Section: Visualization Toolkit Graphics Classes
This class does not create vtkUnstructuredGrid output because it is intended for use in mappers as well as filters. Instead, it calls the registered function with simplices as they are created.
The subdivision algorithm should change the vertex coordinates (it must change both geometric and, if desired, parametric coordinates) of the midpoint. These coordinates need not be changed unless the EvaluateEdge() member returns true. The vtkStreamingTessellator itself has no way of creating a more accurate midpoint vertex.
Here's how to use this class: - Call AdaptivelySample1Facet, AdaptivelySample2Facet, or AdaptivelySample3Facet, with an edge, triangle, or tetrahedron you want tessellated. - The adaptive tessellator classifies each edge by passing the midpoint values to the vtkEdgeSubdivisionCriterion. - After each edge is classified, the tessellator subdivides edges as required until the subdivision criterion is satisfied or the maximum subdivision depth has been reached. - Edges, triangles, or tetrahedra connecting the vertices generated by the subdivision algorithm are processed by calling the user-defined callback functions (set with SetTetrahedronCallback(), SetTriangleCallback(), or SetEdgeCallback() ).
.SECTION Warning Note that the vertices passed to AdaptivelySample3Facet, AdaptivelySample2Facet, or AdaptivelySample1Facet must be at least 6, 5, or 4 entries long, respectively! This is because the <r,s,t>, <r,s>, or <r> parametric coordinates of the vertices are maintained as the facet is subdivided. This information is often required by the subdivision algorithm in order to compute an error metric. You may change the number of parametric coordinates associated with each vertex using vtkStreamingTessellator::SetEmbeddingDimension().
.SECTION Interpolating Field Values If you wish, you may also use \p vtkStreamingTessellator to interpolate field values at newly created vertices. Interpolated field values are stored just beyond the parametric coordinates associated with a vertex. They will always be \p double values; it does not make sense to interpolate a boolean or string value and your output and subdivision subroutines may always cast to a \p float or use \p floor() to truncate an interpolated value to an integer.
To create an instance of class vtkStreamingTessellator, simply invoke its constructor as follows
obj = vtkStreamingTessellator
obj
is an instance of the vtkStreamingTessellator class.
string = obj.GetClassName ()
int = obj.IsA (string name)
vtkStreamingTessellator = obj.NewInstance ()
vtkStreamingTessellator = obj.SafeDownCast (vtkObject o)
obj.SetSubdivisionAlgorithm (vtkEdgeSubdivisionCriterion )
- Get/Set the algorithm used to determine whether an edge should be
subdivided or left as-is. This is used once for each call to
AdaptivelySample1Facet (which is recursive and will call itself
resulting in additional edges to be checked) or three times for
each call to AdaptivelySample2Facet (also recursive).
vtkEdgeSubdivisionCriterion = obj.GetSubdivisionAlgorithm ()
- Get/Set the algorithm used to determine whether an edge should be
subdivided or left as-is. This is used once for each call to
AdaptivelySample1Facet (which is recursive and will call itself
resulting in additional edges to be checked) or three times for
each call to AdaptivelySample2Facet (also recursive).
obj.SetEmbeddingDimension (int k, int d)
- Get/Set the number of parameter-space coordinates associated with each input and output point.
The default is \a k for \a k -facets. You may
specify a different dimension, \a d, for each type of \a k -facet to be processed.
For example, \p SetEmbeddingDimension( \p 2, \p 3 ) would associate \a r, \a s, and
\a t coordinates with each input and output point generated by \p AdaptivelySample2Facet
but does not say anything about input or output points generated by
\p AdaptivelySample1Facet.
Call \p SetEmbeddingDimension( \p -1, \a d ) to specify the same dimension for
all possible \a k values.
\a d may not exceed 8, as that would be plain silly.
int = obj.GetEmbeddingDimension (int k) const
- Get/Set the number of parameter-space coordinates associated with each input and output point.
The default is \a k for \a k -facets. You may
specify a different dimension, \a d, for each type of \a k -facet to be processed.
