DAMASK_EICMD/python/damask/result.py

1139 lines
42 KiB
Python
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

import multiprocessing
import re
import glob
import os
from functools import partial
import vtk
from vtk.util import numpy_support
import h5py
import numpy as np
from . import util
from . import version
from . import mechanics
from . import Rotation
from . import Orientation
from . import Environment
from . import grid_filters
class Result:
"""
Read and write to DADF5 files.
DADF5 (DAKMASK HDF5) files contain DAMASK results.
"""
def __init__(self,fname):
"""
Opens an existing DADF5 file.
Parameters
----------
fname : str
name of the DADF5 file to be openend.
"""
with h5py.File(fname,'r') as f:
try:
self.version_major = f.attrs['DADF5_version_major']
self.version_minor = f.attrs['DADF5_version_minor']
except KeyError:
self.version_major = f.attrs['DADF5-major']
self.version_minor = f.attrs['DADF5-minor']
if self.version_major != 0 or not 2 <= self.version_minor <= 6:
raise TypeError('Unsupported DADF5 version {}.{} '.format(self.version_major,
self.version_minor))
self.structured = 'grid' in f['geometry'].attrs.keys()
if self.structured:
self.grid = f['geometry'].attrs['grid']
self.size = f['geometry'].attrs['size']
self.origin = f['geometry'].attrs['origin'] if self.version_major == 0 and self.version_minor >= 5 else \
np.zeros(3)
r=re.compile('inc[0-9]+')
increments_unsorted = {int(i[3:]):i for i in f.keys() if r.match(i)}
self.increments = [increments_unsorted[i] for i in sorted(increments_unsorted)]
self.times = [round(f[i].attrs['time/s'],12) for i in self.increments]
self.Nmaterialpoints, self.Nconstituents = np.shape(f['mapping/cellResults/constituent'])
self.materialpoints = [m.decode() for m in np.unique(f['mapping/cellResults/materialpoint']['Name'])]
self.constituents = [c.decode() for c in np.unique(f['mapping/cellResults/constituent'] ['Name'])]
self.con_physics = []
for c in self.constituents:
self.con_physics += f['/'.join([self.increments[0],'constituent',c])].keys()
self.con_physics = list(set(self.con_physics)) # make unique
self.mat_physics = []
for m in self.materialpoints:
self.mat_physics += f['/'.join([self.increments[0],'materialpoint',m])].keys()
self.mat_physics = list(set(self.mat_physics)) # make unique
self.selection= {'increments': self.increments,
'constituents': self.constituents,
'materialpoints': self.materialpoints,
'con_physics': self.con_physics,
'mat_physics': self.mat_physics}
self.fname = fname
def _manage_selection(self,action,what,datasets):
"""
Manages the visibility of the groups.
Parameters
----------
action : str
select from 'set', 'add', and 'del'
what : str
attribute to change (must be from self.selection)
datasets : list of str or Boolean
name of datasets as list, supports ? and * wildcards.
True is equivalent to [*], False is equivalent to []
"""
# allow True/False and string arguments
if datasets is True:
datasets = ['*']
elif datasets is False:
datasets = []
choice = datasets if hasattr(datasets,'__iter__') and not isinstance(datasets,str) else \
[datasets]
if what == 'increments':
choice = [c if isinstance(c,str) and c.startswith('inc') else
'inc{}'.format(c) for c in choice]
elif what == 'times':
what = 'increments'
if choice == ['*']:
choice = self.increments
else:
iterator = map(float,choice)
choice = []
for c in iterator:
idx=np.searchsorted(self.times,c)
if np.isclose(c,self.times[idx]):
choice.append(self.increments[idx])
elif np.isclose(c,self.times[idx+1]):
choice.append(self.increments[idx+1])
valid = [e for e_ in [glob.fnmatch.filter(getattr(self,what),s) for s in choice] for e in e_]
existing = set(self.selection[what])
if action == 'set':
self.selection[what] = valid
elif action == 'add':
add=existing.union(valid)
add_sorted=sorted(add, key=lambda x: int("".join([i for i in x if i.isdigit()])))
self.selection[what] = add_sorted
elif action == 'del':
diff=existing.difference(valid)
diff_sorted=sorted(diff, key=lambda x: int("".join([i for i in x if i.isdigit()])))
self.selection[what] = diff_sorted
def incs_in_range(self,start,end):
selected = []
for i,inc in enumerate([int(i[3:]) for i in self.increments]):
s,e = map(lambda x: int(x[3:] if isinstance(x,str) and x.startswith('inc') else x), (start,end))
if s <= inc <= e:
selected.append(self.increments[i])
return selected
def times_in_range(self,start,end):
selected = []
for i,time in enumerate(self.times):
if start <= time <= end:
selected.append(self.increments[i])
return selected
def iter_selection(self,what):
"""
Iterate over selection items by setting each one selected.
