with the hazard site collection and stored in the datastore.

openquake.calculators.base.build_hmaps(hcurves_by_kind, slice_, imtls, poes, monitor)[source]

Build hazard maps from a slice of hazard curves. :returns: a pair ({kind: hmaps}, slice)

openquake.calculators.base.build_weights(realizations, imt_dt)[source]
Returns:an array with the realization weights of dtype imt_dt
openquake.calculators.base.check_time_event(oqparam, occupancy_periods)[source]

Check the time_event parameter in the datastore, by comparing with the periods found in the exposure.


Physically, an extremely small intensity measure level can have an extremely large probability of exceedence, however that probability cannot be exactly 1 unless the level is exactly 0. Numerically, the PoE can be 1 and this give issues when calculating the damage (there is a log(0) in openquake.risklib.scientific.annual_frequency_of_exceedence). Here we solve the issue by replacing the unphysical probabilities 1 with .9999999999999999 (the float64 closest to 1).

openquake.calculators.base.get_gmv_data(sids, gmfs, events)[source]

Convert an array of shape (N, E, M) into an array of type gmv_data_dt

openquake.calculators.base.get_idxs(data, eid2idx)[source]

Convert from event IDs to event indices.

  • data – an array with a field eid
  • eid2idx – a dictionary eid -> idx

the array of event indices

openquake.calculators.base.import_gmfs(dstore, fname, sids)[source]

Import in the datastore a ground motion field CSV file.

  • dstore – the datastore
  • fname – the CSV file
  • sids – the site IDs (complete)

event_ids, num_rlzs

openquake.calculators.base.only_filter(srcs, srcfilter, dummy, monitor)[source]

Filter the given sources. Yield a pair (filtered_sources, { 0}) if there are filtered sources.

openquake.calculators.base.parallel_split_filter(csm, srcfilter, split, monitor)[source]

Apply split_filter() in parallel to the composite source model.

Returns:a new openquake.commonlib.source.CompositeSourceModel
openquake.calculators.base.save_gmf_data(dstore, sitecol, gmfs, imts, events=())[source]
Parameters:calculator – a scenario_risk/damage or event_based_risk calculator
Returns:a pair (eids, R) where R is the number of realizations
openquake.calculators.base.set_array(longarray, shortarray)[source]
  • longarray – a numpy array of floats of length L >= l
  • shortarray – a numpy array of floats of length l

Fill longarray with the values of shortarray, starting from the left. If shortarry is shorter than longarray, then the remaining elements on the right are filled with numpy.nan values.

openquake.calculators.base.split_filter(srcs, srcfilter, seed, monitor)[source]

Split the given source and filter the subsources by distance and by magnitude. Perform sampling if a nontrivial sample_factor is passed. Yields a pair (split_sources, split_time) if split_sources is non-empty.

getters module

class openquake.calculators.getters.AssetGetter(dstore)[source]

Bases: object

An object which is able to read the assets on a given site.

get(site_id, tagnames)[source]
Parameters:site_id – the site of interest
Returns:assets, ass_by_aid
class openquake.calculators.getters.GmfDataGetter(dstore, sids, num_rlzs)[source]


A dictionary-like object {sid: dictionary by realization index}

Parameters:gsim – ignored
Returns:an dict rlzi -> datadict
class openquake.calculators.getters.GmfGetter(rupgetter, srcfilter, oqparam)[source]

Bases: object

An hazard getter with methods .get_gmfdata and .get_hazard returning ground motion values.

Returns:a dict with keys gmfdata, indices, hcurves

Compute the GMFs for the given realization and yields arrays of the dtype (sid, eid, imti, gmv), one for rupture

Returns:an array of the dtype (sid, eid, imti, gmv)
Parameters:data – if given, an iterator of records of dtype gmf_dt
Returns:sid -> records

Initialize the computers. Should be called on the workers

class openquake.calculators.getters.PmapGetter(dstore, rlzs_assoc=None, sids=None, poes=())[source]

Bases: object

Read hazard curves from the datastore for all realizations or for a specific realization.

  • dstore – a DataStore instance or file system path to it
  • sids – the subset of sites to consider (if None, all sites)
  • rlzs_assoc – a RlzsAssoc instance (if None, infers it)
get(rlzi, grp=None)[source]
  • rlzi – a realization index
  • grp – None (all groups) or a string of the form “grp-XX”

the hazard curves for the given realization

Parameters:gsim – ignored
Returns:an dict rlzi -> datadict
Parameters:imtls – intensity measure types and levels
Returns:an array of (R, N) hazard curves

Compute the mean curve as a ProbabilityMap

Parameters:grp – if not None must be a string of the form “grp-XX”; in that case returns the mean considering only the contribution for group XX
Returns:a list of R probability maps

Read the poes and set the .data attribute with the hazard curves


Extract probability maps from the datastore, possibly generating on the fly the ones corresponding to the individual realizations. Yields pairs (tag, pmap).