For example, \p SetEmbeddingDimension( \p 2, \p 3 ) would associate \a r, \a s, and
\a t coordinates with each input and output point generated by \p AdaptivelySample2Facet
but does not say anything about input or output points generated by
\p AdaptivelySample1Facet.
Call \p SetEmbeddingDimension( \p -1, \a d ) to specify the same dimension for
all possible \a k values.
\a d may not exceed 8, as that would be plain silly.
obj.SetFieldSize (int k, int s)
- Get/Set the number of field value coordinates associated with each input and output point.
The default is 0; no field values are interpolated.
You may specify a different size, \a s, for each type of \a k -facet to be processed.
For example, \p SetFieldSize( \p 2, \p 3 ) would associate 3 field value coordinates
with each input and output point of an \p AdaptivelySample2Facet call,
but does not say anything about input or output points of \p AdaptivelySample1Facet.
Call \p SetFieldSize( \p -1, \a s ) to specify the same dimension for all possible \a k values.
\a s may not exceed vtkStreamingTessellator::MaxFieldSize.
This is a compile-time constant that defaults to 18, which is large enough for
a scalar, vector, tensor, normal, and texture coordinate to be included at each point.
Normally, you will not call \a SetFieldSize() directly; instead, subclasses of
vtkEdgeSubdivisionCriterion, such as vtkShoeMeshSubdivisionAlgorithm, will call it
for you.
In any event, setting \a FieldSize to a non-zero value means you must pass field
values to the \p AdaptivelySamplekFacet routines; For example,
@verbatim
vtkStreamingTessellator* t = vtkStreamingTessellator::New();
t->SetFieldSize( 1, 3 );
t->SetEmbeddingDimension( 1, 1 ); // not really required, this is the default
double p0[3+1+3] = { x0, y0, z0, r0, fx0, fy0, fz0 };
double p1[3+1+3] = { x1, y1, z1, r1, fx1, fy1, fz1 };
t->AdaptivelySample1Facet( p0, p1 );
@endverbatim
This would adaptively sample an curve (1-facet) with geometry and
a vector field at every output point on the curve.
int = obj.GetFieldSize (int k) const
- Get/Set the number of field value coordinates associated with each input and output point.
The default is 0; no field values are interpolated.
You may specify a different size, \a s, for each type of \a k -facet to be processed.
For example, \p SetFieldSize( \p 2, \p 3 ) would associate 3 field value coordinates
with each input and output point of an \p AdaptivelySample2Facet call,
but does not say anything about input or output points of \p AdaptivelySample1Facet.
Call \p SetFieldSize( \p -1, \a s ) to specify the same dimension for all possible \a k values.
\a s may not exceed vtkStreamingTessellator::MaxFieldSize.
This is a compile-time constant that defaults to 18, which is large enough for
a scalar, vector, tensor, normal, and texture coordinate to be included at each point.
Normally, you will not call \a SetFieldSize() directly; instead, subclasses of
vtkEdgeSubdivisionCriterion, such as vtkShoeMeshSubdivisionAlgorithm, will call it
for you.
In any event, setting \a FieldSize to a non-zero value means you must pass field
values to the \p AdaptivelySamplekFacet routines; For example,
@verbatim
vtkStreamingTessellator* t = vtkStreamingTessellator::New();
t->SetFieldSize( 1, 3 );
t->SetEmbeddingDimension( 1, 1 ); // not really required, this is the default
double p0[3+1+3] = { x0, y0, z0, r0, fx0, fy0, fz0 };
double p1[3+1+3] = { x1, y1, z1, r1, fx1, fy1, fz1 };
t->AdaptivelySample1Facet( p0, p1 );
@endverbatim
This would adaptively sample an curve (1-facet) with geometry and
a vector field at every output point on the curve.
obj.SetMaximumNumberOfSubdivisions (int num_subdiv_in)
- Get/Set the maximum number of subdivisions that may occur.
int = obj.GetMaximumNumberOfSubdivisions ()
- Get/Set the maximum number of subdivisions that may occur.
obj.AdaptivelySample3Facet (double v1, double v2, double v3, double v4) const
- This will adaptively subdivide the tetrahedron (3-facet),
triangle (2-facet), or edge (1-facet) until the subdivision
algorithm returns false for every edge or the maximum recursion
depth is reached.