Parameters
----------
what : str
attribute to change (must be from self.selection)
"""
datasets = self.selection[what]
last_datasets = datasets.copy()
for dataset in datasets:
if last_datasets != self.selection[what]:
self._manage_selection('set',what,datasets)
raise Exception
self._manage_selection('set',what,dataset)
last_datasets = self.selection[what]
yield dataset
self._manage_selection('set',what,datasets)
def pick(self,what,datasets):
"""
Set selection.
Parameters
----------
what : str
attribute to change (must be from self.selection)
datasets : list of str or Boolean
name of datasets as list, supports ? and * wildcards.
True is equivalent to [*], False is equivalent to []
"""
self._manage_selection('set',what,datasets)
def pick_more(self,what,datasets):
"""
Add to selection.
Parameters
----------
what : str
attribute to change (must be from self.selection)
datasets : list of str or Boolean
name of datasets as list, supports ? and * wildcards.
True is equivalent to [*], False is equivalent to []
"""
self._manage_selection('add',what,datasets)
def pick_less(self,what,datasets):
"""
Delete from selection.
Parameters
----------
what : str
attribute to change (must be from self.selection)
datasets : list of str or Boolean
name of datasets as list, supports ? and * wildcards.
True is equivalent to [*], False is equivalent to []
"""
self._manage_selection('del',what,datasets)
def groups_with_datasets(self,datasets):
"""
Return groups that contain all requested datasets.
Only groups within
- inc?????/constituent/*_*/*
- inc?????/materialpoint/*_*/*
- inc?????/geometry/*
are considered as they contain user-relevant data.
Single strings will be treated as list with one entry.
Wild card matching is allowed, but the number of arguments need to fit.
Parameters
----------
datasets : iterable or str or Boolean
Examples
--------
datasets = False matches no group
datasets = True matches all groups
datasets = ['F','P'] matches a group with ['F','P','sigma']
datasets = ['*','P'] matches a group with ['F','P']
datasets = ['*'] does not match a group with ['F','P','sigma']
datasets = ['*','*'] does not match a group with ['F','P','sigma']
datasets = ['*','*','*'] matches a group with ['F','P','sigma']
"""
if datasets is False: return []
sets = datasets if isinstance(datasets,bool) or (hasattr(datasets,'__iter__') and not isinstance(datasets,str)) else \
[datasets]
groups = []
with h5py.File(self.fname,'r') as f:
for i in self.iter_selection('increments'):
for o,p in zip(['constituents','materialpoints'],['con_physics','mat_physics']):
for oo in self.iter_selection(o):
for pp in self.iter_selection(p):
group = '/'.join([i,o[:-1],oo,pp]) # o[:-1]: plural/singular issue
if sets is True:
groups.append(group)
else:
match = [e for e_ in [glob.fnmatch.filter(f[group].keys(),s) for s in sets] for e in e_]
if len(set(match)) == len(sets) : groups.append(group)
return groups
def list_data(self):
"""Return information on all active datasets in the file."""
message = ''
with h5py.File(self.fname,'r') as f:
for i in self.iter_selection('increments'):
message+='\n{} ({}s)\n'.format(i,self.times[self.increments.index(i)])
for o,p in zip(['constituents','materialpoints'],['con_physics','mat_physics']):
for oo in self.iter_selection(o):
message+=' {}\n'.format(oo)
for pp in self.iter_selection(p):
message+=' {}\n'.format(pp)
group = '/'.join([i,o[:-1],oo,pp]) # o[:-1]: plural/singular issue
for d in f[group].keys():
try:
dataset = f['/'.join([group,d])]
message+=' {} / ({}): {}\n'.\
format(d,dataset.attrs['Unit'].decode(),dataset.attrs['Description'].decode())
except KeyError:
pass
return message
def get_dataset_location(self,label):
"""Return the location of all active datasets with given label."""
