cugl::Geometry Class Reference

#include <CUGeometry.h>

Public Types
enum	Value : uint8_t { IMPLICIT, POINTS, PATH, SOLID }

Public Member Functions
	Geometry ()=default
constexpr	Geometry (Value val)
	operator Value () const
	operator bool ()=delete
GLuint	glCommand () const
bool	matches (const std::vector< Uint32 > &indices) const
bool	matches (const Uint32 *indices, size_t size) const
std::vector< Uint32 >	index (const std::vector< Vec2 > &vertices) const
std::vector< Uint32 >	index (const std::vector< float > &vertices) const
std::vector< Uint32 >	index (const float *vertices, size_t size) const
std::unordered_set< Uint32 >	exterior (const std::vector< Uint32 > indices) const
std::unordered_set< Uint32 >	exterior (const Uint32 *indices, size_t size) const
std::vector< std::vector< Uint32 > >	boundaries (const std::vector< Uint32 > &indices) const
std::vector< std::vector< Uint32 > >	boundaries (const Uint32 *indices, size_t size) const

Static Public Member Functions
static Geometry	categorize (const std::vector< Uint32 > &indices)
static Geometry	categorize (const Uint32 *indices, size_t size)

Detailed Description

This class (enum) is used to provide meaning to a set of vertex indices.

Rendering classes should use this type as a hint for how to render a polygon. In particular, the method glCommand() returns an OpenGL command corresponding the index geometry.

There is not a one-to-one correspondence between geometries and OpenGL drawing commands. Geometries are concerned with structure, not the optimal organization of indices. So we do not make a distinction between triangles, triangle strips, or triangle fans.

In fact, the geometries supported all correspond to decomposable indices. These are the OpenGL commands TRIANGLES, LINES, or POINTS. These correspond to indices that we can easily split and combine on vertex concatenation.

Member Enumeration Documentation

Value

enum cugl::Geometry::Value : uint8_t

The geometry values

This enum provides the values for this enum class, but the code in this class allows us to "skip over" it. That is, we can write Geometry::IMPLICIT in place of Geometry::Value::IMPLICIT, and so on. So we can ignore this enum.

Enumerator
IMPLICIT	This geometry corresponds to an empty set of vertices. In this case the geometry is inferred from the vertex order. Such an object is useful for pure math computations, but has no drawing mesh.
POINTS	This geometry corresponds to an unstructured set of indices.. In this case indices represent the vertices that are relevant. Vertices not in the index set are ignored.
PATH	This geometry represents an open path outline. The indices are a sequence of line segments. That is, the number of indices is divisible by two, with each pair forming a segment. These segments may or be linked to each other.
SOLID	This geometry represents a solid shape with a triangulation. The indices are a sequence of triangles. That is, the number of indices is divisible by three, with each triplet forming a triangle.

Constructor & Destructor Documentation

Geometry() [1/2]

cugl::Geometry::Geometry ( )

default

Disable the explicity Geomtry constructor

Geometry() [2/2]

constexpr cugl::Geometry::Geometry ( Value  val )

inlineconstexpr

Construct a geometry with one of the enumeration values

Parameters

val	The enumeration value

Member Function Documentation

boundaries() [1/2]

std::vector<std::vector<Uint32> > cugl::Geometry::boundaries

(

const std::vector< Uint32 > &

indices

)

const

Returns the connected boundary components for this geometry.

This method allows us to reconstruct the exterior boundary of a solid shape, or to compose a pathwise connected curve into components.

For the geometry PATH, the elements of the outer array are connected links of line segments, with the indices listed in order of the path traversal. A closed path only has one component, so the outer array will only have one element. An open path may have several components if there are discontinuities in the path. IMPLICIT geometries are treated as if they were closed a closed path.

For SOLID, this method detriangulates the mesh, returning the outer hull, and discarding any interior points. This hull need not be convex (as concave or convex cannot be inferred from indices alone). If the mesh represents a simple polygon, only one boundary will be returned. If the mesh is not continuous, the outer array will contain the boundary of each disjoint polygon. If the mesh has holes, each hole will be returned as a separate boundary. There is no guarantee on the order of boundaries returned.