Parameters:kind – the kind of PoEs to extract; if not given, returns the realization if there is only one or the statistics otherwise.
Returns:dictionary “grp-XXX” -> ProbabilityMap instance
class openquake.calculators.getters.RuptureGetter(filename, rup_indices, grp_id, trt, samples, rlzs_by_gsim)[source]

Bases: object

Iterable over ruptures.

  • filename – path to an HDF5 file with a dataset names ruptures
  • rup_indices – a list of rupture indices of the same group
E2R(array, rlzi)[source]
  • array – an array of shape (E, …)
  • rlzi – an array of E realization indices

an aggregated array of shape (R, …)

Returns:a composite array with the associations eid->rlz
Returns:a dictionary with the parameters of the rupture
get_ruptures(srcfilter=<openquake.hazardlib.calc.filters.SourceFilter object>)[source]
Returns:a list of EBRuptures filtered by bounding box
set_weights(src_filter, num_taxonomies_by_site)[source]
Returns:the weights of the ruptures in the getter
Yields:RuptureGetters with weight <= maxweight
openquake.calculators.getters.gen_rupture_getters(dstore, slc=slice(None, None, None), concurrent_tasks=1, hdf5cache=None)[source]
openquake.calculators.getters.get_maxloss_rupture(dstore, loss_type)[source]
  • dstore – a DataStore instance
  • loss_type – a loss type string

EBRupture instance corresponding to the maximum loss for the given loss type

classical module

class openquake.calculators.classical.ClassicalCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.HazardCalculator

Classical PSHA calculator


Initial accumulator, a dict grp_id -> ProbabilityMap(L, G)

accept_precalc = ['psha']
agg_dicts(acc, dic)[source]

Aggregate dictionaries of hazard curves by updating the accumulator.

  • acc – accumulator dictionary
  • dic – dictionary with keys pmap, calc_times, eff_ruptures
core_task(group, src_filter, gsims, param, monitor=<Monitor [jenkins]>)

Compute the hazard curves for a set of sources belonging to the same tectonic region type for all the GSIMs associated to that TRT. The arguments are the same as in calc_hazard_curves(), except for gsims, which is a list of GSIM instances.

Returns:a dictionary {grp_id: pmap} with attributes .grp_ids, .calc_times, .eff_ruptures

Run in parallel core_task(sources, sitecol, monitor), by parallelizing on the sources according to their weight and tectonic region type.


Used in the case of large source model logic trees. :yields: (sources, sites, gsims) triples


Collect the hazard curves by realization and export them.

Parameters:pmap_by_grp_id – a dictionary grp_id -> hazard curves
save_hazard_stats(acc, pmap_by_kind)[source]

Works by side effect by saving statistical hcurves and hmaps on the datastore.

  • acc – ignored
  • pmap_by_kind – a dictionary of ProbabilityMaps

kind can be (‘hcurves’, ‘mean’), (‘hmaps’, ‘mean’), …

class openquake.calculators.classical.PreCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.classical.ClassicalCalculator

Calculator to filter the sources and compute the number of effective ruptures

openquake.calculators.classical.build_hazard_stats(pgetter, N, hstats, individual_curves, monitor)[source]
  • pgetter – an openquake.commonlib.getters.PmapGetter
  • N – the total number of sites
  • hstats – a list of pairs (statname, statfunc)
  • individual_curves – if True, also build the individual curves
  • monitor – instance of Monitor

a dictionary kind -> ProbabilityMap

The “kind” is a string of the form ‘rlz-XXX’ or ‘mean’ of ‘quantile-XXX’ used to specify the kind of output.

openquake.calculators.classical.get_extreme_poe(array, imtls)[source]
  • array – array of shape (L, G) with L=num_levels, G=num_gsims
  • imtls – DictArray imt -> levels

the maximum PoE corresponding to the maximum level for IMTs and GSIMs

Returns:a string with the source IDs of the given sources, stripping the extension after the colon, if any

classical_bcr module

class openquake.calculators.classical_bcr.ClassicalBCRCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.classical_risk.ClassicalRiskCalculator

Classical BCR Risk calculator

accept_precalc = ['classical']
core_task(riskinputs, riskmodel, param, monitor)

Compute and return the average losses for each asset.

openquake.calculators.classical_bcr.classical_bcr(riskinputs, riskmodel, param, monitor)[source]

Compute and return the average losses for each asset.


classical_damage module

class openquake.calculators.classical_damage.ClassicalDamageCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.classical_risk.ClassicalRiskCalculator

Scenario damage calculator

accept_precalc = ['classical']
core_task(riskinputs, riskmodel, param, monitor)

Core function for a classical damage computation.


a nested dictionary lt_idx, rlz_idx -> asset_idx -> <damage array>


Export the result in CSV format.

