def _cell_quality (dataset, quality) :
if not dataset : raise RuntimeError, 'Need a dataset to compute _cell_quality'
# create a dataset with only our array but the same geometry/topology
ds = dataset.NewInstance()
ds.UnRegister(None)
ds.CopyStructure(dataset.VTKObject)
filter = vtk.vtkCellQuality()
filter.SetInputData(ds)
if "area" == quality : filter.SetQualityMeasureToArea()
elif "aspect" == quality : filter.SetQualityMeasureToAspectRatio()
elif "aspect_gamma" == quality : filter.SetQualityMeasureToAspectGamma()
elif "condition" == quality : filter.SetQualityMeasureToCondition()
elif "diagonal" == quality : filter.SetQualityMeasureToDiagonal()
elif "jacobian" == quality : filter.SetQualityMeasureToJacobian()
elif "max_angle" == quality : filter.SetQualityMeasureToMaxAngle()
elif "shear" == quality : filter.SetQualityMeasureToShear()
elif "skew" == quality : filter.SetQualityMeasureToSkew()
elif "min_angle" == quality : filter.SetQualityMeasureToMinAngle()
elif "volume" == quality : filter.SetQualityMeasureToVolume()
else : raise RuntimeError, 'Unknown cell quality ['+quality+'].'
filter.Update()
varray = filter.GetOutput().GetCellData().GetArray("CellQuality")
ans = dsa.vtkDataArrayToVTKArray(varray, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = dsa.ArrayAssociation.CELL
return ans
def _cell_quality(dataset, quality):
if not dataset:
raise RuntimeError, 'Need a dataset to compute _cell_quality'
# create a dataset with only our array but the same geometry/topology
ds = dataset.NewInstance()
ds.UnRegister(None)
ds.CopyStructure(dataset.VTKObject)
filter = vtk.vtkCellQuality()
filter.SetInputData(ds)
if "area" == quality: filter.SetQualityMeasureToArea()
elif "aspect" == quality: filter.SetQualityMeasureToAspectRatio()
elif "aspect_gamma" == quality: filter.SetQualityMeasureToAspectGamma()
elif "condition" == quality: filter.SetQualityMeasureToCondition()
elif "diagonal" == quality: filter.SetQualityMeasureToDiagonal()
elif "jacobian" == quality: filter.SetQualityMeasureToJacobian()
elif "max_angle" == quality: filter.SetQualityMeasureToMaxAngle()
elif "shear" == quality: filter.SetQualityMeasureToShear()
elif "skew" == quality: filter.SetQualityMeasureToSkew()
elif "min_angle" == quality: filter.SetQualityMeasureToMinAngle()
elif "volume" == quality: filter.SetQualityMeasureToVolume()
else: raise RuntimeError, 'Unknown cell quality [' + quality + '].'
filter.Update()
varray = filter.GetOutput().GetCellData().GetArray("CellQuality")
ans = dsa.vtkDataArrayToVTKArray(varray, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = dsa.ArrayAssociation.CELL
return ans
def vertex_normal (dataset) :
"Returns the vertex normal of each point in a dataset."
if not dataset : raise RuntimeError, 'Need a dataset to compute vertex_normal'
ds = dataset.NewInstance()
ds.UnRegister(None)
ds.CopyStructure(dataset.VTKObject)
filter = vtk.vtkPolyDataNormals()
filter.SetInputData(ds)
filter.ComputeCellNormalsOff()
filter.ComputePointNormalsOn()
filter.SetFeatureAngle(180)
filter.SplittingOff()
filter.ConsistencyOff()
filter.AutoOrientNormalsOff()
filter.FlipNormalsOff()
filter.NonManifoldTraversalOff()
filter.Update()
varray = filter.GetOutput().GetPointData().GetNormals()
ans = dsa.vtkDataArrayToVTKArray(varray, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = dsa.ArrayAssociation.POINT
return ans
def gradient(narray, dataset=None):
"Returns the gradient of an array of scalars/vectors."
