DAMASK_EICMD/processing/pre/geom_vicinityOffset.py

118 lines
4.9 KiB
Python
Raw Normal View History

#!/usr/bin/env python
# -*- coding: UTF-8 no BOM -*-
import os,sys,string,math
import numpy as np
from scipy import ndimage
from optparse import OptionParser
import damask
scriptID = string.replace('$Id$','\n','\\n')
2014-11-18 21:01:39 +05:30
scriptName = os.path.splitext(scriptID.split()[1])[0]
2013-05-14 23:21:53 +05:30
#--------------------------------------------------------------------------------------------------
# MAIN
#--------------------------------------------------------------------------------------------------
parser = OptionParser(option_class=damask.extendableOption, usage='%prog options [file[s]]', description = """
Offset microstructure index for points which see a microstructure different from themselves within a given (cubic) vicinity,
i.e. within the region close to a grain/phase boundary.
""", version = scriptID)
parser.add_option('-v', '--vicinity',
dest = 'vicinity',
type = 'int', metavar = 'int',
help = 'voxel distance checked for presence of other microstructure [%default]')
parser.add_option('-m', '--microstructureoffset',
dest='offset',
type = 'int', metavar = 'int',
help = 'offset (positive or negative) for tagged microstructure indices. '+
'"0" selects maximum microstructure index [%default]')
parser.set_defaults(vicinity = 1,
offset = 0,
)
(options, filenames) = parser.parse_args()
# --- loop over input files -------------------------------------------------------------------------
if filenames == []: filenames = [None]
for name in filenames:
try:
table = damask.ASCIItable(name = name,
buffered = False, labeled = False)
except: continue
2015-09-24 22:22:58 +05:30
damask.util.report(scriptName,name)
# --- interpret header ----------------------------------------------------------------------------
table.head_read()
info,extra_header = table.head_getGeom()
2015-09-24 22:22:58 +05:30
damask.util.croak(['grid a b c: %s'%(' x '.join(map(str,info['grid']))),
'size x y z: %s'%(' x '.join(map(str,info['size']))),
'origin x y z: %s'%(' : '.join(map(str,info['origin']))),
'homogenization: %i'%info['homogenization'],
'microstructures: %i'%info['microstructures'],
])
errors = []
if np.any(info['grid'] < 1): errors.append('invalid grid a b c.')
if np.any(info['size'] <= 0.0): errors.append('invalid size x y z.')
if errors != []:
2015-09-24 22:22:58 +05:30
damask.util.croak(errors)
table.close(dismiss = True)
continue
# --- read data ------------------------------------------------------------------------------------
microstructure = table.microstructure_read(info['grid']).reshape(info['grid'],order='F') # read microstructure
# --- do work ------------------------------------------------------------------------------------
2013-05-14 23:21:53 +05:30
newInfo = {
'microstructures': 0,
}
if options.offset == 0: options.offset = microstructure.max()
microstructure = np.where(ndimage.filters.maximum_filter(microstructure,size=1+2*options.vicinity,mode='wrap') ==
ndimage.filters.minimum_filter(microstructure,size=1+2*options.vicinity,mode='wrap'),
microstructure, microstructure + options.offset)
2013-05-14 23:21:53 +05:30
newInfo['microstructures'] = microstructure.max()
# --- report ---------------------------------------------------------------------------------------
remarks = []
if ( newInfo['microstructures'] != info['microstructures']): remarks.append('--> microstructures: %i'%newInfo['microstructures'])
2015-09-24 22:22:58 +05:30
if remarks != []: damask.util.croak(remarks)
# --- write header ---------------------------------------------------------------------------------
table.labels_clear()
table.info_clear()
2015-09-24 22:22:58 +05:30
table.info_append([
2014-01-20 20:11:56 +05:30
scriptID + ' ' + ' '.join(sys.argv[1:]),
"grid\ta {grid[0]}\tb {grid[1]}\tc {grid[2]}".format(grid=info['grid']),
"size\tx {size[0]}\ty {size[1]}\tz {size[2]}".format(size=info['size']),
"origin\tx {origin[0]}\ty {origin[1]}\tz {origin[2]}".format(origin=info['origin']),
"homogenization\t{homog}".format(homog=info['homogenization']),
"microstructures\t{microstructures}".format(microstructures=newInfo['microstructures']),
2015-09-24 22:22:58 +05:30
extra_header
])
table.head_write()
2013-05-14 23:21:53 +05:30
# --- write microstructure information ------------------------------------------------------------
formatwidth = int(math.floor(math.log10(microstructure.max())+1))
table.data = microstructure.reshape((info['grid'][0],info['grid'][1]*info['grid'][2]),order='F').transpose()
table.data_writeArray('%%%ii'%(formatwidth),delimiter = ' ')
# --- output finalization --------------------------------------------------------------------------
table.close() # close ASCII table