Parameters:result – a dictionary (l, r) -> asset_ordinal -> fractions per damage state
openquake.calculators.classical_damage.classical_damage(riskinputs, riskmodel, param, monitor)[source]

Core function for a classical damage computation.


a nested dictionary lt_idx, rlz_idx -> asset_idx -> <damage array>

classical_risk module

class openquake.calculators.classical_risk.ClassicalRiskCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.RiskCalculator

Classical Risk calculator

accept_precalc = ['classical']
core_task(riskinputs, riskmodel, param, monitor)

Compute and return the average losses for each asset.


Saving loss curves in the datastore.

Parameters:result – aggregated result of the task classical_risk

Associate the assets to the sites and build the riskinputs.

precalc = 'classical'
openquake.calculators.classical_risk.classical_risk(riskinputs, riskmodel, param, monitor)[source]

Compute and return the average losses for each asset.


disaggregation module

Disaggregation calculator core functionality

class openquake.calculators.disaggregation.DisaggregationCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.HazardCalculator

Classical PSHA disaggregation calculator

POE_TOO_BIG = "You are trying to disaggregate for poe=%s.\nHowever the source model #%d, '%s',\nproduces at most probabilities of %.7f for rlz=#%d, IMT=%s.\nThe disaggregation PoE is too big or your model is wrong,\nproducing too small PoEs."
accept_precalc = ['psha']
agg_result(acc, result)[source]

Collect the results coming from compute_disagg into self.results, a dictionary with key (sid, rlzi, poe, imt, trti) and values which are probability arrays.

  • acc – dictionary k -> dic accumulating the results
  • result – dictionary with the result coming from a task
  • dstore – a datastore
  • iml4 – 4D array of IMLs with shape (N, 1, M, P)
build_stats(results, hstats)[source]
  • results – dict key -> 6D disagg_matrix
  • hstats – (statname, statfunc) pairs

Raise an error if the given poes_disagg are too small compared to the hazard curves.


Performs the disaggregation


Run the disaggregation phase.

Parameters:curves – a list of hazard curves, one per site

The curves can be all None if iml_disagg is set in the job.ini

Returns:(num_sites, num_rlzs, num_poes, num_imts)

Get all the relevant hazard curves for the given site ordinal. Returns a dictionary rlz_id -> curve_by_imt.


Save all the results of the disaggregation. NB: the number of results to save is #sites * #rlzs * #disagg_poes * #IMTs.

Parameters:results – a dictionary (sid, rlzi, poe, imt) -> trti -> disagg matrix

Save disagg-bins

save_disagg_result(dskey, results)[source]

Save the computed PMFs in the datastore

  • dskey – dataset key; can be ‘disagg’ or ‘disagg-stats’
  • results – a dictionary sid, rlz, poe, imt -> 6D disagg_matrix

Aggregate probabilities withe the usual formula 1 - (1 - P1) … (1 - Pn)

openquake.calculators.disaggregation.compute_disagg(sitecol, sources, cmaker, iml4, trti, bin_edges, oqparam, monitor)[source]
  • sitecol – a instance
  • sources – list of hazardlib source objects
  • cmaker – a openquake.hazardlib.gsim.base.ContextMaker instance
  • iml4 – an array of intensities of shape (N, R, M, P)
  • trti (dict) – tectonic region type index
  • bin_egdes – a dictionary site_id -> edges
  • oqparam – the parameters in the job.ini file
  • monitor – monitor of the currently running job

a dictionary of probability arrays, with composite key (sid, rlzi, poe, imt, iml, trti).

event_based module

class openquake.calculators.event_based.EventBasedCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.HazardCalculator

Event based PSHA calculator generating the ground motion fields and the hazard curves from the ruptures, depending on the configuration parameters.