if not dataset: dataset = narray.DataSet
if not dataset: raise RuntimeError, 'Need a dataset to compute gradient'
try:
ncomp = narray.shape[1]
except IndexError:
ncomp = 1
if ncomp != 1 and ncomp != 3:
raise RuntimeError, 'Gradient only works with scalars (1 component) and vectors (3 component)'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
if ncomp == 1 : attribute_type = 'scalars'
else : attribute_type = 'vectors'
res = _cell_derivatives(narray, dataset, attribute_type, cd)
if ncomp == 1 : retVal = res.GetVectors()
else : retVal = res.GetTensors()
try:
if narray.GetName() : retVal.SetName("gradient of " + narray.GetName())
else : retVal.SetName("gradient")
except AttributeError : retVal.SetName("gradient")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
def vertex_normal(dataset):
"Returns the vertex normal of each point in a dataset."
if not dataset:
raise RuntimeError, 'Need a dataset to compute vertex_normal'
ds = dataset.NewInstance()
ds.UnRegister(None)
ds.CopyStructure(dataset.VTKObject)
filter = vtk.vtkPolyDataNormals()
filter.SetInputData(ds)
filter.ComputeCellNormalsOff()
filter.ComputePointNormalsOn()
filter.SetFeatureAngle(180)
filter.SplittingOff()
filter.ConsistencyOff()
filter.AutoOrientNormalsOff()
filter.FlipNormalsOff()
filter.NonManifoldTraversalOff()
filter.Update()
varray = filter.GetOutput().GetPointData().GetNormals()
ans = dsa.vtkDataArrayToVTKArray(varray, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = dsa.ArrayAssociation.POINT
return ans
def gradient(narray, dataset=None):
"Returns the gradient of an array of scalars/vectors."
if not dataset: dataset = narray.DataSet
if not dataset: raise RuntimeError, 'Need a dataset to compute gradient'
try:
ncomp = narray.shape[1]
except IndexError:
ncomp = 1
if ncomp != 1 and ncomp != 3:
raise RuntimeError, 'Gradient only works with scalars (1 component) and vectors (3 component)'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
if ncomp == 1: attribute_type = 'scalars'
else: attribute_type = 'vectors'
res = _cell_derivatives(narray, dataset, attribute_type, cd)
if ncomp == 1: retVal = res.GetVectors()
else: retVal = res.GetTensors()
try:
if narray.GetName(): retVal.SetName("gradient of " + narray.GetName())
else: retVal.SetName("gradient")
except AttributeError:
retVal.SetName("gradient")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
def _matrix_math_filter(narray, operation):
if operation not in [
'Determinant', 'Inverse', 'Eigenvalue', 'Eigenvector'
]:
raise RuntimeError, 'Unknown quality measure ['+operation+']'+\
'Supported are [Determinant, Inverse, Eigenvalue, Eigenvector]'
if narray.ndim != 3:
raise RuntimeError, operation+' only works for an array of matrices(3D array).'\
'Input shape ' + narray.shape
elif narray.shape[1] != narray.shape[2]:
raise RuntimeError, operation+' requires an array of 2D square matrices.'\
'Input shape ' + narray.shape
# numpy_to_vtk converts only contiguous arrays
if not narray.flags.contiguous: narray = narray.copy()
# Reshape is necessary because numpy_support.numpy_to_vtk only works with 2D or
# less arrays.
nrows = narray.shape[0]
ncols = narray.shape[1] * narray.shape[2]
narray = narray.reshape(nrows, ncols)
ds = vtk.vtkImageData()
ds.SetDimensions(nrows, 1, 1)
varray = numpy_support.numpy_to_vtk(narray)
varray.SetName('tensors')
ds.GetPointData().SetTensors(varray)
filter = vtk.vtkMatrixMathFilter()
if operation == 'Determinant': filter.SetOperationToDeterminant()
elif operation == 'Inverse': filter.SetOperationToInverse()
elif operation == 'Eigenvalue': filter.SetOperationToEigenvalue()
elif operation == 'Eigenvector': filter.SetOperationToEigenvector()
filter.SetInputData(ds)
filter.Update()
varray = filter.GetOutput().GetPointData().GetArray(operation)
ans = dsa.vtkDataArrayToVTKArray(varray)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
ans.DataSet = narray.DataSet
return ans
def _matrix_math_filter (narray, operation) :
if operation not in ['Determinant', 'Inverse', 'Eigenvalue', 'Eigenvector'] :
raise RuntimeError, 'Unknown quality measure ['+operation+']'+\
'Supported are [Determinant, Inverse, Eigenvalue, Eigenvector]'
if narray.ndim != 3 :
raise RuntimeError, operation+' only works for an array of matrices(3D array).'\
'Input shape ' + narray.shape
elif narray.shape[1] != narray.shape[2] :
raise RuntimeError, operation+' requires an array of 2D square matrices.'\
'Input shape ' + narray.shape
# numpy_to_vtk converts only contiguous arrays
if not narray.flags.contiguous : narray = narray.copy()
