198 lines
10 KiB
Python
Executable File
198 lines
10 KiB
Python
Executable File
#!/usr/bin/env python
|
|
# -*- coding: UTF-8 no BOM -*-
|
|
|
|
import os,sys,string,math,operator
|
|
import numpy as np
|
|
from collections import defaultdict
|
|
from optparse import OptionParser
|
|
import damask
|
|
|
|
scriptID = string.replace('$Id$','\n','\\n')
|
|
scriptName = os.path.splitext(scriptID.split()[1])[0]
|
|
|
|
#--------------------------------------------------------------------------------------------------
|
|
#> @brief calculates curl field using differentation in Fourier space
|
|
#> @todo enable odd resolution
|
|
#--------------------------------------------------------------------------------------------------
|
|
|
|
def curlFFT(geomdim,field):
|
|
grid = np.array(np.shape(field)[0:3])
|
|
wgt = 1.0/np.array(grid).prod()
|
|
|
|
if len(np.shape(field)) == 4:
|
|
dataType = 'vector'
|
|
elif len(np.shape(field)) == 5:
|
|
dataType = 'tensor'
|
|
|
|
field_fourier=np.fft.fftpack.rfftn(field,axes=(0,1,2))
|
|
curl_fourier=np.zeros(field_fourier.shape,'c8')
|
|
|
|
# differentiation in Fourier space
|
|
k_s=np.zeros([3],'i')
|
|
TWOPIIMG = (0.0+2.0j*math.pi)
|
|
for i in xrange(grid[0]):
|
|
k_s[0] = i
|
|
if(i > grid[0]/2 ): k_s[0] = k_s[0] - grid[0]
|
|
for j in xrange(grid[1]):
|
|
k_s[1] = j
|
|
if(j > grid[1]/2 ): k_s[1] = k_s[1] - grid[1]
|
|
for k in xrange(grid[2]/2+1):
|
|
k_s[2] = k
|
|
if(k > grid[2]/2 ): k_s[2] = k_s[2] - grid[2]
|
|
xi=np.array([k_s[2]/geomdim[2]+0.0j,k_s[1]/geomdim[1]+0.j,k_s[0]/geomdim[0]+0.j],'c8')
|
|
if dataType == 'tensor':
|
|
for l in xrange(3):
|
|
curl_fourier[i,j,k,0,l] = ( field_fourier[i,j,k,l,2]*xi[1]\
|
|
-field_fourier[i,j,k,l,1]*xi[2]) *TWOPIIMG
|
|
curl_fourier[i,j,k,1,l] = (-field_fourier[i,j,k,l,2]*xi[0]\
|
|
+field_fourier[i,j,k,l,0]*xi[2]) *TWOPIIMG
|
|
curl_fourier[i,j,k,2,l] = ( field_fourier[i,j,k,l,1]*xi[0]\
|
|
-field_fourier[i,j,k,l,0]*xi[1]) *TWOPIIMG
|
|
elif dataType == 'vector':
|
|
curl_fourier[i,j,k,0] = ( field_fourier[i,j,k,2]*xi[1]\
|
|
-field_fourier[i,j,k,1]*xi[2]) *TWOPIIMG
|
|
curl_fourier[i,j,k,1] = (-field_fourier[i,j,k,2]*xi[0]\
|
|
+field_fourier[i,j,k,0]*xi[2]) *TWOPIIMG
|
|
curl_fourier[i,j,k,2] = ( field_fourier[i,j,k,1]*xi[0]\
|
|
-field_fourier[i,j,k,0]*xi[1]) *TWOPIIMG
|
|
curl=np.fft.fftpack.irfftn(curl_fourier,axes=(0,1,2))
|
|
if dataType == 'tensor':
|
|
return curl.reshape([grid.prod(),9])
|
|
if dataType == 'vector':
|
|
return curl.reshape([grid.prod(),3])
|
|
|
|
|
|
# --------------------------------------------------------------------
|
|
# MAIN
|
|
# --------------------------------------------------------------------
|
|
|
|
parser = OptionParser(option_class=damask.extendableOption, usage='%prog options [file[s]]', description = """
|
|
Add column(s) containing curl of requested column(s).
|
|
Operates on periodic ordered three-dimensional data sets.
|
|
Deals with both vector- and tensor-valued fields.
|
|
|
|
""", version = scriptID)
|
|
|
|
parser.add_option('-c','--coordinates', dest='coords', metavar='string',
|
|
help='column heading for coordinates [%default]')
|
|
parser.add_option('-v','--vector', dest='vector', action='extend', metavar='<string LIST>',
|
|
help='heading of columns containing vector field values')
|
|
parser.add_option('-t','--tensor', dest='tensor', action='extend', metavar='<string LIST>',
|
|
help='heading of columns containing tensor field values')
|
|
parser.set_defaults(coords = 'ipinitialcoord')
|
|
parser.set_defaults(vector = [])
|
|
parser.set_defaults(tensor = [])
|
|
|
|
(options,filenames) = parser.parse_args()
|
|
|
|
if len(options.vector) + len(options.tensor) == 0:
|
|
parser.error('no data column specified...')