path = []
with h5py.File(self.fname,'r') as f:
for i in self.iter_selection('increments'):
k = '/'.join([i,'geometry',label])
try:
f[k]
path.append(k)
except KeyError as e:
pass
for o,p in zip(['constituents','materialpoints'],['con_physics','mat_physics']):
for oo in self.iter_selection(o):
for pp in self.iter_selection(p):
k = '/'.join([i,o[:-1],oo,pp,label])
try:
f[k]
path.append(k)
except KeyError as e:
pass
return path
def get_constituent_ID(self,c=0):
"""Pointwise constituent ID."""
with h5py.File(self.fname,'r') as f:
names = f['/mapping/cellResults/constituent']['Name'][:,c].astype('str')
return np.array([int(n.split('_')[0]) for n in names.tolist()],dtype=np.int32)
def get_crystal_structure(self): # ToDo: extension to multi constituents/phase
"""Info about the crystal structure."""
with h5py.File(self.fname,'r') as f:
return f[self.get_dataset_location('orientation')[0]].attrs['Lattice'].astype('str') # np.bytes_ to string
def read_dataset(self,path,c=0,plain=False):
"""
Dataset for all points/cells.
If more than one path is given, the dataset is composed of the individual contributions.
"""
with h5py.File(self.fname,'r') as f:
shape = (self.Nmaterialpoints,) + np.shape(f[path[0]])[1:]
if len(shape) == 1: shape = shape +(1,)
dataset = np.full(shape,np.nan,dtype=np.dtype(f[path[0]]))
for pa in path:
label = pa.split('/')[2]
if (pa.split('/')[1] == 'geometry'):
dataset = np.array(f[pa])
continue
p = np.where(f['mapping/cellResults/constituent'][:,c]['Name'] == str.encode(label))[0]
if len(p)>0:
u = (f['mapping/cellResults/constituent']['Position'][p,c])
a = np.array(f[pa])
if len(a.shape) == 1:
a=a.reshape([a.shape[0],1])
dataset[p,:] = a[u,:]
p = np.where(f['mapping/cellResults/materialpoint']['Name'] == str.encode(label))[0]
if len(p)>0:
u = (f['mapping/cellResults/materialpoint']['Position'][p.tolist()])
a = np.array(f[pa])
if len(a.shape) == 1:
a=a.reshape([a.shape[0],1])
dataset[p,:] = a[u,:]
if plain and dataset.dtype.names is not None:
return dataset.view(('float64',len(dataset.dtype.names)))
else:
return dataset
def cell_coordinates(self):
"""Return initial coordinates of the cell centers."""
if self.structured:
return grid_filters.cell_coord0(self.grid,self.size,self.origin)
else:
with h5py.File(self.fname,'r') as f:
return f['geometry/x_c'][()]
@staticmethod
def _add_absolute(x):
return {
'data': np.abs(x['data']),
'label': '|{}|'.format(x['label']),
'meta': {
'Unit': x['meta']['Unit'],
'Description': 'Absolute value of {} ({})'.format(x['label'],x['meta']['Description']),
'Creator': 'result.py:add_abs v{}'.format(version)
}
}
def add_absolute(self,x):
"""
Add absolute value.
Parameters
----------
x : str
Label of scalar, vector, or tensor dataset to take absolute value of.
"""
self._add_generic_pointwise(self._add_absolute,{'x':x})
@staticmethod
def _add_calculation(**kwargs):
formula = kwargs['formula']
for d in re.findall(r'#(.*?)#',formula):
formula = formula.replace('#{}#'.format(d),"kwargs['{}']['data']".format(d))
return {
'data': eval(formula),
'label': kwargs['label'],
'meta': {
'Unit': kwargs['unit'],
'Description': '{} (formula: {})'.format(kwargs['description'],kwargs['formula']),
'Creator': 'result.py:add_calculation v{}'.format(version)
}
}
def add_calculation(self,label,formula,unit='n/a',description=None,vectorized=True):
"""
Add result of a general formula.
Parameters
----------
label : str
Label of resulting dataset.
formula : str
Formula to calculate resulting dataset. Existing datasets are referenced by #TheirLabel#.
unit : str, optional
Physical unit of the result.
description : str, optional
Human-readable description of the result.
vectorized : bool, optional
Indicate whether the formula can be used in vectorized form. Defaults to True.
