DAMASK_EICMD/processing/post/addDivergence

289 lines
14 KiB
Python
Executable File

#!/usr/bin/python
import os,re,sys,math,string,numpy,postprocessingMath
from optparse import OptionParser, Option
# -----------------------------
class extendableOption(Option):
# -----------------------------
# used for definition of new option parser action 'extend', which enables to take multiple option arguments
# taken from online tutorial http://docs.python.org/library/optparse.html
ACTIONS = Option.ACTIONS + ("extend",)
STORE_ACTIONS = Option.STORE_ACTIONS + ("extend",)
TYPED_ACTIONS = Option.TYPED_ACTIONS + ("extend",)
ALWAYS_TYPED_ACTIONS = Option.ALWAYS_TYPED_ACTIONS + ("extend",)
def take_action(self, action, dest, opt, value, values, parser):
if action == "extend":
lvalue = value.split(",")
values.ensure_value(dest, []).extend(lvalue)
else:
Option.take_action(self, action, dest, opt, value, values, parser)
def location(idx,res):
return ( idx % res[0], \
(idx // res[0]) % res[1], \
(idx // res[0] // res[1]) % res[2] )
def index(location,res):
return ( location[0] % res[0] + \
(location[1] % res[1]) * res[0] + \
(location[2] % res[2]) * res[0] * res[1] )
def prefixMultiply(what,len):
return {True: ['%i_%s'%(i+1,what) for i in range(len)],
False:[what]}[len>1]
# --------------------------------------------------------------------
# MAIN
# --------------------------------------------------------------------
FDcoefficients = [ \
[1.0/2.0, 0.0, 0.0, 0.0],\
[2.0/3.0,-1.0/12.0, 0.0, 0.0],\
[3.0/4.0,-3.0/20.0,1.0/ 60.0, 0.0],\
[4.0/5.0,-1.0/ 5.0,4.0/105.0,-1.0/280.0],\
]
parser = OptionParser(option_class=extendableOption, usage='%prog options [file[s]]', description = """
Add column(s) containing divergence of requested column(s).
Operates on periodic ordered three-dimensional data sets.
Deals with both vector- and tensor-valued fields.
""" + string.replace('$Id$','\n','\\n')
)
parser.add_option('--fdm', dest='accuracy', action='extend', type='string', \
help='degree of central difference accuracy')
parser.add_option('--fft', dest='fft', action='store_true', \
help='calculate divergence in Fourier space [%default]')
parser.add_option('-m','--memory', dest='memory', action='store_true', \
help='memory efficient calculation (not possible for FFT based divergency [%default]')
parser.add_option('-v','--vector', dest='vector', action='extend', type='string', \
help='heading of columns containing vector field values')
parser.add_option('-t','--tensor', dest='tensor', action='extend', type='string', \
help='heading of columns containing tensor field values')
parser.add_option('-d','--dimension', dest='dim', type='float', nargs=3, \
help='physical dimension of data set in x (fast) y z (slow) [%default]')
parser.add_option('-r','--resolution', dest='res', type='int', nargs=3, \
help='resolution of data set in x (fast) y z (slow)')
parser.add_option('-s','--skip', dest='skip', type='int', nargs=3, \
help='items skipped due to periodicity in x (fast) y z (slow)')
parser.set_defaults(accuracy = [])
parser.set_defaults(memory = False)
parser.set_defaults(fft = False)
parser.set_defaults(vector = [])
parser.set_defaults(tensor = [])
parser.set_defaults(dim = [])
parser.set_defaults(skip = [0,0,0])
accuracyChoices = [2,4,6,8]
(options,filenames) = parser.parse_args()
if len(options.vector) + len(options.tensor) == 0:
parser.error('no data column specified...')
if len(options.dim) < 3:
parser.error('improper dimension specification...')
if not options.res or len(options.res) < 3:
parser.error('improper resolution specification...')
for choice in options.accuracy:
if int(choice) not in accuracyChoices:
parser.error('accuracy must be chosen from %s...'%(', '.join(accuracyChoices)))
if options.fft: options.accuracy.append('fft')
resSkip = [lambda a,b: a+b for a,b in zip(options.res,options.skip)]
datainfo = { # list of requested labels per datatype
'vector': {'len':3,
'label':[]},
'tensor': {'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 = []
if filenames == []:
files.append({'name':'STDIN', 'handle':sys.stdin})
else:
for name in filenames:
if os.path.exists(name):
files.append({'name':name, 'handle':open(name)})
# ------------------------------------------ loop over input files ---------------------------------------
for file in files:
print file['name']
content = file['handle'].readlines()
file['handle'].close()
# get labels by either read the first row, or - if keyword header is present - the last line of the header
headerlines = 1
m = re.search('(\d+)\s*head', content[0].lower())
if m:
headerlines = int(m.group(1))
passOn = content[1:headerlines]
headers = content[headerlines].split()
data = content[headerlines+1:]
regexp = re.compile('1_\d+_')
for i,l in enumerate(headers):
if regexp.match(l):
headers[i] = l[2:]
active = {}
column = {}
values = {}
div_field ={}
head = []
for datatype,info in datainfo.items():
for label in info['label']:
key = {True :'1_%s',
False:'%s' }[info['len']>1]%label
if key not in headers:
print 'column %s not found...'%key
else:
if datatype not in active: active[datatype] = []
if datatype not in column: column[datatype] = {}
if datatype not in values: values[datatype] = {}
if datatype not in div_field: div_field[datatype] = {}