Initial accumulator, a dictionary (grp_id, gsim) -> curves

accept_precalc = ['event_based', 'event_based_risk', 'ucerf_hazard']
agg_dicts(acc, result)[source]
  • acc – accumulator dictionary
  • result – an AccumDict with events, ruptures, gmfs and hcurves

Prefilter the composite source model and store the source_info

build_ruptures(sources, srcfilter, param, monitor=<Monitor [jenkins]>)
  • sources – a sequence of (prefiltered) sources of the same group
  • srcfilter – SourceFilter instance used for bounding box post filtering
  • param – a dictionary of additional parameters including ses_per_logic_tree_path
  • monitor – monitor instance

dictionaries with keys rup_array, calc_times, eff_ruptures


Raise a ValueError if the number of sites is larger than 65,536 or the number of IMTs is larger than 256 or the number of ruptures is larger than 4,294,967,296. The limits are due to the numpy dtype used to store the GMFs (gmv_dt). They could be relaxed in the future.

core_task(rupgetter, srcfilter, param, monitor)

Compute GMFs and optionally hazard curves

Returns:a cached CompositionInfo object
Returns:a dict eid -> index in the events table
Returns:a list of RuptureGetters
is_stochastic = True
Parameters:rup_array – an array of ruptures with fields grp_id
Returns:a list of RuptureGetters
openquake.calculators.event_based.compute_gmfs(rupgetter, srcfilter, param, monitor)[source]

Compute GMFs and optionally hazard curves


Extract the mean hazard curves from the datastore, as a composite array of length nsites.


Save in the datastore a composite array with fields (grp_id, gsim_id, rlzs)

openquake.calculators.event_based.update_nbytes(dstore, key, array)[source]

event_based_risk module

class openquake.calculators.event_based_risk.EbrCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.RiskCalculator

Event based PSHA calculator generating the total losses by taxonomy

accept_precalc = ['event_based', 'event_based_risk', 'ucerf_hazard', 'ebrisk']
combine(dummy, res)[source]
  • dummy – unused parameter
  • res – a result dictionary
core_task(riskinputs, riskmodel, param, monitor)

a dictionary of numpy arrays of shape (L, R)

Returns:a callable (start, stop) producing a slice of epsilons
is_stochastic = True

Save risk data and build the aggregate loss curves


Build aggregate loss curves in process

precalc = 'event_based'

Save the event loss tables incrementally.

Parameters:dic – dictionary with agglosses, avglosses

Compute the total losses by rupture and losses by rlzi.

openquake.calculators.event_based_risk.event_based_risk(riskinputs, riskmodel, param, monitor)[source]

a dictionary of numpy arrays of shape (L, R)

reportwriter module

Utilities to build a report writer generating a .rst report for a calculation

class openquake.calculators.reportwriter.ReportWriter(dstore)[source]

Bases: object

A particularly smart view over the datastore

add(name, obj=None)[source]

Add the view named name to the report text


Build the report and return a restructed text string


Save the report

title = {'params': 'Parameters', 'inputs': 'Input files', 'csm_info': 'Composite source model', 'dupl_sources': 'Duplicated sources', 'required_params_per_trt': 'Required parameters per tectonic region type', 'ruptures_per_trt': 'Number of ruptures per tectonic region type', 'ruptures_events': 'Specific information for event based', 'rlzs_assoc': 'Realizations per (GRP, GSIM)', 'job_info': 'Data transfer', 'biggest_ebr_gmf': 'Maximum memory allocated for the GMFs', 'avglosses_data_transfer': 'Estimated data transfer for the avglosses', 'exposure_info': 'Exposure model', 'slow_sources': 'Slowest sources', 'task_hazard:0': 'Fastest task', 'task_hazard:-1': 'Slowest task', 'task_info': 'Information about the tasks', 'times_by_source_class': 'Computation times by source typology', 'performance': 'Slowest operations'}
openquake.calculators.reportwriter.build_report(job_ini, output_dir=None)[source]

Write a report.csv file with information about the calculation without running it

  • job_ini – full pathname of the job.ini file
  • output_dir – the directory where the report is written (default the input directory)

scenario module

class openquake.calculators.scenario.ScenarioCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.HazardCalculator

Scenario hazard calculator


Compute the GMFs and return a dictionary gsim -> array(N, E, I)

is_stochastic = True

Read the site collection and initialize GmfComputer and seeds

scenario_damage module

class openquake.calculators.scenario_damage.ScenarioDamageCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.RiskCalculator

Scenario damage calculator

accept_precalc = ['scenario']
core_task(riskinputs, riskmodel, param, monitor)

Core function for a damage computation.


a dictionary {‘d_asset’: [(l, r, a, mean-stddev), …],

’d_event’: damage array of shape R, L, E, D, ‘c_asset’: [(l, r, a, mean-stddev), …], ‘c_event’: damage array of shape R, L, E}

d_asset and d_tag are related to the damage distributions whereas c_asset and c_tag are the consequence distributions. If there is no consequence model c_asset is an empty list and c_tag is a zero-valued array.

is_stochastic = True

Compute stats for the aggregated distributions and save the results on the datastore.

