Types are organized in a hierarchy based on the basic interface mi::IData. More...
Modules | |
| Simple Types | |
| This module lists all interfaces related to simple types. | |
| Collections | |
| This module and its submodule Compounds list all interfaces related to collections. | |
Classes | |
| class | mi::IData |
| This interface is the base interface of all types. More... | |
| class | mi::IEnum_decl |
| An enum declaration is used to describe enum types. More... | |
| class | mi::neuraylib::IFactory |
| This API component allows the creation, assignment, and cloning of instances of types. More... | |
| class | mi::IStructure_decl |
| A structure declaration is used to describe structure types. More... | |
| struct | mi::Type_traits<I> |
| Type traits relating interfaces, corresponding primitive types, and their type names. More... | |
Types are organized in a hierarchy based on the basic interface mi::IData.
The hierarchy of types based on mi::IData is split into two branches, simple types (derived from mi::IData_simple) and collections (derived from mi::IData_collection). Simple types are e.g. numbers and strings. Collections are e.g. arrays, structures, maps, vectors, and matrices.
Types are identified by type names. The method mi::IData::get_type_name() returns the type name of an instance of mi::IData. Most notably type names are used to create instances of types via mi::neuraylib::IFactory::create() , and to create attributes or to query their type.
Since there is a very high number of interfaces related to types they are not listed in this module. Rather they have been split into three submodules Simple Types, Collections, and Compounds (compounds are a subset of the collections).
Simple types are booleans, integral and floating-point numbers, strings, UUIDs, void, pointers, and references. All interfaces of the simple types are derived from mi::IData_simple. For most simple types there is an equivalent C++ class. With the exception of pointers, references, and enums the type name of simple types equals the interface name without the leading "I". See also the module Simple Types.
| Type name | Attribute type | Interface | C++ class | Description |
|---|---|---|---|---|
"Boolean" | yes | mi::IBoolean | bool | Boolean |
"Sint8" | yes | mi::ISint8 | mi::Sint8 | Signed 8-bit integer |
"Sint16" | yes | mi::ISint16 | mi::Sint16 | Signed 16-bit integer |
"Sint32" | yes | mi::ISint32 | mi::Sint32 | Signed 32-bit integer |
"Sint64" | yes | mi::ISint64 | mi::Sint64 | Signed 64-bit integer |
"Uint8" | no | mi::IUint8 | mi::Uint8 | Unsigned 8-bit integer |
"Uint16" | no | mi::IUint16 | mi::Uint16 | Unsigned 16-bit integer |
"Uint32" | no | mi::IUint32 | mi::Uint32 | Unsigned 32-bit integer |
"Uint64" | no | mi::IUint64 | mi::Uint64 | Unsigned 64-bit integer |
"Float32" | yes | mi::IFloat32 | mi::Float32 | 32-bit IEEE-754 single-precision floating-point number |
"Float64" | yes | mi::IFloat64 | mi::Float64 | 64-bit IEEE-754 double-precision floating-point number |
"Difference" | no | mi::IDifference | mi::Difference | Signed 32-bit or 64-bit integer, depending on the architecture |
"Size" | no | mi::ISize | mi::Size | Unsigned 32-bit or 64-bit integer, depending on the architecture |
"String" | yes | mi::IString | const char* | String representation in ISO-8859-1 encoding |
"Uuid" | no | mi::IUuid | mi::base::Uuid | Universally unique identifier (UUID) |
"Void" | no | mi::IVoid | void | Void type |
"Pointer<T>" | no | mi::IPointer | - | Mutable pointer to an instance of type T |
"Const_pointer<T>" | no | mi::IConst_pointer | - | Const pointer to an instance of type T |
| as registered (1) | yes | mi::IEnum | - | Subsets of mi::Sint32 values, identified by strings |
(1) The type name of an enum is specific to the actual enum type. There is no naming pattern as for the other simple types. User-defined enum types require the registration of a corresponding enum declaration with a type name (see mi::neuraylib::IExtension_api::register_enum_decl()).
Collections are static and dynamic arrays, maps, structures, and compounds (for compounds see separate section below). All interfaces of collections are derived from mi::IData_collection. For all collection types there are similar constructs in C++, but there are no directly equivalent C++ classes as for the simple types. The type names of collections are typically constructed according to certain rules, typically involving the type name of the contained elements or values. See the table below for examples or the corresponding interfaces for an exact description. See also the module Collections.
| Type name | Attribute type | Interface | C++ class | Description |
|---|---|---|---|---|
"T[N]" | yes (2)(3) | mi::IArray | - | Static array of N elements of type T, N > 0 |
"T[]" | yes (2) | mi::IDynamic_array | - | Dynamic array of elements of type T |
"Map<T>" | no | mi::IMap | - | Set of key-value pairs, values are of type T |
| as registered (4) | yes | mi::IStructure | - | Ordered set of key-value pairs, values are of arbitrary types |
(2) For attributes, the type of array elements must not be an array (but it can be a structure with arrays as members).
