Source code for pyrocko.model.geometry

# https://pyrocko.org - GPLv3
#
# The Pyrocko Developers, 21st Century
# ---|P------/S----------~Lg----------

'''
Representation of geometrical objects built from vertices, faces, outlines,
etc.
'''

import numpy as num

from pyrocko.guts import Object, List
from pyrocko.guts_array import Array
from pyrocko.table import Table, LocationRecipe

from .event import Event

from logging import getLogger


logger = getLogger('model.geometry')


[docs]def reduce_array_dims(array): ''' Support function to reduce output array ndims from table ''' if array.shape[0] == 1 and array.ndim > 1: return num.squeeze(array, axis=0) else: return array
[docs]class Geometry(Object): ''' Spatial (and temporal) distributed properties of an event The Geometry object allows to store properties of an event in a spatial (and temporal) distributed manner. For a set of planes ("faces"), characterized by their corner points ("vertices"), properties are stored. Also information on the outline of the source are stored. ''' properties = Table.T( default=Table.D(), help='Properties that should be displayable on the surface of the' ' geometry. If 2d time dependency in column directions.', optional=True) vertices = Table.T( default=Table.D(), help='Vertices of the mesh of the geometry. ' 'Expected to be (lat,lon,north,east,depth) for each vertex.', optional=True) faces = Table.T( default=Table.D(), help='Face integer indexes to the respective vertices. ' 'Indexes belonging to one polygon have to be given row-wise.', optional=True) outlines = List.T( Table.T(), default=[], help='List of vertices of the mesh of the outlines of the geometry. ' 'Expected to be (lat,lon,north,east,depth) for each vertex' '(outline).', optional=True) event = Event.T(default=Event.D()) times = Array.T( shape=(None,), dtype='float64', help='1d vector of times [s] wrt. event time for which ' 'properties have value. Must have constant delta t', optional=True) def __init__(self, **kwargs): Object.__init__(self, **kwargs) self._ntimes = None @property def deltat(self): ''' Sampling rate of properties (time difference) [s] ''' if self.times.size > 2: return self.times[1] - self.times[0] else: return 1. @property def ntimes(self): if self._ntimes is None: if self.times is not None: self._ntimes = len(self.times) else: return 0 return self._ntimes def time2idx(self, time): if self.times.size > 2: idx = int(round((time - self.times.min() - ( self.deltat / 2.)) / self.deltat)) if idx < 0: return 0 elif idx > self.ntimes: return self.ntimes else: return idx else: return 0 def set_vertices(self, vertices): self.vertices.add_recipe(LocationRecipe()) self.vertices.add_col(( 'c5', '', ('ref_lat', 'ref_lon', 'north_shift', 'east_shift', 'depth')), vertices) def get_vertices(self, col='c5'): if self.vertices: return reduce_array_dims(self.vertices.get_col(col)) def set_faces(self, faces): ncorners = faces.shape[1] sub_headers = tuple(['f{}'.format(i) for i in range(ncorners)]) self.faces.add_col(('faces', '', sub_headers), faces) def get_faces(self, col='faces'): if self.faces: return reduce_array_dims(self.faces.get_col(col)) def no_faces(self): return self.faces.get_nrows() def setup(self, vertices, faces, outlines=None): self.set_vertices(vertices) self.set_faces(faces) if outlines is not None: self.set_outlines(outlines) def set_outlines(self, outlines): if outlines is not None: self.outlines = [] for outline in outlines: outl = Table() outl.add_recipe(LocationRecipe()) outl.add_col(( 'c5', '', ('ref_lat', 'ref_lon', 'north_shift', 'east_shift', 'depth')), outline) self.outlines.append(outl) def get_outlines(self): if self.outlines: return self.outlines else: logger.warning('No outlines set!') def add_property(self, name, values): if values.ndim == 1 or values.shape[1] == 1: self.properties.add_col(name, values.reshape(-1,)) elif (values.ndim == 2) and (self.times is not None): assert values.shape[1] == self.times.shape[0] self.properties.add_col(name, values) else: raise AttributeError( 'Please give either 1D array or the associated times first.') def get_property(self, name): return reduce_array_dims(self.properties.get_col(name)) def has_property(self, name): return self.properties.has_col(name)