Use \p SetMaximumNumberOfSubdivisions to change the maximum
recursion depth.
The AdaptivelySample0Facet method is provided as a convenience.
Obviously, there is no way to adaptively subdivide a vertex.
Instead the input vertex is passed unchanged to the output
via a call to the registered VertexProcessorFunction callback.
.SECTION Warning
This assumes that you have called SetSubdivisionAlgorithm(),
SetEdgeCallback(), SetTriangleCallback(), and SetTetrahedronCallback()
with valid values!
obj.AdaptivelySample2Facet (double v1, double v2, double v3) const
- This will adaptively subdivide the tetrahedron (3-facet),
triangle (2-facet), or edge (1-facet) until the subdivision
algorithm returns false for every edge or the maximum recursion
depth is reached.
Use \p SetMaximumNumberOfSubdivisions to change the maximum
recursion depth.
The AdaptivelySample0Facet method is provided as a convenience.
Obviously, there is no way to adaptively subdivide a vertex.
Instead the input vertex is passed unchanged to the output
via a call to the registered VertexProcessorFunction callback.
.SECTION Warning
This assumes that you have called SetSubdivisionAlgorithm(),
SetEdgeCallback(), SetTriangleCallback(), and SetTetrahedronCallback()
with valid values!
obj.AdaptivelySample1Facet (double v1, double v2) const
- This will adaptively subdivide the tetrahedron (3-facet),
triangle (2-facet), or edge (1-facet) until the subdivision
algorithm returns false for every edge or the maximum recursion
depth is reached.
Use \p SetMaximumNumberOfSubdivisions to change the maximum
recursion depth.
The AdaptivelySample0Facet method is provided as a convenience.
Obviously, there is no way to adaptively subdivide a vertex.
Instead the input vertex is passed unchanged to the output
via a call to the registered VertexProcessorFunction callback.
.SECTION Warning
This assumes that you have called SetSubdivisionAlgorithm(),
SetEdgeCallback(), SetTriangleCallback(), and SetTetrahedronCallback()
with valid values!
obj.AdaptivelySample0Facet (double v1) const
- This will adaptively subdivide the tetrahedron (3-facet),
triangle (2-facet), or edge (1-facet) until the subdivision
algorithm returns false for every edge or the maximum recursion
depth is reached.
Use \p SetMaximumNumberOfSubdivisions to change the maximum
recursion depth.
The AdaptivelySample0Facet method is provided as a convenience.
Obviously, there is no way to adaptively subdivide a vertex.
Instead the input vertex is passed unchanged to the output
via a call to the registered VertexProcessorFunction callback.
.SECTION Warning
This assumes that you have called SetSubdivisionAlgorithm(),
SetEdgeCallback(), SetTriangleCallback(), and SetTetrahedronCallback()
with valid values!
obj.ResetCounts ()
- Reset/access the histogram of subdivision cases encountered.
The histogram may be used to examine coverage during testing as well as characterizing the
tessellation algorithm's performance.
You should call ResetCounts() once, at the beginning of a stream of tetrahedra.
It must be called before AdaptivelySample3Facet() to prevent uninitialized memory reads.
These functions have no effect (and return 0) when PARAVIEW_DEBUG_TESSELLATOR has not been defined.
By default, PARAVIEW_DEBUG_TESSELLATOR is not defined, and your code will be fast and efficient. Really!
vtkIdType = obj.GetCaseCount (int c)
- Reset/access the histogram of subdivision cases encountered.
The histogram may be used to examine coverage during testing as well as characterizing the
tessellation algorithm's performance.
You should call ResetCounts() once, at the beginning of a stream of tetrahedra.
It must be called before AdaptivelySample3Facet() to prevent uninitialized memory reads.
These functions have no effect (and return 0) when PARAVIEW_DEBUG_TESSELLATOR has not been defined.
By default, PARAVIEW_DEBUG_TESSELLATOR is not defined, and your code will be fast and efficient. Really!
vtkIdType = obj.GetSubcaseCount (int casenum, int sub)