# Reshape is necessary because numpy_support.numpy_to_vtk only works with 2D or
# less arrays.
nrows = narray.shape[0]
ncols = narray.shape[1] * narray.shape[2]
narray = narray.reshape(nrows, ncols)
ds = vtk.vtkImageData()
ds.SetDimensions(nrows, 1, 1)
varray = numpy_support.numpy_to_vtk(narray)
varray.SetName('tensors')
ds.GetPointData().SetTensors(varray)
filter = vtk.vtkMatrixMathFilter()
if operation == 'Determinant' : filter.SetOperationToDeterminant()
elif operation == 'Inverse' : filter.SetOperationToInverse()
elif operation == 'Eigenvalue' : filter.SetOperationToEigenvalue()
elif operation == 'Eigenvector' : filter.SetOperationToEigenvector()
filter.SetInputData(ds)
filter.Update()
varray = filter.GetOutput().GetPointData().GetArray(operation)
ans = dsa.vtkDataArrayToVTKArray(varray)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
ans.DataSet = narray.DataSet
return ans
def strain (narray, dataset=None) :
"Returns the strain of an array of 3D vectors."
if not dataset : dataset = narray.DataSet
if not dataset : raise RuntimeError, 'Need a dataset to compute strain'
if 2 != narray.ndim or 3 != narray.shape[1] :
raise RuntimeError, 'strain only works with an array of 3D vectors'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
cd.SetTensorModeToComputeStrain()
res = _cell_derivatives(narray, dataset, 'vectors', cd)
retVal = res.GetTensors()
retVal.SetName("strain")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
def curl (narray, dataset=None):
"Returns the curl of an array of 3D vectors."
if not dataset : dataset = narray.DataSet
if not dataset : raise RuntimeError, 'Need a dataset to compute curl.'
if narray.ndim != 2 or narray.shape[1] != 3 :
raise RuntimeError, 'Curl only works with an array of 3D vectors.'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
cd.SetVectorModeToComputeVorticity()
res = _cell_derivatives(narray, dataset, 'vectors', cd)
retVal = res.GetVectors()
retVal.SetName("vorticity")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
def strain(narray, dataset=None):
"Returns the strain of an array of 3D vectors."
if not dataset: dataset = narray.DataSet
if not dataset: raise RuntimeError, 'Need a dataset to compute strain'
if 2 != narray.ndim or 3 != narray.shape[1]:
raise RuntimeError, 'strain only works with an array of 3D vectors'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
cd.SetTensorModeToComputeStrain()
res = _cell_derivatives(narray, dataset, 'vectors', cd)
retVal = res.GetTensors()
retVal.SetName("strain")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
def curl(narray, dataset=None):
"Returns the curl of an array of 3D vectors."
if not dataset: dataset = narray.DataSet
if not dataset: raise RuntimeError, 'Need a dataset to compute curl.'
if narray.ndim != 2 or narray.shape[1] != 3:
raise RuntimeError, 'Curl only works with an array of 3D vectors.'\
'Input shape ' + narray.shape
cd = vtk.vtkCellDerivatives()
cd.SetVectorModeToComputeVorticity()
res = _cell_derivatives(narray, dataset, 'vectors', cd)
retVal = res.GetVectors()
retVal.SetName("vorticity")
ans = dsa.vtkDataArrayToVTKArray(retVal, dataset)
# The association information has been lost over the vtk filter
# we must reconstruct it otherwise lower pipeline will be broken.
ans.Association = narray.Association
return ans
