TiledShape.h

Classes

TiledShape -- Define the shape and tile shape (full description)

class TiledShape

Interface

Public Members
TiledShape()
TiledShape (const IPosition& shape)
TiledShape (const IPosition& shape, const IPosition& tileShape)
TiledShape (const TiledShape& that)
~TiledShape()
TiledShape& operator= (const TiledShape& that)
Bool isTileShapeDefined() const
const IPosition& shape() const
IPosition tileShape (uInt nrPixelsPerTile = 32768, Double tolerance = 0.5) const
IPosition defaultTileShape (uInt nrPixelsPerTile, Double tolerance) const
IPosition defaultTileShape (uInt nrPixelsPerTile, const Vector<Double>& tolerance, const Vector<Double>& weight) const

Description

Review Status

Reviewed By:
Peter Barnes
Date Reviewed:
1999/10/30
Programs:
Tests:

Prerequisite

Etymology

TiledShape defines the shape and tile shape of a tiled array.

Synopsis

TiledShape is a class defining the shape and optionally the tile shape of a lattice. It is used in the constructors of PagedArray and PagedImage.

In principle it serves as a place holder for the lattice shape and tile shape. The functions shape and tileShape can be used to retrieve the shapes. However, when the tile shape is not given, the function tileShape calculates a default tile shape using the given maximum tile size in pixel elements. The default tile shape is calculated in such a way that the sizes of its axes are proportional to the sizes of the lattice axes. Per axis it is tried as much as possible to fit an integral number of tiles in the lattice.
In this way getting the tile shape is completely transparent.

Example

    // Do not explicitly define a tile shape.
    // This results in a default tile shape (of 32,32,32).
    TiledShape shape(IPosition(3,128,128,128));
    cout << shape.shape() << ' ' << shape.tileShape() << endl;
    
    // Use with an explicitly given tile shape.
    TiledShape shape(IPosition(3,128,128,128), IPosition(3,64,32,8));
    cout << shape.shape() << ' ' << shape.tileShape() << endl;

Motivation

Classes PagedArray and PagedImage contained several duplicated constructors to be able to pass a tile shape. This class makes it possible to have only one constructor instead of two. Furthermore it contains the logic to check if the shapes are conforming and the logic to calculate a default tile shape.

Member Description

TiledShape()

Default constructor has empty shape and tile shape.

TiledShape (const IPosition& shape)

Use the given shape. No tile shape is given, so function tileShape will calculate it using the size of a tile.

TiledShape (const IPosition& shape, const IPosition& tileShape)

Use the given shape and tile shape. Both shapes must be conforming (i.e. have same number of elements).

TiledShape (const TiledShape& that)

Copy constructor (copy semantics).

~TiledShape()

TiledShape& operator= (const TiledShape& that)

Assignment (copy semantics).

Bool isTileShapeDefined() const

Is the tile shape defined?

const IPosition& shape() const

Return the shape.

IPosition tileShape (uInt nrPixelsPerTile = 32768, Double tolerance = 0.5) const

Return the tile shape. When the tile shape is undefined, the default tile shape will be calculated using the given tile size and tolerance.
The tolerance is used to determine the boundaries where it is tried to fit an integral number of tiles.

IPosition defaultTileShape (uInt nrPixelsPerTile, Double tolerance) const
IPosition defaultTileShape (uInt nrPixelsPerTile, const Vector<Double>& tolerance, const Vector<Double>& weight) const

Derive the default tile shape from the shape for the given number of pixels per tile. It is tried to get the same number of tiles for each dimension. When a weight vector is given, the number of tiles for a dimension is proportional to the weight.
After the initial guess it tries to optimize it by trying to waste as little space as possible, while trying to keep as close as possible to the initial guess. The given tolerance (possibly per axis) gives the minimum and maximum possible length of a tile axis (minimum = initial_guess*tolerance; maximum = initial_guess/tolerance). The heuristic is such that a tile axis length dividing the cube length exactly is always favoured. The test program tTiledShape can be used to see how the algorithm works out for a given shape and tile size.