Interface NativeType<T extends NativeType<T>>

All Superinterfaces:
Type<T>, ValueEquals<T>
All Known Implementing Classes:
AbstractBit64Type, AbstractBitType, AbstractIntegerBitType, AbstractNativeType, ARGBType, BasePairBitType, BasePairCharType, BitType, ByteType, ComplexDoubleType, ComplexFloatType, DoubleType, FloatType, GenericByteType, GenericIntType, GenericLongType, GenericShortType, IntType, LongType, NativeARGBDoubleType, ShortType, Unsigned128BitType, Unsigned12BitType, Unsigned2BitType, Unsigned4BitType, UnsignedByteType, UnsignedIntType, UnsignedLongType, UnsignedShortType, UnsignedVariableBitLengthType

public interface NativeType<T extends NativeType<T>> extends Type<T>
A NativeType is a Type that that provides access to data stored in Java primitive arrays. To this end, implementations maintain a reference to the current storage array and the index of an element in that array. The NativeType is positioned on the correct storage array and index by accessors (Cursors and RandomAccesses ).

The NativeType is the only class that is aware of the actual data type, i.e., which Java primitive type is used to store the data. On the other hand it does not know the storage layout, i.e., how n-dimensional pixel coordinates map to indices in the current array. It also doesn't know whether and how the data is split into multiple chunks. This is determined by the container implementation (e.g., ArrayImg, CellImg, ...). Separating the storage layout from access and operations on the Type avoids re-implementation for each container type.

Author:
Stephan Preibisch, Stephan Saalfeld, Tobias Pietzsch
  • Method Details

    • getEntitiesPerPixel

      Fraction getEntitiesPerPixel()
      Get the number of entities in the storage array required to store one pixel value. A pixel value may be spread over several or less than one entity. For example, a complex number may require 2 entries of a float[] array to store one pixel. Or a 12-bit type might need 12/64th entries of a long[] array.
      Returns:
      the number of storage type entities required to store one pixel value.
    • createSuitableNativeImg

      NativeImg<T,?> createSuitableNativeImg(NativeImgFactory<T> storageFactory, long[] dim)
      The NativeType creates the NativeImg used for storing image data; based on the given storage strategy and its size. It basically only decides here which BasicType it uses (float, int, byte, bit, ...) and how many entities per pixel it needs (e.g. 2 floats per pixel for a complex number). This enables the separation of containers and the basic types.
      Parameters:
      storageFactory - which storage strategy is used
      dim - the dimensions
      Returns:
      the instantiated NativeImg where only the Type knows the BasicType it contains.
    • duplicateTypeOnSameNativeImg

      T duplicateTypeOnSameNativeImg()
      Creates a new NativeType which stores in the same physical array. This is only used internally.
      Returns:
      a new NativeType instance working on the same NativeImg
    • updateContainer

      void updateContainer(Object c)
      This method is used by an accessor (e.g., a Cursor) to request an update of the current data array.

      As an example consider a CellCursor moving on a CellImg. The cursor maintains a NativeType which provides access to the image data. When the cursor moves from one cell to the next, the underlying data array of the NativeType must be switched to the data array of the new cell.

      To achieve this, the CellCursor calls updateContainer() with itself as the argument. updateContainer() in turn will call NativeImg.update(Object) on it's container, passing along the reference to the cursor. In this example, the container would be a CellImg. While the NativeType does not know about the type of the cursor, the container does. CellImg knows that it is passed a CellCursor instance, which can be used to figure out the current cell and the underlying data array, which is then returned to the NativeType.

      The idea behind this concept is maybe not obvious. The NativeType knows which basic type is used (float, int, byte, ...). However, it does not know how the data is stored (ArrayImg, CellImg, ...). This prevents the need for multiple implementations of NativeType .

      Parameters:
      c - reference to an accessor which can be passed on to the container (which will know what to do with it).
    • updateIndex

      void updateIndex(int i)
      Set the index into the current data array.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

      Parameters:
      i - the new array index
    • getIndex

      int getIndex()
      Get the current index into the current data array.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

      Returns:
      the current index into the underlying data array
    • incIndex

      void incIndex()
      Increment the index into the current data array.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

    • incIndex

      void incIndex(int increment)
      Increases the index into the current data array by increment steps.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

      Parameters:
      increment - how many steps
    • decIndex

      void decIndex()
      Decrement the index into the current data array.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

    • decIndex

      void decIndex(int decrement)
      Decrease the index into the current data array by decrement steps.

      This is used by accessors (e.g., a Cursor) to position the NativeType in the container.

      Parameters:
      decrement - how many steps