DAMASK_EICMD/processing/post/DADF5_postResults.py

97 lines
3.7 KiB
Python
Raw Normal View History

#!/usr/bin/env python3
import os
import argparse
import numpy as np
import damask
scriptName = os.path.splitext(os.path.basename(__file__))[0]
scriptID = ' '.join([scriptName,damask.version])
# --------------------------------------------------------------------
# MAIN
# --------------------------------------------------------------------
parser = argparse.ArgumentParser()
#ToDo: We need to decide on a way of handling arguments of variable lentght
#https://stackoverflow.com/questions/15459997/passing-integer-lists-to-python
#parser.add_argument('--version', action='version', version='%(prog)s {}'.format(scriptID))
parser.add_argument('filenames', nargs='+',
help='DADF5 files')
parser.add_argument('-d','--dir', dest='dir',default='postProc',metavar='string',
help='name of subdirectory relative to the location of the DADF5 file to hold output')
2019-05-16 03:57:06 +05:30
parser.add_argument('--mat', nargs='+',
help='labels for materialpoint',dest='mat')
2019-05-16 03:57:06 +05:30
parser.add_argument('--con', nargs='+',
help='labels for constituent',dest='con')
options = parser.parse_args()
2019-05-16 03:57:06 +05:30
if options.mat is None: options.mat=[]
if options.con is None: options.con=[]
# --- loop over input files ------------------------------------------------------------------------
for filename in options.filenames:
results = damask.DADF5(filename)
if not results.structured: continue
delta = results.size/results.grid*0.5
x, y, z = np.meshgrid(np.linspace(delta[2],results.size[2]-delta[2],results.grid[2]),
np.linspace(delta[1],results.size[1]-delta[1],results.grid[1]),
np.linspace(delta[0],results.size[0]-delta[0],results.grid[0]),
indexing = 'ij')
coords = np.concatenate((z[:,:,:,None],y[:,:,:,None],x[:,:,:,None]),axis = 3)
for i,inc in enumerate(results.iter_visible('increments')):
print('Output step {}/{}'.format(i+1,len(results.increments)))
header = '1 header\n'
data = np.array([int(inc[3:]) for j in range(np.product(results.grid))]).reshape([np.product(results.grid),1])
header+= 'inc'
coords = coords.reshape([np.product(results.grid),3])
data = np.concatenate((data,coords),1)
header+=' 1_pos 2_pos 3_pos'
2019-05-16 03:57:06 +05:30
for label in options.con:
for p in results.iter_visible('con_physics'):
2019-09-16 04:30:19 +05:30
for c in results.iter_visible('constituents'):
x = results.get_dataset_location(label)
if len(x) == 0:
continue
array = results.read_dataset(x,0)
2019-05-16 03:57:06 +05:30
d = int(np.product(np.shape(array)[1:]))
data = np.concatenate((data,np.reshape(array,[np.product(results.grid),d])),1)
2019-05-16 03:57:06 +05:30
2019-05-16 13:01:13 +05:30
if d>1:
header+= ''.join([' {}_{}'.format(j+1,label) for j in range(d)])
else:
header+=' '+label
2019-05-16 15:14:03 +05:30
for label in options.mat:
for p in results.iter_visible('mat_physics'):
2019-09-16 04:30:19 +05:30
for m in results.iter_visible('materialpoints'):
2019-05-16 03:57:06 +05:30
x = results.get_dataset_location(label)
if len(x) == 0:
continue
array = results.read_dataset(x,0)
d = int(np.product(np.shape(array)[1:]))
data = np.concatenate((data,np.reshape(array,[np.product(results.grid),d])),1)
2019-05-07 18:48:12 +05:30
2019-05-16 13:01:13 +05:30
if d>1:
header+= ''.join([' {}_{}'.format(j+1,label) for j in range(d)])
else:
header+=' '+label
2019-05-07 18:48:12 +05:30
dirname = os.path.abspath(os.path.join(os.path.dirname(filename),options.dir))
if not os.path.isdir(dirname):
os.mkdir(dirname,0o755)
file_out = '{}_{}.txt'.format(os.path.splitext(os.path.split(filename)[-1])[0],inc)
2019-05-07 18:48:12 +05:30
np.savetxt(os.path.join(dirname,file_out),data,header=header,comments='')