For POINTS, there are no connected components, so this method returns an array of single-element arrays. IMPLICIT returns null, as it is impossible to infer the boundaries without the vertices.

Parameters

indices

The indices to process

Returns

the connected boundary components for this geometry.

boundaries() [2/2]

std::vector<std::vector<Uint32> > cugl::Geometry::boundaries	(	const Uint32 *	indices,
		size_t	size
	)		const

Returns the connected boundary components for this geometry.

This method allows us to reconstruct the exterior boundary of a solid shape, or to compose a pathwise connected curve into components.

For the geometry PATH, the elements of the outer array are connected links of line segments, with the indices listed in order of the path traversal. A closed path only has one component, so the outer array will only have one element. An open path may have several components if there are discontinuities in the path. IMPLICIT geometries are treated as if they were closed a closed path.

For SOLID, this method detriangulates the mesh, returning the outer hull, and discarding any interior points. This hull need not be convex (as concave or convex cannot be inferred from indices alone). If the mesh represents a simple polygon, only one boundary will be returned. If the mesh is not continuous, the outer array will contain the boundary of each disjoint polygon. If the mesh has holes, each hole will be returned as a separate boundary. There is no guarantee on the order of boundaries returned.

For POINTS, there are no connected components, so this method returns an array of single-element arrays. IMPLICIT returns null, as it is impossible to infer the boundaries without the vertices.

Parameters

indices	The indices to process
size	The index size

Returns

the connected boundary components for this geometry.

categorize() [1/2]

static Geometry cugl::Geometry::categorize ( const std::vector< Uint32 > &  indices )

static

Returns the best guess for the geometry of these indices

If the indices form a continuous link of pairs, then it will identify them as a PATH. Otherwise, this method can only guess at the geometry from the parity. If the number of indices is a multiple of 3, it assumes it is SOLID. If it is a multiple of 2, it assumes it is a PATH. If neither is true, it identifies these as POINTS. It is only IMPLICIT if indices are null or empty.

Parameters

indices

The indices to analyze

Returns

the best guess for the geometry of these indices

categorize() [2/2]

static Geometry cugl::Geometry::categorize ( const Uint32 *  indices, size_t  size  )

static

Returns the best guess for the geometry of these indices

If the indices form a continuous link of pairs, then it will identify them as a PATH. Otherwise, this method can only guess at the geometry from the parity. If the number of indices is a multiple of 3, it assumes it is SOLID. If it is a multiple of 2, it assumes it is a PATH. If neither is true, it identifies these as POINTS. It is only IMPLICIT if indices are null or empty.

Parameters

indices	The indices to analyze
size	The index size

Returns

the best guess for the geometry of these indices

exterior() [1/2]

std::unordered_set<Uint32> cugl::Geometry::exterior

(

const std::vector< Uint32 >

indices

)

const

Returns the set of indices that are on a boundary of this geometry

This method is really only relevant for SOLID geometry. For other geometries, it simply returns a set of the indices. In the case of solid geometries, it identifies the outer hull (not necessarily convex).

This method can identify the outer hull using the graph properties of the triangle mesh. An internal node if the number of neighbors is the same as the number of attached triangles. An index that is not internal is external.

Unlike boundaries, this method does not order the boundary indices or decompose them into connected components.

Parameters

indices

The indices to process

Returns

the set of indices that are on a boundary of this geometry

exterior() [2/2]

std::unordered_set<Uint32> cugl::Geometry::exterior	(	const Uint32 *	indices,
		size_t	size
	)		const

Returns the set of indices that are on a boundary of this geometry

This method is really only relevant for SOLID geometry. For other geometries, it simply returns a set of the indices. In the case of solid geometries, it identifies the outer hull (not necessarily convex).

This method can identify the outer hull using the graph properties of the triangle mesh. An internal node if the number of neighbors is the same as the number of attached triangles. An index that is not internal is external.

Unlike boundaries, this method does not order the boundary indices or decompose them into connected components.