"""
if not vectorized:
raise NotImplementedError
dataset_mapping = {d:d for d in set(re.findall(r'#(.*?)#',formula))} # datasets used in the formula
args = {'formula':formula,'label':label,'unit':unit,'description':description}
self._add_generic_pointwise(self._add_calculation,dataset_mapping,args)
@staticmethod
def _add_Cauchy(P,F):
return {
'data': mechanics.Cauchy(P['data'],F['data']),
'label': 'sigma',
'meta': {
'Unit': P['meta']['Unit'],
'Description': 'Cauchy stress calculated from {} ({}) '.format(P['label'],
P['meta']['Description'])+\
'and {} ({})'.format(F['label'],F['meta']['Description']),
'Creator': 'result.py:add_Cauchy v{}'.format(version)
}
}
def add_Cauchy(self,P='P',F='F'):
"""
Add Cauchy stress calculated from first Piola-Kirchhoff stress and deformation gradient.
Parameters
----------
P : str, optional
Label of the dataset containing the first Piola-Kirchhoff stress. Defaults to P.
F : str, optional
Label of the dataset containing the deformation gradient. Defaults to F.
"""
self._add_generic_pointwise(self._add_Cauchy,{'P':P,'F':F})
@staticmethod
def _add_determinant(T):
return {
'data': np.linalg.det(T['data']),
'label': 'det({})'.format(T['label']),
'meta': {
'Unit': T['meta']['Unit'],
'Description': 'Determinant of tensor {} ({})'.format(T['label'],T['meta']['Description']),
'Creator': 'result.py:add_determinant v{}'.format(version)
}
}
def add_determinant(self,T):
"""
Add the determinant of a tensor.
Parameters
----------
T : str
Label of tensor dataset.
"""
self._add_generic_pointwise(self._add_determinant,{'T':T})
@staticmethod
def _add_deviator(T):
if not T['data'].shape[1:] == (3,3):
raise ValueError
return {
'data': mechanics.deviatoric_part(T['data']),
'label': 's_{}'.format(T['label']),
'meta': {
'Unit': T['meta']['Unit'],
'Description': 'Deviator of tensor {} ({})'.format(T['label'],T['meta']['Description']),
'Creator': 'result.py:add_deviator v{}'.format(version)
}
}
def add_deviator(self,T):
"""
Add the deviatoric part of a tensor.
Parameters
----------
T : str
Label of tensor dataset.
"""
self._add_generic_pointwise(self._add_deviator,{'T':T})
@staticmethod
def _add_eigenvalue(T_sym):
return {
'data': mechanics.eigenvalues(T_sym['data']),
'label': 'lambda({})'.format(T_sym['label']),
'meta' : {
'Unit': T_sym['meta']['Unit'],
'Description': 'Eigenvalues of {} ({})'.format(T_sym['label'],T_sym['meta']['Description']),
'Creator': 'result.py:add_eigenvalues v{}'.format(version)
}
}
def add_eigenvalues(self,T_sym):
"""
Add eigenvalues of symmetric tensor.
Parameters
----------
T_sym : str
Label of symmetric tensor dataset.
"""
self._add_generic_pointwise(self._add_eigenvalue,{'T_sym':T_sym})
@staticmethod
def _add_eigenvector(T_sym):
return {
'data': mechanics.eigenvectors(T_sym['data']),
'label': 'v({})'.format(T_sym['label']),
'meta' : {
'Unit': '1',
'Description': 'Eigenvectors of {} ({})'.format(T_sym['label'],T_sym['meta']['Description']),
'Creator': 'result.py:add_eigenvectors v{}'.format(version)
}
}
def add_eigenvectors(self,T_sym):
"""
Add eigenvectors of symmetric tensor.
Parameters
----------
T_sym : str
Label of symmetric tensor dataset.
"""
self._add_generic_pointwise(self._add_eigenvector,{'T_sym':T_sym})
@staticmethod
def _add_IPFcolor(q,l):
d = np.array(l)
d_unit = d/np.linalg.norm(d)
m = util.scale_to_coprime(d)
colors = np.empty((len(q['data']),3),np.uint8)
lattice = q['meta']['Lattice']
for i,q in enumerate(q['data']):
o = Orientation(np.array([q['w'],q['x'],q['y'],q['z']]),lattice).reduced()
colors[i] = np.uint8(o.IPFcolor(d_unit)*255)
return {
'data': colors,
'label': 'IPFcolor_[{} {} {}]'.format(*m),
'meta' : {
'Unit': 'RGB (8bit)',
'Lattice': lattice,
'Description': 'Inverse Pole Figure (IPF) colors for direction/plane [{} {} {})'.format(*m),
'Creator': 'result.py:add_IPFcolor v{}'.format(version)
}
}
def add_IPFcolor(self,q,l):
"""
Add RGB color tuple of inverse pole figure (IPF) color.