(3) For attributes, the array length N must not be zero.
(4) The type name of a structure is specific to the actual structure type. There is no naming pattern as for arrays or maps. User-defined structure types require the registration of a corresponding structure declaration with a type name (see mi::neuraylib::IExtension_api::register_structure_decl()).
Compound types are vectors, matrices, colors, spectrums, and bounding boxes. Actually, compounds belong to the set of collections, they are not a third kind of type besides simple types and collections.In the documentation they are split into their own module due to the very high number of interfaces and typedefs for compounds. See also the module Compounds.
For all compound types there is a C++ class from the math API as counterpart. Type names for vectors and matrices are based on the type name of the element followed by the dimension in angle brackets ("<N>" or "<M,N>", N = 2, 3, or 4).
| Type name | Attribute type | Interface | C++ class | Description |
|---|---|---|---|---|
"Boolean<N>" | yes | mi::IBoolean_2, ... | mi::Boolean_2, ... | N x bool vector, N = 2, 3, or 4 |
"Sint32<N>" | yes | mi::ISint32_2, ... | mi::Sint32_2, ... | N x Sint32 vector, N = 2, 3, or 4 |
"Uint32<N>" | no | mi::IUint32_2, ... | mi::Uint32_2, ... | N x Uint32 vector, N = 2, 3, or 4 |
"Float32<N>" | yes | mi::IFloat32_2, ... | mi::Float32_2, ... | N x Float32 vector, N = 2, 3, or 4 |
"Float64<N>" | yes | mi::IFloat64_2, ... | mi::Float64_2, ... | N x Float64 vector, N = 2, 3, or 4 |
"Boolean<M,N>" | no | mi::IBoolean_2_2, ... | mi::Boolean_2_2, ... | M x N matrix of bool, M, N = 2, 3, or 4 |
"Sint32<M,N>" | no | mi::ISint32_2_2, ... | mi::Sint32_2_2, ... | M x N matrix of Sint32, M, N = 2, 3, or 4 |
"Uint32<M,N>" | no | mi::IUint32_2_2, ... | mi::Uint32_2_2, ... | M x N matrix of Uint32, M, N = 2, 3, or 4 |
"Float32<M,N>" | yes | mi::IFloat32_2_2, ... | mi::Float32_2_2, ... | M x N matrix of Float32, M, N = 2, 3, or 4 |
"Float64<M,N>" | yes | mi::IFloat64_2_2, ... | mi::Float64_2_2, ... | M x N matrix of Float64, M, N = 2, 3, or 4 |
"Color" | yes | mi::IColor | mi::Color | 4 x Float32 representing RGBA color |
"Color3" | yes | mi::IColor3 | mi::Color | 3 x Float32 representing RGB color |
"Spectrum" | yes | mi::ISpectrum | mi::Spectrum | 3 x Float32 representing three color bands |
"Bbox3" | no | mi::IBbox3 | mi::Bbox3 | Bounding box, represented by two mi::Float32_3 |
Pixel types denote the type of the pixel of a canvas or image. The table below lists the valid pixel types. See also mi::neuraylib::ICanvas and mi::neuraylib::ITile.
| Type name | Interface | C++ class | Description |
|---|---|---|---|
"Sint8" | mi::ISint8 | mi::Sint8 | Signed 8-bit integer (6) |
"Sint32" | mi::ISint32 | mi::Sint32 | Signed 32-bit integer |
"Float32" | mi::IFloat32 | mi::Float32 | 32-bit IEEE-754 single-precision floating-point number |
"Float32<2>" | mi::IFloat32_2 | mi::Float32_2 | 2 x Float32 |
"Float32<3>" | mi::IFloat32_3 | mi::Float32_3 | 3 x Float32 |
"Float32<4>" | mi::IFloat32_4 | mi::Float32_4 | 4 x Float32 |
"Rgb" | - | - | 3 x Uint8 representing RGB color |
"Rgba" | - | - | 4 x Uint8 representing RGBA color |
"Rgbe" | - | - | 4 x Uint8 representing RGBE color |
"Rgbea" | - | - | 5 x Uint8 representing RGBEA color |
"Rgb_16" | - | - | 3 x Uint16 representing RGB color |
"Rgba_16" | - | - | 4 x Uint16 representing RGBA color |
"Rgb_fp" | - | - | 3 x Float32 representing RGB color |
"Color" | mi::IColor | mi::Color | 4 x Float32 representing RGBA color |
(6) For most purposes, in particular for pixel type conversion, the data is actually treated as unsigned 8-bit integer.