Parameters

indices	The indices to process
size	The index size

Returns

the set of indices that are on a boundary of this geometry

glCommand()

GLuint cugl::Geometry::glCommand

(

)

const

Returns the OpenGL drawing code for this geometry

If the geometry is IMPLICIT, this method returns -1.

Returns

the OpenGL drawing code for this geometry

index() [1/3]

std::vector<Uint32> cugl::Geometry::index	(	const float *	vertices,
		size_t	size
	)		const

Returns a canonical set of indices for the given vertex list

This takes a vertex list and generates a canonical index list that matches this geometry. For IMPLICIT, this will return an empty index set. For POINTS it is the index of all the vertices. For PATH, it is a linked set of segment indices of the vertices in order.

For SOLID it triangulates the vertices. For simplicity, this method does not support vertex lists with self-crossings. Such polygons will need to be decomposed first. This method always uses a SimpleTriangulator. If you wish to use another algorithm (such as ComplexTriangulator), you will need to generate the indices separately.

Parameters

vertices	The vertex list
size	The number of vertices

Returns

a canonical set of indices for the given vertex list

index() [2/3]

std::vector<Uint32> cugl::Geometry::index

(

const std::vector< float > &

vertices

)

const

Returns a canonical set of indices for the given vertex list

This takes a vertex list and generates a canonical index list that matches this geometry. For IMPLICIT, this will return an empty index set. For POINTS it is the index of all the vertices. For PATH, it is a linked set of segment indices of the vertices in order.

For SOLID it triangulates the vertices. For simplicity, this method does not support vertex lists with self-crossings. Such polygons will need to be decomposed first. This method always uses a SimpleTriangulator. If you wish to use another algorithm (such as ComplexTriangulator), you will need to generate the indices separately.

Parameters

vertices

The vertex list

Returns

a canonical set of indices for the given vertex list

index() [3/3]

std::vector<Uint32> cugl::Geometry::index

(

const std::vector< Vec2 > &

vertices

)

const

Returns a canonical set of indices for the given vertex list

This takes a vertex list and generates a canonical index list that matches this geometry. For IMPLICIT, this will return an empty index set. For POINTS it is the index of all the vertices. For PATH, it is a linked set of segment indices of the vertices in order.

For SOLID it triangulates the vertices. For simplicity, this method does not support vertex lists with self-crossings. Such polygons will need to be decomposed first. This method always uses a SimpleTriangulator. If you wish to use another algorithm (such as ComplexTriangulator), you will need to generate the indices separately.

Parameters

vertices

The vertex list

Returns

a canonical set of indices for the given vertex list

matches() [1/2]

bool cugl::Geometry::matches

(

const std::vector< Uint32 > &

indices

)

const

Returns true if the indices match this geometry.

If the geometry is SOLID, this method will only detect if the number of indices is divisible by three. For PATH, it will detect if the number of indices is not divisible by 2. Finally, IMPLICIT requires the indices be empty, but the geometry POINTS matches anything.

This method does not validate that the indices are with in range, as that requires knowledge of the vertices.

Parameters

indices

The indices to test

Returns

true if the indices match this geometry.

matches() [2/2]

bool cugl::Geometry::matches	(	const Uint32 *	indices,
		size_t	size
	)		const

Returns true if the indices match this geometry.

If the geometry is SOLID, this method will only detect if the number of indices is divisible by three. For PATH, it will detect if the number of indices is not divisible by 2. Finally, IMPLICIT requires the indices be empty, but the geometry POINTS matches anything.

This method does not validate that the indices are with in range, as that requires knowledge of the vertices.

Parameters

indices	The indices to test
size	The index size

Returns

true if the indices match this geometry.

operator bool()

cugl::Geometry::operator bool ( )

explicitdelete

Prevent usage: if(geom)

operator Value()

cugl::Geometry::operator Value ( ) const

inline

Returns the current value of this geometry.

This method enables both direct comparisons and swith statements.

Returns

the current value of this geometry.

---

The documentation for this class was generated from the following file:

• include/cugl/math/CUGeometry.h