Source code for openquake.hmtk.seismicity.catalogue

# -*- coding: utf-8 -*-
# vim: tabstop=4 shiftwidth=4 softtabstop=4

#
# LICENSE
#
# Copyright (C) 2010-2021 GEM Foundation, G. Weatherill, M. Pagani,
# D. Monelli.
#
# The Hazard Modeller's Toolkit is free software: you can redistribute
# it and/or modify it under the terms of the GNU Affero General Public
# License as published by the Free Software Foundation, either version
# 3 of the License, or (at your option) any later version.
#
# You should have received a copy of the GNU Affero General Public License
# along with OpenQuake. If not, see <http://www.gnu.org/licenses/>
#
# DISCLAIMER
#
# The software Hazard Modeller's Toolkit (openquake.hmtk) provided herein
# is released as a prototype implementation on behalf of
# scientists and engineers working within the GEM Foundation (Global
# Earthquake Model).
#
# It is distributed for the purpose of open collaboration and in the
# hope that it will be useful to the scientific, engineering, disaster
# risk and software design communities.
#
# The software is NOT distributed as part of GEM’s OpenQuake suite
# (https://www.globalquakemodel.org/tools-products) and must be considered as a
# separate entity. The software provided herein is designed and implemented
# by scientific staff. It is not developed to the design standards, nor
# subject to same level of critical review by professional software
# developers, as GEM’s OpenQuake software suite.
#
# Feedback and contribution to the software is welcome, and can be
# directed to the hazard scientific staff of the GEM Model Facility
# (hazard@globalquakemodel.org).
#
# The Hazard Modeller's Toolkit (openquake.hmtk) is therefore distributed WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
# FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
# for more details.
#
# The GEM Foundation, and the authors of the software, assume no
# liability for use of the software.

"""
Prototype of a 'Catalogue' class
"""

import csv
import numpy as np
from openquake.hazardlib.pmf import PMF
from openquake.hazardlib.geo.mesh import Mesh
from openquake.hazardlib.geo.utils import spherical_to_cartesian
from openquake.hmtk.seismicity.utils import (decimal_time, bootstrap_histogram_1D,
                                             bootstrap_histogram_2D)


[docs]class Catalogue(object): """ General Catalogue Class """ FLOAT_ATTRIBUTE_LIST = [ 'second', 'timeError', 'longitude', 'latitude', 'SemiMajor90', 'SemiMinor90', 'ErrorStrike', 'depth', 'depthError', 'magnitude', 'sigmaMagnitude'] INT_ATTRIBUTE_LIST = ['year', 'month', 'day', 'hour', 'minute', 'flag'] STRING_ATTRIBUTE_LIST = ['eventID', 'Agency', 'magnitudeType', 'comment'] TOTAL_ATTRIBUTE_LIST = list( (set(FLOAT_ATTRIBUTE_LIST).union( set(INT_ATTRIBUTE_LIST))).union( set(STRING_ATTRIBUTE_LIST))) SORTED_ATTRIBUTE_LIST = [ 'eventID', 'Agency', 'year', 'month', 'day', 'hour', 'minute', 'second', 'timeError', 'longitude', 'latitude', 'SemiMajor90', 'SemiMinor90', 'ErrorStrike', 'depth', 'depthError', 'magnitude', 'sigmaMagnitude', 'magnitudeType'] def __init__(self): """ Initialise the catalogue dictionary """ self.data = {} self.end_year = None self.start_year = None self.processes = {'declustering': None, 'completeness': None, 'recurrence': None, 'Poisson Tests': None} for attribute in self.TOTAL_ATTRIBUTE_LIST: if attribute in self.FLOAT_ATTRIBUTE_LIST: self.data[attribute] = np.array([], dtype=float) elif attribute in self.INT_ATTRIBUTE_LIST: self.data[attribute] = np.array([], dtype=int) else: self.data[attribute] = [] self.number_earthquakes = 0
[docs] def get_number_events(self): return len(self.data['eventID'])