precalc = 'scenario'
openquake.calculators.scenario_damage.dist_by_asset(data, multi_stat_dt, number)[source]
  • data – array of shape (N, R, L, 2, …)
  • multi_stat_dt – numpy dtype for statistical outputs
  • number – expected number of units per asset

array of shape (N, R) with records of type multi_stat_dt

openquake.calculators.scenario_damage.scenario_damage(riskinputs, riskmodel, param, monitor)[source]

Core function for a damage computation.


a dictionary {‘d_asset’: [(l, r, a, mean-stddev), …],

’d_event’: damage array of shape R, L, E, D, ‘c_asset’: [(l, r, a, mean-stddev), …], ‘c_event’: damage array of shape R, L, E}

d_asset and d_tag are related to the damage distributions whereas c_asset and c_tag are the consequence distributions. If there is no consequence model c_asset is an empty list and c_tag is a zero-valued array.

scenario_risk module

class openquake.calculators.scenario_risk.ScenarioRiskCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.base.RiskCalculator

Run a scenario risk calculation

accept_precalc = ['scenario']
core_task(riskinputs, riskmodel, param, monitor)

Core function for a scenario computation.


a dictionary { ‘agg’: array of shape (E, L, R, 2), ‘avg’: list of tuples (lt_idx, rlz_idx, asset_ordinal, statistics) } where E is the number of simulated events, L the number of loss types, R the number of realizations and statistics is an array of shape (n, R, 4), with n the number of assets in the current riskinput object

is_stochastic = True

Compute stats for the aggregated distributions and save the results on the datastore.


Compute the GMFs, build the epsilons, the riskinputs, and a dictionary with the unit of measure, used in the export phase.

precalc = 'scenario'
openquake.calculators.scenario_risk.scenario_risk(riskinputs, riskmodel, param, monitor)[source]

Core function for a scenario computation.


a dictionary { ‘agg’: array of shape (E, L, R, 2), ‘avg’: list of tuples (lt_idx, rlz_idx, asset_ordinal, statistics) } where E is the number of simulated events, L the number of loss types, R the number of realizations and statistics is an array of shape (n, R, 4), with n the number of assets in the current riskinput object

ucerf_event_classical module

class openquake.calculators.ucerf_classical.UcerfClassicalCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.classical.ClassicalCalculator

UCERF classical calculator.

accept_precalc = ['ucerf_psha']

Run in parallel core_task(sources, sitecol, monitor), by parallelizing on the sources according to their weight and tectonic region type.


ucerf_event_based module

class openquake.calculators.ucerf_event_based.UCERFHazardCalculator(oqparam, calc_id=None)[source]

Bases: openquake.calculators.event_based.EventBasedCalculator

Event based PSHA calculator generating the ruptures and GMFs together

accept_precalc = ['ucerf_hazard']
build_ruptures(sources, src_filter, param, monitor)
  • sources – a list with a single UCERF source
  • param – extra parameters
  • monitor – a Monitor instance

an AccumDict grp_id -> EBRuptures


parse the logic tree and source model input

openquake.calculators.ucerf_event_based.build_ruptures(sources, src_filter, param, monitor)[source]
  • sources – a list with a single UCERF source
  • param – extra parameters
  • monitor – a Monitor instance

an AccumDict grp_id -> EBRuptures

openquake.calculators.ucerf_event_based.generate_event_set(ucerf, background_sids, src_filter, ses_idx, seed)[source]

Generates the event set corresponding to a particular branch

openquake.calculators.ucerf_event_based.sample_background_model(hdf5, branch_key, tom, seed, filter_idx, min_mag, npd, hdd, upper_seismogenic_depth, lower_seismogenic_depth, msr=<WC1994>, aspect=1.5, trt='Active Shallow Crust')[source]

Generates a rupture set from a sample of the background model

  • branch_key – Key to indicate the branch for selecting the background model
  • tom – Temporal occurrence model as instance of :class: openquake.hazardlib.tom.TOM
  • seed – Random seed to use in the call to tom.sample_number_of_occurrences
  • filter_idx – Sites for consideration (can be None!)
  • min_mag (float) – Minimim magnitude for consideration of background sources
  • npd – Nodal plane distribution as instance of :class: openquake.hazardlib.pmf.PMF
  • hdd – Hypocentral depth distribution as instance of :class: openquake.hazardlib.pmf.PMF
  • aspect (float) – Aspect ratio
  • upper_seismogenic_depth (float) – Upper seismogenic depth (km)
  • lower_seismogenic_depth (float) – Lower seismogenic depth (km)
  • msr – Magnitude scaling relation
  • integration_distance (float) – Maximum distance from rupture to site for consideration

views module

class openquake.calculators.views.Source(source_id, code, geom, num_ruptures)

Bases: tuple


Alias for field number 1


Alias for field number 2


Alias for field number 3


Alias for field number 0

openquake.calculators.views.avglosses_data_transfer(token, dstore)[source]

Determine the amount of average losses transferred from the workers to the controller node in a risk calculation.

openquake.calculators.views.ebr_data_transfer(token, dstore)[source]

Display the data transferred in an event based risk calculation


Format numbers in a nice way.