|
|
|
|
datainfo = { # list of requested labels per datatype
|
|
'vector': {'shape':[3],
|
|
'len':3,
|
|
'label':[]},
|
|
'tensor': {'shape':[3,3],
|
|
'len':9,
|
|
'label':[]},
|
|
}
|
|
|
|
if options.vector != None: datainfo['vector']['label'] = options.vector
|
|
if options.tensor != None: datainfo['tensor']['label'] = options.tensor
|
|
|
|
# ------------------------------------------ setup file handles ------------------------------------
|
|
files = []
|
|
for name in filenames:
|
|
if os.path.exists(name):
|
|
files.append({'name':name, 'input':open(name), 'output':open(name+'_tmp','w'), 'croak':sys.stderr})
|
|
|
|
#--- loop over input files -------------------------------------------------------------------------
|
|
for file in files:
|
|
file['croak'].write('\033[1m'+scriptName+'\033[0m: '+file['name']+'\n')
|
|
|
|
table = damask.ASCIItable(file['input'],file['output'],True) # make unbuffered ASCII_table
|
|
table.head_read() # read ASCII header info
|
|
table.info_append(scriptID + '\t' + ' '.join(sys.argv[1:]))
|
|
|
|
# --------------- figure out columns for coordinates and vector/tensor fields to process ---------
|
|
column = defaultdict(dict)
|
|
pos = 0 # when reading in the table via data_readArray, the first key is at colum 0
|
|
try:
|
|
column['coords'] = pos
|
|
pos+=3 # advance by data len (columns) for next key
|
|
keys=['%i_%s'%(i+1,options.coords) for i in xrange(3)] # store labels for column keys
|
|
except ValueError:
|
|
try:
|
|
column['coords'] = pos
|
|
pos+=3 # advance by data len (columns) for next key
|
|
directions = ['x','y','z']
|
|
keys=['%s.%s'%(options.coords,directions[i]) for i in xrange(3)] # store labels for column keys
|
|
except ValueError:
|
|
file['croak'].write('no coordinate data (1_%s) found...\n'%options.coords)
|
|
continue
|
|
|
|
active = defaultdict(list)
|
|
for datatype,info in datainfo.items():
|
|
for label in info['label']:
|
|
key = '1_%s'%label
|
|
if key not in table.labels:
|
|
file['croak'].write('column %s not found...\n'%key)
|
|
else:
|
|
active[datatype].append(label)
|
|
column[label] = pos
|
|
pos+=datainfo[datatype]['len']
|
|
keys+=['%i_%s'%(i+1,label) for i in xrange(datainfo[datatype]['len'])] # extend ASCII header with new labels
|
|
|
|
table.data_readArray(keys)
|
|
|
|
# --------------- assemble new header (columns containing curl) -----------------------------------
|
|
for datatype,labels in active.items(): # loop over vector,tensor
|
|
for label in labels:
|
|
table.labels_append(['%i_curlFFT(%s)'%(i+1,label) for i in xrange(datainfo[datatype]['len'])])# extend ASCII header with new labels
|
|
table.head_write()
|
|
|
|
# --------------- figure out size and grid ---------------------------------------------------------
|
|
coords = [{},{},{}]
|
|
for i in xrange(table.data.shape[0]):
|
|
for j in xrange(3):
|
|
coords[j][str(table.data[i,j])] = True # remember coordinate along x,y,z
|
|
grid = np.array([len(coords[0]),\
|
|
len(coords[1]),\
|
|
len(coords[2]),],'i') # grid is number of distinct coordinates found
|
|
size = grid/np.maximum(np.ones(3,'d'),grid-1.0)* \
|
|
np.array([max(map(float,coords[0].keys()))-min(map(float,coords[0].keys())),\
|
|
max(map(float,coords[1].keys()))-min(map(float,coords[1].keys())),\
|
|
max(map(float,coords[2].keys()))-min(map(float,coords[2].keys())),\
|
|
],'d') # size from bounding box, corrected for cell-centeredness
|
|
for i, points in enumerate(grid):
|
|
if points == 1:
|
|
mask = np.ones(3,dtype=bool)
|
|
mask[i]=0
|
|
size[i] = min(size[mask]/grid[mask]) # third spacing equal to smaller of other spacing
|
|
|
|
|
|
# ------------------------------------------ process value field -----------------------------------
|
|
curl = defaultdict(dict)
|
|
for datatype,labels in active.items(): # loop over vector,tensor
|
|
for label in labels: # loop over all requested curls
|
|
curl[datatype][label] = curlFFT(size[::-1], # we need to reverse order here, because x is fastest,ie rightmost, but leftmost in our x,y,z notation
|
|
table.data[:,column[label]:column[label]+datainfo[datatype]['len']].\
|
|
reshape([grid[2],grid[1],grid[0]]+datainfo[datatype]['shape']))
|
|
# ------------------------------------------ process data ------------------------------------------
|
|
table.data_rewind()
|
|
idx = 0
|
|
outputAlive = True
|
|
while outputAlive and table.data_read(): # read next data line of ASCII table
|
|
for datatype,labels in active.items(): # loop over vector,tensor
|
|
for label in labels: # loop over all requested norms
|
|
table.data_append(list(curl[datatype][label][idx,:]))
|
|
idx+=1
|
|
outputAlive = table.data_write() # output processed line
|
|
|
|
# ------------------------------------------ output result -----------------------------------------
|
|
outputAlive and table.output_flush() # just in case of buffered ASCII table
|
|
|
|
table.input_close() # close input ASCII table
|
|
table.output_close() # close output ASCII table
|
|
os.rename(file['name']+'_tmp',file['name']) # overwrite old one with tmp new
|