def __len__(self): return self.get_number_events() def __str__(self): """ Returns a shortened print of the catalogue """ neq = self.get_number_events() if not neq: return "<Catalogue Object>No events" elif neq > 20: # Too many events to print, show 1st 10 and last 10 row_set = ["<Catalogue Object>{:g} events".format(neq)] for i in range(10): row_set.append(self._get_row_str(i)) row_set.append("...") for i in range(-10, 0, 1): row_set.append(self._get_row_str(i)) else: # Show all events row_set = ["<Catalogue Object>{:g} events".format(neq)] for i in range(neq): row_set.append(self._get_row_str(i)) return "\n".join(row_set) def _get_row_str(self, i): """ Returns a string representation of the key information in a row """ row_data = ["{:s}".format(self.data['eventID'][i]), "{:g}".format(self.data['year'][i]), "{:g}".format(self.data['month'][i]), "{:g}".format(self.data['day'][i]), "{:g}".format(self.data['hour'][i]), "{:g}".format(self.data['minute'][i]), "{:.1f}".format(self.data['second'][i]), "{:.3f}".format(self.data['longitude'][i]), "{:.3f}".format(self.data['latitude'][i]), "{:.1f}".format(self.data['depth'][i]), "{:.1f}".format(self.data['magnitude'][i])] return " ".join(row_data) def __getitem__(self, key): """ If the key is provided as an int, return a data for that index, otherwise if it is a string then return the data column """ if isinstance(key, int): # Gets the row specied row =[] for attr in self.SORTED_ATTRIBUTE_LIST: if len(self.data[attr]): row.append(self.data[attr][key]) else: # For empty columns just append None row.append(None) return row elif isinstance(key, str): return self.data[key] else: raise ValueError("__getitem__ requires integer or string") def __iter__(self): """ Iteration yields for each event a list of data """ for i in range(len(self)): row =[] for key in self.SORTED_ATTRIBUTE_LIST: if len(self.data[key]): row.append(self.data[key][i]) else: # For empty columns just append None row.append(None) yield row
[docs] def add_event(self): raise NotImplementedError
[docs] def write_catalogue(self, output_file, key_list=SORTED_ATTRIBUTE_LIST): """ Writes the catalogue to file using HTMK format (CSV). :param output_file: Name of the output file :param key_list: Optional list of attribute keys to be exported """ with open(output_file, 'w') as of: writer = csv.DictWriter(of, fieldnames=key_list) writer.writeheader() for i in range(self.get_number_events()): row_dict = {} for key in key_list: if len(self.data[key]) > 0: data = self.data[key][i] if key in self.INT_ATTRIBUTE_LIST: if np.isnan(data): data = '' else: data = int(data) if key in self.FLOAT_ATTRIBUTE_LIST: if np.isnan(data): data = '' else: data = float(data) row_dict[key] = data writer.writerow(row_dict)
[docs] def load_to_array(self, keys): """ This loads the data contained in the catalogue into a numpy array. The method works only for float data :param keys: A list of keys to be uploaded into the array :type list: """ # Preallocate the numpy array data = np.empty((len(self.data[keys[0]]), len(keys))) for i in range(0, len(self.data[keys[0]])): for j, key in enumerate(keys): data[i, j] = self.data[key][i] return data
[docs] def load_from_array(self, keys, data_array): """ This loads the data contained in an array into the catalogue object :param keys: A list of keys explaining the content of the columns in the array :type list: """ if len(keys) != np.shape(data_array)[1]: raise ValueError('Key list does not match shape of array!') for i, key in enumerate(keys): if key in self.INT_ATTRIBUTE_LIST: self.data[key] = data_array[:, i].astype(int) else: self.data[key] = data_array[:, i] if key not in self.TOTAL_ATTRIBUTE_LIST: print('Key %s not a recognised catalogue attribute' % key) self.update_end_year()
[docs] @classmethod def make_from_dict(cls, data): cat = cls() cat.data = data cat.update_end_year() return cat
[docs] def update_end_year(self): """ NOTE: To be called only when the catalogue is loaded (not when it is modified by declustering or completeness-based filtering) """ self.end_year = np.max(self.data['year'])
[docs] def update_start_year(self): """ NOTE: To be called only when the catalogue is loaded (not when it is modified by declustering or completeness-based filtering) """ self.start_year = np.min(self.data['year'])