>>> form(0)
>>> form(0.0)
>>> form(0.0001)
>>> form(1003.4)
>>> form(103.4)
>>> form(9.3)
>>> form(-1.2)

Returns the performance view as a numpy array.

openquake.calculators.views.rst_table(data, header=None, fmt=None)[source]

Build a .rst table from a matrix.

>>> tbl = [['a', 1], ['b', 2]]
>>> print(rst_table(tbl, header=['Name', 'Value']))
==== =====
Name Value
==== =====
a    1    
b    2    
==== =====
openquake.calculators.views.stats(name, array, *extras)[source]

Returns statistics from an array of numbers.

Parameters:name – a descriptive string
Returns:(name, mean, std, min, max, len)

Used to compute summaries. The records are assumed to have numeric fields, except the first field which is ignored, since it typically contains a label. Here is an example:

>>> sum_table([('a', 1), ('b', 2)])
['total', 3]
openquake.calculators.views.sum_tbl(tbl, kfield, vfields)[source]

Aggregate a composite array and compute the totals on a given key.

>>> dt = numpy.dtype([('name', (bytes, 10)), ('value', int)])
>>> tbl = numpy.array([('a', 1), ('a', 2), ('b', 3)], dt)
>>> sum_tbl(tbl, 'name', ['value'])['value']
array([3, 3])
openquake.calculators.views.view_act_ruptures_by_src(token, dstore)[source]

Display the actual number of ruptures by source in event based calculations

openquake.calculators.views.view_assets_by_site(token, dstore)[source]

Display statistical information about the distribution of the assets

openquake.calculators.views.view_contents(token, dstore)[source]

Returns the size of the contents of the datastore and its total size

openquake.calculators.views.view_csm_info(token, dstore)[source]
openquake.calculators.views.view_dupl_sources(token, dstore)[source]

Show the sources with the same ID and the truly duplicated sources

openquake.calculators.views.view_dupl_sources_time(token, dstore)[source]

Display the time spent computing duplicated sources

openquake.calculators.views.view_elt(token, dstore)[source]

Display the event loss table averaged by event

openquake.calculators.views.view_exposure_info(token, dstore)[source]

Display info about the exposure model

openquake.calculators.views.view_extreme_groups(token, dstore)[source]

Show the source groups contributing the most to the highest IML

openquake.calculators.views.view_fullreport(token, dstore)[source]

Display an .rst report about the computation

openquake.calculators.views.view_global_gmfs(token, dstore)[source]

Display GMFs averaged on everything for debugging purposes

openquake.calculators.views.view_global_hcurves(token, dstore)[source]

Display the global hazard curves for the calculation. They are used for debugging purposes when comparing the results of two calculations. They are the mean over the sites of the mean hazard curves.

openquake.calculators.views.view_global_hmaps(token, dstore)[source]

Display the global hazard maps for the calculation. They are used for debugging purposes when comparing the results of two calculations. They are the mean over the sites of the mean hazard maps.

openquake.calculators.views.view_global_poes(token, dstore)[source]

Display global probabilities averaged on all sites and all GMPEs

openquake.calculators.views.view_hmap(token, dstore)[source]

Display the highest 20 points of the mean hazard map. Called as $ oq show hmap:0.1 # 10% PoE

openquake.calculators.views.view_inputs(token, dstore)[source]
openquake.calculators.views.view_job_info(token, dstore)[source]

Determine the amount of data transferred from the controller node to the workers and back in a classical calculation.

openquake.calculators.views.view_mean_disagg(token, dstore)[source]

Display mean quantities for the disaggregation. Useful for checking differences between two calculations.

openquake.calculators.views.view_num_units(token, dstore)[source]

Display the number of units by taxonomy

openquake.calculators.views.view_params(token, dstore)[source]
openquake.calculators.views.view_performance(token, dstore)[source]

Display performance information

openquake.calculators.views.view_pmap(token, dstore)[source]

Display the mean ProbabilityMap associated to a given source group name

openquake.calculators.views.view_portfolio_loss(token, dstore)[source]

The mean and stddev loss for the full portfolio for each loss type, extracted from the event loss table, averaged over the realizations

openquake.calculators.views.view_portfolio_losses(token, dstore)[source]

The losses for the full portfolio, for each realization and loss type, extracted from the event loss table.

openquake.calculators.views.view_required_params_per_trt(token, dstore)[source]