Parameters
----------
q : str
Label of the dataset containing the crystallographic orientation as quaternions.
l : numpy.array of shape (3)
Lab frame direction for inverse pole figure.
"""
self._add_generic_pointwise(self._add_IPFcolor,{'q':q},{'l':l})
@staticmethod
def _add_maximum_shear(T_sym):
return {
'data': mechanics.maximum_shear(T_sym['data']),
'label': 'max_shear({})'.format(T_sym['label']),
'meta': {
'Unit': T_sym['meta']['Unit'],
'Description': 'Maximum shear component of {} ({})'.format(T_sym['label'],T_sym['meta']['Description']),
'Creator': 'result.py:add_maximum_shear v{}'.format(version)
}
}
def add_maximum_shear(self,T_sym):
"""
Add maximum shear components of symmetric tensor.
Parameters
----------
T_sym : str
Label of symmetric tensor dataset.
"""
self._add_generic_pointwise(self._add_maximum_shear,{'T_sym':T_sym})
@staticmethod
def _add_Mises(T_sym):
t = 'strain' if T_sym['meta']['Unit'] == '1' else \
'stress'
return {
'data': mechanics.Mises_strain(T_sym['data']) if t=='strain' else mechanics.Mises_stress(T_sym['data']),
'label': '{}_vM'.format(T_sym['label']),
'meta': {
'Unit': T_sym['meta']['Unit'],
'Description': 'Mises equivalent {} of {} ({})'.format(t,T_sym['label'],T_sym['meta']['Description']),
'Creator': 'result.py:add_Mises v{}'.format(version)
}
}
def add_Mises(self,T_sym):
"""
Add the equivalent Mises stress or strain of a symmetric tensor.
Parameters
----------
T_sym : str
Label of symmetric tensorial stress or strain dataset.
"""
self._add_generic_pointwise(self._add_Mises,{'T_sym':T_sym})
@staticmethod
def _add_norm(x,ord):
o = ord
if len(x['data'].shape) == 2:
axis = 1
t = 'vector'
if o is None: o = 2
elif len(x['data'].shape) == 3:
axis = (1,2)
t = 'tensor'
if o is None: o = 'fro'
else:
raise ValueError
return {
'data': np.linalg.norm(x['data'],ord=o,axis=axis,keepdims=True),
'label': '|{}|_{}'.format(x['label'],o),
'meta': {
'Unit': x['meta']['Unit'],
'Description': '{}-norm of {} {} ({})'.format(o,t,x['label'],x['meta']['Description']),
'Creator': 'result.py:add_norm v{}'.format(version)
}
}
def add_norm(self,x,ord=None):
"""
Add the norm of vector or tensor.
Parameters
----------
x : str
Label of vector or tensor dataset.
ord : {non-zero int, inf, -inf, fro, nuc}, optional
Order of the norm. inf means NumPys inf object. For details refer to numpy.linalg.norm.
"""
self._add_generic_pointwise(self._add_norm,{'x':x},{'ord':ord})
@staticmethod
def _add_PK2(P,F):
return {
'data': mechanics.PK2(P['data'],F['data']),
'label': 'S',
'meta': {
'Unit': P['meta']['Unit'],
'Description': '2. Kirchhoff stress calculated from {} ({}) '.format(P['label'],
P['meta']['Description'])+\
'and {} ({})'.format(F['label'],F['meta']['Description']),
'Creator': 'result.py:add_PK2 v{}'.format(version)
}
}
def add_PK2(self,P='P',F='F'):
"""
Add second Piola-Kirchhoff calculated from first Piola-Kirchhoff stress and deformation gradient.
Parameters
----------
P : str, optional
Label first Piola-Kirchhoff stress dataset. Defaults to P.
F : str, optional
Label of deformation gradient dataset. Defaults to F.
"""
self._add_generic_pointwise(self._add_PK2,{'P':P,'F':F})
@staticmethod
def _add_pole(q,p,polar):
pole = np.array(p)
unit_pole = pole/np.linalg.norm(pole)
m = util.scale_to_coprime(pole)
coords = np.empty((len(q['data']),2))
for i,q in enumerate(q['data']):
o = Rotation(np.array([q['w'],q['x'],q['y'],q['z']]))
rotatedPole = o*unit_pole # rotate pole according to crystal orientation
(x,y) = rotatedPole[0:2]/(1.+abs(unit_pole[2])) # stereographic projection
coords[i] = [np.sqrt(x*x+y*y),np.arctan2(y,x)] if polar else [x,y]
return {
'data': coords,
'label': 'p^{}_[{} {} {})'.format(u'' if polar else 'xy',*m),
'meta' : {
'Unit': '1',
'Description': '{} coordinates of stereographic projection of pole (direction/plane) in crystal frame'\
.format('Polar' if polar else 'Cartesian'),
'Creator' : 'result.py:add_pole v{}'.format(version)
}
}
def add_pole(self,q,p,polar=False):
"""
Add coordinates of stereographic projection of given pole in crystal frame.