active[datatype].append(label)
column[datatype][label] = headers.index(key)
values[datatype][label] = numpy.array([0.0 for i in xrange(datainfo[datatype]['len']*\
options.res[0]*options.res[1]*options.res[2])]).\
reshape((options.res[0],options.res[1],options.res[2],\
3,datainfo[datatype]['len']//3))
for what in options.accuracy: # loop over all requested degrees of accuracy (plus potentially fft)
if not options.memory or what != 'fft': # FFT divergence excluded in memory saving mode
head += prefixMultiply('div%s(%s)'%(what,label),datainfo[datatype]['len']//3)
# ------------------------------------------ assemble header ---------------------------------------
output = '%i\theader'%(headerlines+1) + '\n' + \
''.join(passOn) + \
string.replace('$Id$','\n','\\n')+ '\t' + \
' '.join(sys.argv[1:]) + '\n' + \
'\t'.join(headers + head) + '\n' # build extended header
# ------------------------------------------ read value field ---------------------------------------
idx = 0
for line in data:
items = line.split()[:len(headers)]
if len(items) < len(headers): # skip too short lines (probably comments or invalid)
continue
locSkip = location(idx,resSkip)
if ( locSkip[0] < options.res[0]
and locSkip[1] < options.res[1]
and locSkip[2] < options.res[2] ): # only take values that are not periodic images
for datatype,labels in active.items():
for label in labels:
values[datatype][label][locSkip[0]][locSkip[1]][locSkip[2]]\
= numpy.reshape(items[column[datatype][label]:
column[datatype][label]+datainfo[datatype]['len']],(3,datainfo[datatype]['len']//3))
idx += 1
# ------------------------------------------ read file ---------------------------------------
if options.memory:
idx = 0
for line in data:
items = line.split()[:len(headers)]
if len(items) < len(headers):
continue
output += '\t'.join(items)
(x,y,z) = location(idx,options.res)
for datatype,labels in active.items():
for label in labels:
for accuracy in options.accuracy:
if accuracy == 'fft': continue
for k in range(datainfo[datatype]['len']/3): # formulas from Mikhail Itskov: Tensor Algebra and Tensor Analysis for Engineers, Springer 2009, p 52
theDiv = 0.0
for a in range(int(accuracy)//2):
theDiv += FDcoefficients[int(accuracy)//2-1][a] * \
( \
(values[datatype][label][location(index([x+1+a,y,z],options.res),options.res)[0]] \
[location(index([x+1+a,y,z],options.res),options.res)[1]] \
[location(index([x+1+a,y,z],options.res),options.res)[2]][k][0] - \
values[datatype][label][location(index([x-1-a,y,z],options.res),options.res)[0]] \
[location(index([x-1-a,y,z],options.res),options.res)[1]] \
[location(index([x-1-a,y,z],options.res),options.res)[2]][k][0]) * options.res[0] / options.dim[0] + \
(values[datatype][label][location(index([x,y+1+a,z],options.res),options.res)[0]] \
[location(index([x,y+1+a,z],options.res),options.res)[1]] \
[location(index([x,y+1+a,z],options.res),options.res)[2]][k][1] - \
values[datatype][label][location(index([x,y-1-a,z],options.res),options.res)[0]] \
[location(index([x,y-1-a,z],options.res),options.res)[1]] \
[location(index([x,y-1-a,z],options.res),options.res)[2]][k][1]) * options.res[1] / options.dim[1] + \
(values[datatype][label][location(index([x,y,z+1+a],options.res),options.res)[0]] \
[location(index([x,y,z+1+a],options.res),options.res)[1]] \
[location(index([x,y,z+1+a],options.res),options.res)[2]][k][2]- \
values[datatype][label][location(index([x,y,z-1-a],options.res),options.res)[0]] \
[location(index([x,y,z-1-a],options.res),options.res)[1]] \
[location(index([x,y,z-1-a],options.res),options.res)[2]][k][2]) * options.res[2] / options.dim[2] \
)
output += '\t%f'%theDiv
output += '\n'
idx += 1
else:
for datatype,labels in active.items():
for label in labels:
if label not in div_field[datatype]: div_field[datatype][label] = {}
for accuracy in options.accuracy:
div_field[datatype][label][accuracy] = numpy.array([0.0 for i in range((datainfo[datatype]['len'])//3*\
options.res[0]*options.res[1]*options.res[2])]).\
reshape((options.res[0],options.res[1],options.res[2],\
datainfo[datatype]['len']//3))
if accuracy == 'fft':
div_field[datatype][label][accuracy] = postprocessingMath.divergence_fft(options.res[0],options.res[1],options.res[2],datainfo[datatype]['len']//3,options.dim,values[datatype][label])
else:
div_field[datatype][label][accuracy] = postprocessingMath.divergence(options.res[0],options.res[1],options.res[2],datainfo[datatype]['len']//3,eval(accuracy)//2-1,options.dim,values[datatype][label])
idx = 0
for line in data:
items = line.split()[:len(headers)]
if len(items) < len(headers):
continue
output += '\t'.join(items)
for datatype,labels in active.items():
for label in labels:
for accuracy in options.accuracy:
for i in range(datainfo[datatype]['len']/3):
output += '\t%f'%div_field[datatype][label][accuracy][location(idx,options.res)[0]][location(idx,options.res)[1]][location(idx,options.res)[2]][i]
output += '\n'
idx += 1
# ------------------------------------------ output result ---------------------------------------
if file['name'] == 'STDIN':
print output
else:
file['handle'] = open(file['name']+'_tmp','w')
try:
file['handle'].write(output)
file['handle'].close()
os.rename(file['name']+'_tmp',file['name'])
except:
print 'error during writing',file['name']+'_tmp'