Display the parameters needed by each tectonic region type

openquake.calculators.views.view_ruptures_events(token, dstore)[source]
openquake.calculators.views.view_ruptures_per_trt(token, dstore)[source]
openquake.calculators.views.view_short_source_info(token, dstore, maxrows=20)[source]
openquake.calculators.views.view_slow_sources(token, dstore, maxrows=20)[source]

Returns the slowest sources

openquake.calculators.views.view_task_durations(token, dstore)[source]

Display the raw task durations. Here is an example of usage:

$ oq show task_durations:classical
openquake.calculators.views.view_task_hazard(token, dstore)[source]

Display info about a given task. Here are a few examples of usage:

$ oq show task_hazard:0  # the fastest task
$ oq show task_hazard:-1  # the slowest task
openquake.calculators.views.view_task_info(token, dstore)[source]

Display statistical information about the tasks performance. It is possible to get full information about a specific task with a command like this one, for a classical calculation:

$ oq show task_info:classical
openquake.calculators.views.view_task_risk(token, dstore)[source]

Display info about a given risk task. Here are a few examples of usage:

$ oq show task_risk:0  # the fastest task
$ oq show task_risk:-1  # the slowest task
openquake.calculators.views.view_times_by_source_class(token, dstore)[source]

Returns the calculation times depending on the source typology

openquake.calculators.views.view_totlosses(token, dstore)[source]

This is a debugging view. You can use it to check that the total losses, i.e. the losses obtained by summing the average losses on all assets are indeed equal to the aggregate losses. This is a sanity check for the correctness of the implementation.

extract module

class openquake.calculators.extract.Extract[source]

Bases: dict

A callable dictionary of functions with a single instance called extract. Then extract(dstore, fullkey) dispatches to the function determined by the first part of fullkey (a slash-separated string) by passing as argument the second part of fullkey.

For instance extract(dstore, ‘sitecol’), extract(dstore, ‘asset_values/0’) etc.

add(key, cache=False)[source]
class openquake.calculators.extract.Extractor(calc_id)[source]

Bases: object

A class to extract data from a calculation.

Parameters:calc_id – a calculation ID

NB: instantiating the Extractor opens the datastore.


Close the datastore

Parameters:what – what to extract
Returns:an ArrayWrapper instance
exception openquake.calculators.extract.WebAPIError[source]

Bases: RuntimeError

Wrapper for an error on a WebAPI server

class openquake.calculators.extract.WebExtractor(calc_id, server=None, username=None, password=None)[source]

Bases: openquake.calculators.extract.Extractor

A class to extract data from the WebAPI.

  • calc_id – a calculation ID
  • server – hostname of the webapi server (can be ‘’)
  • username – login username (can be ‘’)
  • password – login password (can be ‘’)

NB: instantiating the WebExtractor opens a session.


Close the session


Dump the remote datastore on a local path.

Parameters:what – what to extract
Returns:an ArrayWrapper instance

Array of bytes

openquake.calculators.extract.build_damage_array(data, damage_dt)[source]
  • data – an array of length N with fields ‘mean’ and ‘stddev’
  • damage_dt – a damage composite data type loss_type -> states

a composite array of length N and dtype damage_dt

openquake.calculators.extract.build_damage_dt(dstore, mean_std=True)[source]
  • dstore – a datastore instance
  • mean_std – a flag (default True)

a composite dtype loss_type -> (mean_ds1, stdv_ds1, …) or loss_type -> (ds1, ds2, …) depending on the flag mean_std

openquake.calculators.extract.cast(loss_array, loss_dt)[source]
openquake.calculators.extract.crm_attrs(dstore, what)[source]
Returns:the attributes of the risk model, i.e. limit_states, loss_types, min_iml and covs, needed by the risk exporters.
openquake.calculators.extract.curves_by_tag(dstore, tag)[source]

Statistical loss curves by tag. For instance call

$ oq extract curves_by_tag/occupancy

openquake.calculators.extract.extract_(dstore, dspath)[source]

Extracts an HDF5 path object from the datastore, for instance extract(dstore, ‘sitecol’).

openquake.calculators.extract.extract_agg_curves(dstore, what)[source]

Aggregate loss curves of the given loss type and tags for event based risk calculations. Use it as /extract/agg_curves/structural?taxonomy=RC&zipcode=20126

Returns:array of shape (S, P), being P the number of return periods and S the number of statistics
openquake.calculators.extract.extract_agg_damages(dstore, what)[source]

Aggregate damages of the given loss type and tags. Use it as /extract/agg_damages/structural?taxonomy=RC&zipcode=20126

Returns:array of shape (R, D), being R the number of realizations and D the number of damage states or array of length 0 if there is no data for the given tags
openquake.calculators.extract.extract_agg_losses(dstore, what)[source]