Parameters
----------
q : str
Label of the dataset containing the crystallographic orientation as quaternions.
p : numpy.array of shape (3)
Crystallographic direction or plane.
polar : bool, optional
Give pole in polar coordinates. Defaults to False.
"""
self._add_generic_pointwise(self._add_pole,{'q':q},{'p':p,'polar':polar})
@staticmethod
def _add_rotational_part(F):
if not F['data'].shape[1:] == (3,3):
raise ValueError
return {
'data': mechanics.rotational_part(F['data']),
'label': 'R({})'.format(F['label']),
'meta': {
'Unit': F['meta']['Unit'],
'Description': 'Rotational part of {} ({})'.format(F['label'],F['meta']['Description']),
'Creator': 'result.py:add_rotational_part v{}'.format(version)
}
}
def add_rotational_part(self,F):
"""
Add rotational part of a deformation gradient.
Parameters
----------
F : str, optional
Label of deformation gradient dataset.
"""
self._add_generic_pointwise(self._add_rotational_part,{'F':F})
@staticmethod
def _add_spherical(T):
if not T['data'].shape[1:] == (3,3):
raise ValueError
return {
'data': mechanics.spherical_part(T['data']),
'label': 'p_{}'.format(T['label']),
'meta': {
'Unit': T['meta']['Unit'],
'Description': 'Spherical component of tensor {} ({})'.format(T['label'],T['meta']['Description']),
'Creator': 'result.py:add_spherical v{}'.format(version)
}
}
def add_spherical(self,T):
"""
Add the spherical (hydrostatic) part of a tensor.
Parameters
----------
T : str
Label of tensor dataset.
"""
self._add_generic_pointwise(self._add_spherical,{'T':T})
@staticmethod
def _add_strain_tensor(F,t,m):
if not F['data'].shape[1:] == (3,3):
raise ValueError
return {
'data': mechanics.strain_tensor(F['data'],t,m),
'label': 'epsilon_{}^{}({})'.format(t,m,F['label']),
'meta': {
'Unit': F['meta']['Unit'],
'Description': 'Strain tensor of {} ({})'.format(F['label'],F['meta']['Description']),
'Creator': 'result.py:add_strain_tensor v{}'.format(version)
}
}
def add_strain_tensor(self,F='F',t='V',m=0.0):
"""
Add strain tensor of a deformation gradient.
For details refer to damask.mechanics.strain_tensor
Parameters
----------
F : str, optional
Label of deformation gradient dataset. Defaults to F.
t : {V, U}, optional
Type of the polar decomposition, V for left stretch tensor and U for right stretch tensor.
Defaults to V.
m : float, optional
Order of the strain calculation. Defaults to 0.0.
"""
self._add_generic_pointwise(self._add_strain_tensor,{'F':F},{'t':t,'m':m})
@staticmethod
def _add_stretch_tensor(F,t):
if not F['data'].shape[1:] == (3,3):
raise ValueError
return {
'data': mechanics.left_stretch(F['data']) if t == 'V' else mechanics.right_stretch(F['data']),
'label': '{}({})'.format(t,F['label']),
'meta': {
'Unit': F['meta']['Unit'],
'Description': '{} stretch tensor of {} ({})'.format('Left' if t == 'V' else 'Right',
F['label'],F['meta']['Description']),
'Creator': 'result.py:add_stretch_tensor v{}'.format(version)
}
}
def add_stretch_tensor(self,F='F',t='V'):
"""
Add stretch tensor of a deformation gradient.
Parameters
----------
F : str, optional
Label of deformation gradient dataset. Defaults to F.
t : {V, U}, optional
Type of the polar decomposition, V for left stretch tensor and U for right stretch tensor.
Defaults to V.