[docs] def catalogue_mt_filter(self, mt_table, flag=None): """ Filter the catalogue using a magnitude-time table. The table has two columns and n-rows. :param nump.ndarray mt_table: Magnitude time table with n-rows where column 1 is year and column 2 is magnitude """ if flag is None: # No flag defined, therefore all events are initially valid flag = np.ones(self.get_number_events(), dtype=bool) for comp_val in mt_table: id0 = np.logical_and(self.data['year'].astype(float) < comp_val[0], self.data['magnitude'] < comp_val[1]) print(id0) flag[id0] = False if not np.all(flag): self.purge_catalogue(flag)
[docs] def get_bounding_box(self): """ Returns the bounding box of the catalogue :returns: (West, East, South, North) """ return (np.min(self.data["longitude"]), np.max(self.data["longitude"]), np.min(self.data["latitude"]), np.max(self.data["latitude"]))
[docs] def get_observed_mmax_sigma(self, default=None): """ :returns: the sigma for the maximum observed magnitude """ if not isinstance(self.data['sigmaMagnitude'], np.ndarray): obsmaxsig = default else: obsmaxsig = self.data['sigmaMagnitude'][ np.argmax(self.data['magnitude'])] return obsmaxsig
[docs] def get_decimal_time(self): ''' Returns the time of the catalogue as a decimal ''' return decimal_time(self.data['year'], self.data['month'], self.data['day'], self.data['hour'], self.data['minute'], self.data['second'])
[docs] def hypocentres_as_mesh(self): ''' Render the hypocentres to a nhlib.geo.mesh.Mesh object ''' return Mesh(self.data['longitude'], self.data['latitude'], self.data['depth'])
[docs] def hypocentres_to_cartesian(self): ''' Render the hypocentres to a cartesian array ''' return spherical_to_cartesian(self.data['longitude'], self.data['latitude'], self.data['depth'])
[docs] def sort_catalogue_chronologically(self): ''' Sorts the catalogue into chronological order ''' dec_time = self.get_decimal_time() idx = np.argsort(dec_time) if np.all((idx[1:] - idx[:-1]) > 0.): # Catalogue was already in chronological order return self.select_catalogue_events(idx)
[docs] def purge_catalogue(self, flag_vector): ''' Purges present catalogue with invalid events defined by flag_vector :param numpy.ndarray flag_vector: Boolean vector showing if events are selected (True) or not (False) ''' id0 = np.where(flag_vector)[0] self.select_catalogue_events(id0) self.get_number_events()
[docs] def select_catalogue_events(self, id0): ''' Orders the events in the catalogue according to an indexing vector. :param np.ndarray id0: Pointer array indicating the locations of selected events ''' for key in self.data: if isinstance( self.data[key], np.ndarray) and len(self.data[key]) > 0: # Dictionary element is numpy array - use logical indexing self.data[key] = self.data[key][id0] elif isinstance( self.data[key], list) and len(self.data[key]) > 0: # Dictionary element is list self.data[key] = [self.data[key][iloc] for iloc in id0] else: continue
[docs] def get_depth_distribution(self, depth_bins, normalisation=False, bootstrap=None): ''' Gets the depth distribution of the earthquake catalogue to return a single histogram. Depths may be normalised. If uncertainties are found in the catalogue the distrbution may be bootstrap sampled :param numpy.ndarray depth_bins: getBin edges for the depths :param bool normalisation: Choose to normalise the results such that the total contributions sum to 1.0 (True) or not (False) :param int bootstrap: Number of bootstrap samples :returns: Histogram of depth values ''' if len(self.data['depth']) == 0: # If depth information is missing raise ValueError('Depths missing in catalogue') if len(self.data['depthError']) == 0: self.data['depthError'] = np.zeros(self.get_number_events(), dtype=float) return bootstrap_histogram_1D(self.data['depth'], depth_bins, self.data['depthError'], normalisation=normalisation, number_bootstraps=bootstrap, boundaries=(0., None))