Aggregate losses of the given loss type and tags. Use it as /extract/agg_losses/structural?taxonomy=RC&zipcode=20126 /extract/agg_losses/structural?taxonomy=RC&zipcode=*

Returns:an array of shape (T, R) if one of the tag names has a * value an array of shape (R,), being R the number of realizations an array of length 0 if there is no data for the given tags
openquake.calculators.extract.extract_aggregate_by(dstore, what)[source]

/extract/aggregate_by/taxonomy,occupancy/curves/structural yield pairs (<stat>, <array of shape (T, O, S, P)>)

/extract/aggregate_by/taxonomy,occupancy/avg_losses/structural yield pairs (<stat>, <array of shape (T, O, S)>)

openquake.calculators.extract.extract_asset_tags(dstore, tagname)[source]

Extract an array of asset tags for the given tagname. Use it as /extract/asset_tags or /extract/asset_tags/taxonomy

openquake.calculators.extract.extract_asset_values(dstore, sid)[source]

Extract an array of asset values for the given sid. Use it as /extract/asset_values/0

Returns:(aid, loss_type1, …, loss_typeN) composite array
openquake.calculators.extract.extract_dmg_by_asset_npz(dstore, what)[source]
openquake.calculators.extract.extract_event_info(dstore, eidx)[source]

Extract information about the given event index. Example:

openquake.calculators.extract.extract_gmf_scenario_npz(dstore, what)[source]
openquake.calculators.extract.extract_hcurves(dstore, what)[source]

Extracts hazard curves. Use it as /extract/hcurves?kind=mean or /extract/hcurves?kind=rlz-0, /extract/hcurves?kind=stats, /extract/hcurves?kind=rlzs etc

openquake.calculators.extract.extract_hmaps(dstore, what)[source]

Extracts hazard maps. Use it as /extract/hmaps?imt=PGA

openquake.calculators.extract.extract_losses_by_asset(dstore, what)[source]
openquake.calculators.extract.extract_losses_by_event(dstore, what)[source]
openquake.calculators.extract.extract_mean_std_curves(dstore, what)[source]

Yield imls/IMT and poes/IMT containg mean and stddev for all sites

openquake.calculators.extract.extract_mfd(dstore, what)[source]

Display num_ruptures by magnitude for event based calculations. Example:

openquake.calculators.extract.extract_realizations(dstore, dummy)[source]

Extract an array of realizations. Use it as /extract/realizations

openquake.calculators.extract.extract_rupture(dstore, serial)[source]

Extract information about the given event index. Example:

openquake.calculators.extract.extract_source_geom(dstore, srcidxs)[source]

Extract the geometry of a given sources Example:,2,3

openquake.calculators.extract.extract_src_loss_table(dstore, loss_type)[source]

Extract the source loss table for a give loss type, ordered in decreasing order. Example:

openquake.calculators.extract.extract_uhs(dstore, what)[source]

Extracts uniform hazard spectra. Use it as /extract/uhs?kind=mean or /extract/uhs?kind=rlz-0, etc

openquake.calculators.extract.get_mesh(sitecol, complete=True)[source]
Returns:a lon-lat or lon-lat-depth array depending if the site collection is at sea level or not
openquake.calculators.extract.get_ruptures_within(dstore, bbox)[source]

Extract the ruptures within the given bounding box, a string minlon,minlat,maxlon,maxlat. Example:,44,10,46

openquake.calculators.extract.hazard_items(dic, mesh, *extras, **kw)[source]
  • dic – dictionary of arrays of the same shape
  • mesh – a mesh array with lon, lat fields of the same length
  • extras – optional triples (field, dtype, values)
  • kw – dictionary of parameters (like investigation_time)

a list of pairs (key, value) suitable for storage in .npz format


ast.literal_eval the string if possible, otherwise returns it unchanged

openquake.calculators.extract.losses_by_tag(dstore, tag)[source]

Statistical average losses by tag. For instance call

$ oq extract losses_by_tag/occupancy

openquake.calculators.extract.parse(query_string, stats, num_rlzs=0)[source]
Returns:a normalized query_dict as in the following examples:
>>> parse('kind=stats', ['mean'])
{'kind': {'mean': 0}}
>>> parse('kind=rlzs', [], 3)
{'kind': {'rlz-000': 0, 'rlz-001': 1, 'rlz-002': 2}}
>>> parse('kind=mean', ['max', 'mean'])
{'kind': {'mean': 1}}
>>> parse('kind=rlz-3&imt=PGA&site_id=0', [])
{'kind': {'rlz-3': 3}, 'imt': ['PGA'], 'site_id': [0]}