"""
self._add_generic_pointwise(self._add_stretch_tensor,{'F':F},{'t':t})
def _job(self,group,func,datasets,args,lock):
"""Execute job for _add_generic_pointwise."""
try:
datasets_in = {}
lock.acquire()
with h5py.File(self.fname,'r') as f:
for arg,label in datasets.items():
loc = f[group+'/'+label]
datasets_in[arg]={'data' :loc[()],
'label':label,
'meta': {k:v.decode() for k,v in loc.attrs.items()}}
lock.release()
r = func(**datasets_in,**args)
return [group,r]
except Exception as err:
print('Error during calculation: {}.'.format(err))
return None
def _add_generic_pointwise(self,func,datasets,args={}):
"""
General function to add pointwise data.
Parameters
----------
func : function
Callback function that calculates a new dataset from one or more datasets per HDF5 group.
datasets : dictionary
Details of the datasets to be used: label (in HDF5 file) and arg (argument to which the data is parsed in func).
args : dictionary, optional
Arguments parsed to func.
"""
pool = multiprocessing.Pool(int(Environment().options['DAMASK_NUM_THREADS']))
lock = multiprocessing.Manager().Lock()
groups = self.groups_with_datasets(datasets.values())
default_arg = partial(self._job,func=func,datasets=datasets,args=args,lock=lock)
for result in util.show_progress(pool.imap_unordered(default_arg,groups),len(groups)):
if not result:
continue
lock.acquire()
with h5py.File(self.fname, 'a') as f:
try:
dataset = f[result[0]].create_dataset(result[1]['label'],data=result[1]['data'])
for l,v in result[1]['meta'].items():
dataset.attrs[l]=v.encode()
except OSError as err:
print('Could not add dataset: {}.'.format(err))
lock.release()
pool.close()
pool.join()
def to_vtk(self,labels,mode='cell'):
"""
Export to vtk cell/point data.
Parameters
----------
labels : str or list of
Labels of the datasets to be exported.
mode : str, either 'cell' or 'point'
Export in cell format or point format.
Defaults to 'cell'.
"""
if mode.lower()=='cell':
if self.structured:
coordArray = [vtk.vtkDoubleArray(),vtk.vtkDoubleArray(),vtk.vtkDoubleArray()]
for dim in [0,1,2]:
for c in np.linspace(0,self.size[dim],1+self.grid[dim]):
coordArray[dim].InsertNextValue(c)
vtk_geom = vtk.vtkRectilinearGrid()
vtk_geom.SetDimensions(*(self.grid+1))
vtk_geom.SetXCoordinates(coordArray[0])
vtk_geom.SetYCoordinates(coordArray[1])
vtk_geom.SetZCoordinates(coordArray[2])
else:
nodes = vtk.vtkPoints()
with h5py.File(self.fname,'r') as f:
nodes.SetData(numpy_support.numpy_to_vtk(f['/geometry/x_n'][()],deep=True))
vtk_geom = vtk.vtkUnstructuredGrid()
vtk_geom.SetPoints(nodes)
vtk_geom.Allocate(f['/geometry/T_c'].shape[0])
if self.version_major == 0 and self.version_minor <= 5:
vtk_type = vtk.VTK_HEXAHEDRON
n_nodes = 8
else:
if f['/geometry/T_c'].attrs['VTK_TYPE'] == b'TRIANGLE':
vtk_type = vtk.VTK_TRIANGLE
n_nodes = 3
elif f['/geometry/T_c'].attrs['VTK_TYPE'] == b'QUAD':
vtk_type = vtk.VTK_QUAD
n_nodes = 4
elif f['/geometry/T_c'].attrs['VTK_TYPE'] == b'TETRA': # not tested
vtk_type = vtk.VTK_TETRA
n_nodes = 4
elif f['/geometry/T_c'].attrs['VTK_TYPE'] == b'HEXAHEDRON':
vtk_type = vtk.VTK_HEXAHEDRON
n_nodes = 8
for i in f['/geometry/T_c']:
vtk_geom.InsertNextCell(vtk_type,n_nodes,i-1)
elif mode.lower()=='point':
Points = vtk.vtkPoints()
Vertices = vtk.vtkCellArray()
for c in self.cell_coordinates():
pointID = Points.InsertNextPoint(c)
Vertices.InsertNextCell(1)
Vertices.InsertCellPoint(pointID)
vtk_geom = vtk.vtkPolyData()
vtk_geom.SetPoints(Points)
vtk_geom.SetVerts(Vertices)
vtk_geom.Modified()
N_digits = int(np.floor(np.log10(int(self.increments[-1][3:]))))+1
for i,inc in enumerate(self.iter_selection('increments')):
vtk_data = []
materialpoints_backup = self.selection['materialpoints'].copy()
self.pick('materialpoints',False)
for label in (labels if isinstance(labels,list) else [labels]):
for p in self.iter_selection('con_physics'):
if p != 'generic':
for c in self.iter_selection('constituents'):
x = self.get_dataset_location(label)
if len(x) == 0:
continue
array = self.read_dataset(x,0)
shape = [array.shape[0],np.product(array.shape[1:])]
vtk_data.append(numpy_support.numpy_to_vtk(num_array=array.reshape(shape),deep=True))
vtk_data[-1].SetName('1_'+x[0].split('/',1)[1]) #ToDo: hard coded 1!
