DAMASK_EICMD/processing/pre/geom_pack.py

170 lines
5.5 KiB
Python
Executable File

#!/usr/bin/env python
# -*- coding: UTF-8 no BOM -*-
import os,sys,string,re,math,numpy
from optparse import OptionParser, OptionGroup, Option, SUPPRESS_HELP
# -----------------------------
class extendedOption(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)
# ----------------------- MAIN -------------------------------
identifiers = {
'resolution': ['a','b','c'],
'dimension': ['x','y','z'],
'origin': ['x','y','z'],
}
mappings = {
'resolution': lambda x: int(x),
'dimension': lambda x: float(x),
'origin': lambda x: float(x),
'homogenization': lambda x: int(x),
'maxmicrostructure': lambda x: int(x),
}
parser = OptionParser(option_class=extendedOption, usage='%prog options [file[s]]', description = """
compress geometry files with ranges "a to b" and/or multiples "n of x".
""" + string.replace('$Id$','\n','\\n')
)
(options, filenames) = parser.parse_args()
# ------------------------------------------ setup file handles ---------------------------------------
files = []
if filenames == []:
files.append({'name':'STDIN', 'input':sys.stdin, 'output':sys.stdout})
else:
for name in filenames:
if os.path.exists(name):
files.append({'name':name, 'input':open(name), 'output':open(name+'_tmp','w')})
# ------------------------------------------ loop over input files ---------------------------------------
for file in files:
if file['name'] != 'STDIN': print file['name']
# get labels by either read the first row, or - if keyword header is present - the last line of the header
firstline = file['input'].readline()
m = re.search('(\d+)\s*head', firstline.lower())
if m:
headerlines = int(m.group(1))
headers = [firstline]+[file['input'].readline() for i in range(headerlines)]
else:
headerlines = 1
headers = firstline
content = file['input'].readlines()
file['input'].close()
info = {'resolution': [0,0,0],
'dimension': [0.0,0.0,0.0],
'origin': [0.0,0.0,0.0],
'homogenization': 1,
'maxmicrostructure': 0,
}
new_header = []
for header in headers:
headitems = map(str.lower,header.split())
if headitems[0] in mappings.keys():
if headitems[0] in identifiers.keys():
for i in xrange(len(identifiers[headitems[0]])):
info[headitems[0]][i] = \
mappings[headitems[0]](headitems[headitems.index(identifiers[headitems[0]][i])+1])
else:
info[headitems[0]] = mappings[headitems[0]](headitems[1])
if info['resolution'] == [0,0,0]:
sys.stderr.write('no resolution info found.\n')
continue
if info['dimension'] == [0.0,0.0,0.0]:
sys.stderr.write('no dimension info found.\n')
continue
if file['name'] != 'STDIN':
print 'resolution: %s'%(' x '.join(map(str,info['resolution'])))
print 'dimension: %s'%(' x '.join(map(str,info['dimension'])))
print 'origin: %s'%(' : '.join(map(str,info['origin'])))
new_header.append("resolution\ta %i\tb %i\tc %i\n"%(
info['resolution'][0],
info['resolution'][1],
info['resolution'][2],))
new_header.append("dimension\tx %f\ty %f\tz %f\n"%(
info['dimension'][0],
info['dimension'][1],
info['dimension'][2]))
new_header.append("origin\tx %f\ty %f\tz %f\n"%(
info['origin'][0],
info['origin'][1],
info['origin'][2]))
new_header.append("homogenization\t%i\n"%info['homogenization'])
if info['maxmicrostructure'] > 0:
new_header.append("maxMicrostructure\t%i\n"%info['maxmicrostructure'])
# ------------------------------------------ assemble header ---------------------------------------
file['output'].write('%i\theader\n'%(len(new_header))+''.join(new_header))
# ------------------------------------------ pack input ---------------------------------------
type = ''
former = -1
start = -1
reps = 0
for line in content:
for current in map(int,line.split()):
if current == former+1 and start+reps == former+1:
type = 'to'
reps += 1
elif current == former and start == former:
type = 'of'
reps += 1
else:
output = {'': '',
'.': str(former)+'\n',
'to': '%i to %i\n'%(former-reps+1,former),
'of': '%i of %i\n'%(reps,former),
}[type]
file['output'].write(output)
type = '.'
start = current
reps = 1
former = current
# write out last item...
output = {'.': str(former),
'to': '%i to %i'%(former-reps+1,former),
'of': '%i of %i'%(reps,former),
}[type]
file['output'].write(output+'\n')
# ------------------------------------------ output finalization ---------------------------------------
if file['name'] != 'STDIN':
file['output'].close()
os.rename(file['name']+'_tmp',file['name'])