[docs] def get_depth_pmf(self, depth_bins, default_depth=5.0, bootstrap=None): """ Returns the depth distribution of the catalogue as a probability mass function """ if len(self.data['depth']) == 0: # If depth information is missing return PMF([(1.0, default_depth)]) # Get the depth distribution depth_hist = self.get_depth_distribution(depth_bins, normalisation=True, bootstrap=bootstrap) # If the histogram does not sum to 1.0 then remove the difference # from the lowest bin depth_hist = np.around(depth_hist, 3) while depth_hist.sum() - 1.0: depth_hist[-1] -= depth_hist.sum() - 1.0 depth_hist = np.around(depth_hist, 3) pmf_list = [] for iloc, prob in enumerate(depth_hist): pmf_list.append((prob, (depth_bins[iloc] + depth_bins[iloc + 1]) / 2.0)) return PMF(pmf_list)
[docs] def get_magnitude_depth_distribution(self, magnitude_bins, depth_bins, normalisation=False, bootstrap=None): ''' Returns a 2-D magnitude-depth histogram for the catalogue :param numpy.ndarray magnitude_bins: Bin edges for the magnitudes :param numpy.ndarray depth_bins: Bin edges for the depths :param bool normalisation: Choose to normalise the results such that the total contributions sum to 1.0 (True) or not (False) :param int bootstrap: Number of bootstrap samples :returns: 2D histogram of events in magnitude-depth bins ''' if len(self.data['depth']) == 0: # If depth information is missing raise ValueError('Depths missing in catalogue') if len(self.data['depthError']) == 0: self.data['depthError'] = np.zeros(self.get_number_events(), dtype=float) if len(self.data['sigmaMagnitude']) == 0: self.data['sigmaMagnitude'] = np.zeros(self.get_number_events(), dtype=float) return bootstrap_histogram_2D(self.data['magnitude'], self.data['depth'], magnitude_bins, depth_bins, boundaries=[(0., None), (None, None)], xsigma=self.data['sigmaMagnitude'], ysigma=self.data['depthError'], normalisation=normalisation, number_bootstraps=bootstrap)
[docs] def get_magnitude_time_distribution(self, magnitude_bins, time_bins, normalisation=False, bootstrap=None): ''' Returns a 2-D histogram indicating the number of earthquakes in a set of time-magnitude bins. Time is in decimal years! :param numpy.ndarray magnitude_bins: Bin edges for the magnitudes :param numpy.ndarray time_bins: Bin edges for the times :param bool normalisation: Choose to normalise the results such that the total contributions sum to 1.0 (True) or not (False) :param int bootstrap: Number of bootstrap samples :returns: 2D histogram of events in magnitude-year bins ''' return bootstrap_histogram_2D( self.get_decimal_time(), self.data['magnitude'], time_bins, magnitude_bins, xsigma=np.zeros(self.get_number_events()), ysigma=self.data['sigmaMagnitude'], normalisation=normalisation, number_bootstraps=bootstrap)
[docs] def concatenate(self, catalogue): """ This method attaches one catalogue to the current one :parameter catalogue: An instance of :class:`htmk.seismicity.catalogue.Catalogue` """ atts = getattr(self, 'data') attn = getattr(catalogue, 'data') data = _merge_data(atts, attn) if data is not None: setattr(self, 'data', data) for attrib in vars(self): atts = getattr(self, attrib) attn = getattr(catalogue, attrib) if attrib == 'end_year': setattr(self, attrib, max(atts, attn)) elif attrib == 'start_year': setattr(self, attrib, min(atts, attn)) elif attrib == 'data': pass elif attrib == 'number_earthquakes': setattr(self, attrib, atts + attn) elif attrib == 'processes': if atts != attn: raise ValueError('The catalogues cannot be merged' + ' since the they have' + ' a different processing history') else: raise ValueError('unknown attribute: %s' % attrib) self.sort_catalogue_chronologically()
def _merge_data(dat1, dat2): """ Merge two data dictionaries containing catalogue data :parameter dictionary dat1: Catalogue data dictionary :parameter dictionary dat2: Catalogue data dictionary :returns: A catalogue data dictionary containing the information originally included in dat1 and dat2 """ cnt = 0 for key in dat1: flg1 = len(dat1[key]) > 0 flg2 = len(dat2[key]) > 0 if flg1 != flg2: cnt += 1 if cnt: raise Warning('Cannot merge catalogues with different' + ' attributes') return None else: for key in dat1: if isinstance(dat1[key], np.ndarray): dat1[key] = np.concatenate((dat1[key], dat2[key]), axis=0) elif isinstance(dat1[key], list): dat1[key] += dat2[key] else: raise ValueError('Unknown type') return dat1