vtk_geom.GetCellData().AddArray(vtk_data[-1])
else:
x = self.get_dataset_location(label)
if len(x) == 0:
continue
array = self.read_dataset(x,0)
shape = [array.shape[0],np.product(array.shape[1:])]
vtk_data.append(numpy_support.numpy_to_vtk(num_array=array.reshape(shape),deep=True))
ph_name = re.compile(r'(?<=(constituent\/))(.*?)(?=(generic))') # identify phase name
dset_name = '1_' + re.sub(ph_name,r'',x[0].split('/',1)[1]) # removing phase name
vtk_data[-1].SetName(dset_name)
vtk_geom.GetCellData().AddArray(vtk_data[-1])
self.pick('materialpoints',materialpoints_backup)
constituents_backup = self.selection['constituents'].copy()
self.pick('constituents',False)
for label in (labels if isinstance(labels,list) else [labels]):
for p in self.iter_selection('mat_physics'):
if p != 'generic':
for m in self.iter_selection('materialpoints'):
x = self.get_dataset_location(label)
if len(x) == 0:
continue
array = self.read_dataset(x,0)
shape = [array.shape[0],np.product(array.shape[1:])]
vtk_data.append(numpy_support.numpy_to_vtk(num_array=array.reshape(shape),deep=True))
vtk_data[-1].SetName('1_'+x[0].split('/',1)[1]) #ToDo: why 1_?
vtk_geom.GetCellData().AddArray(vtk_data[-1])
else:
x = self.get_dataset_location(label)
if len(x) == 0:
continue
array = self.read_dataset(x,0)
shape = [array.shape[0],np.product(array.shape[1:])]
vtk_data.append(numpy_support.numpy_to_vtk(num_array=array.reshape(shape),deep=True))
vtk_data[-1].SetName('1_'+x[0].split('/',1)[1])
vtk_geom.GetCellData().AddArray(vtk_data[-1])
self.pick('constituents',constituents_backup)
if mode.lower()=='cell':
writer = vtk.vtkXMLRectilinearGridWriter() if self.structured else \
vtk.vtkXMLUnstructuredGridWriter()
x = self.get_dataset_location('u_n')
vtk_data.append(numpy_support.numpy_to_vtk(num_array=self.read_dataset(x,0),deep=True))
vtk_data[-1].SetName('u')
vtk_geom.GetPointData().AddArray(vtk_data[-1])
elif mode.lower()=='point':
writer = vtk.vtkXMLPolyDataWriter()
file_out = '{}_inc{}.{}'.format(os.path.splitext(os.path.basename(self.fname))[0],
inc[3:].zfill(N_digits),
writer.GetDefaultFileExtension())
writer.SetCompressorTypeToZLib()
writer.SetDataModeToBinary()
writer.SetFileName(file_out)
writer.SetInputData(vtk_geom)
writer.Write()
###################################################################################################
# BEGIN DEPRECATED
iter_visible = iter_selection
def _time_to_inc(self,start,end):
selected = []
for i,time in enumerate(self.times):
if start <= time <= end:
selected.append(self.increments[i])
return selected
def set_by_time(self,start,end):
"""
Set active increments based on start and end time.
Parameters
----------
start : float
start time (included)
end : float
end time (included)
"""
self._manage_selection('set','increments',self._time_to_inc(start,end))
def set_by_increment(self,start,end):
"""
Set active time increments based on start and end increment.
Parameters
----------
start : int
start increment (included)
end : int
end increment (included)
"""
if self.version_minor >= 4:
self._manage_selection('set','increments',[ 'inc{}'.format(i) for i in range(start,end+1)])
else:
self._manage_selection('set','increments',['inc{:05d}'.format(i) for i in range(start,end+1)])