sim.field.grid.Grid2D Maven / Gradle / Ivy
Show all versions of mason Show documentation
/*
Copyright 2006 by Sean Luke and George Mason University
Licensed under the Academic Free License version 3.0
See the file "LICENSE" for more information
*/
package sim.field.grid;
import sim.util.IntBag;
import java.util.Map;
// (stars down the left are added there to keep the formatting correct)
/**
* Define basic neighborhood functions for 2D Grids. The basic interface defines a width and a height
* (not all grids require a width and a height unless you're doing toroidal grids), and basic math for
* toroidal computation, hex grid location, and triangular grid location.
*
* Toroidal Computation
*
* If you're using the Grid to define a toroidal (wrap-around) world, you can use the tx
* and ty methods to simplify the math for you. For example, to increment in the x direction,
* including wrap-around, you can do: x = tx(x+1).
*
*
If you're sure that the values you'd pass into the toroidal functions would not wander off more than
* a grid dimension in either direction (height, width), you can use the slightly faster toroidal functions
* stx and sty instead. For example, to increment in the x direction,
* including wrap-around, you can do: x = stx(x+1). See the documentation on these functions for
* when they're appropriate to use. Under most common situations, they're okay.
*
*
In HotSpot 1.4.1, stx, and sty are inlined. In Hotspot 1.3.1, they are not (they contain if-statements).
*
*
Hex Grid Computation
* Grids can be used for both squares and hex grids. Hex grids are stored in an ordinary
* rectangular array and are defined as follows:
*
*
*
* (0,0) (2,0) (4,0) (6,0) ...
* (1,0) (3,0) (5,0) (7,0) ...
* (0,1) (2,1) (4,1) (6,1) ...
* (1,1) (3,1) (5,1) (7,1) ...
* (0,2) (2,2) (4,2) (6,2) ...
* (1,2) (3,2) (5,2) (7,2) ...
* ... ... ... ... ...
* ... ... ... ... ...
*
*
*
*The rules moving from a hex location (at CENTER) to another one are as follows:
*
*
*
* UP
* x
* UPLEFT y - 1 UPRIGHT
* x - 1 x + 1
* ((x % 2) == 0) ? y - 1 : y CENTER ((x % 2) == 0) ? y - 1 : y
* x
* DOWNLEFT y DOWNRIGHT
* x - 1 x + 1
* ((x % 2) == 0) ? y : y + 1 DOWN ((x % 2) == 0) ? y : y + 1
* x
* y + 1
*
*
*
* NOTE: (x % 2 == 0), that is, "x is even", may be written instead in this faster way: ((x & 1) == 0)
*
* Because the math is a little hairy, we've provided the math for the UPLEFT, UPRIGHT, DOWNLEFT,
* and DOWNRIGHT directions for you. For example, the UPLEFT location from [x,y] is at
* [ulx(x,y) , uly(x,y)]. Additionally, the toroidal methods can be used in conjunction with the
* hex methods to implement a toroidal hex grid. Be sure to To use a toroidal hex grid properly,
* you must ensure that height of the grid is an even number. For example, the toroidal
* UPLEFT X location is at tx(ulx(x,y)) and the UPLEFT Y location is at ty(uly(x,y)). Similarly,
* you can use stx and sty.
*
*
While this interface defines various methods common to many grids, you should endeavor not to
* call these grids casted into this interface: it's slow. If you call the grids' methods directly
* by their class, their methods are almost certain to be inlined into your code, which is very fast.
*
*
Triangular Grid Computation
*
* Grids can also be used for triangular grids instead of squares. Triangular grids look like this:
*
*
* -------------------------
* \(0,0)/ \(2,0)/ \(4,0)/ \
* \ / \ / \ / \ ...
* \ /(1,0)\ /(3,0)\ /(5,0)\
* -------------------------
* / \(1,1)/ \(3,1)/ \(5,1)/
* / \ / \ / \ / ...
* /(0,1)\ /(2,1)\ /(4,1)\ /
* -------------------------
* \(0,2)/ \(2,2)/ \(4,2)/ \
* \ / \ / \ / \ ...
* \ /(1,2)\ /(3,2)\ /(5,2)\
* -------------------------
* / \(1,3)/ \(3,3)/ \(5,3)/
* / \ / \ / \ / ...
* /(0,3)\ /(2,3)\ /(4,3)\ /
* -------------------------
* .
* .
* .
*
*
* How do you get around such a beast? Piece of cake! Well, to go to your right or left
* neighbor, you just add or subtract the X value. To go to your up or down neighbor, all you
* do is add or subtract the Y value. All you need to know is if your triangle has an edge on
* the top (so you can go up) or an edge on the bottom (so you can go down). The functions TRB
* (triangle with horizontal edge on 'bottom') and TRT (triangle with horizontal edge on 'top')
* will tell you this.
*
*
Like the others, the triangular grid can also be used in toroidal fashion, and the
* toroidal functions should work properly with it. To use a toroidal triangular grid,
* you should ensure that your grid's length and width are both even numbers.
*
*
We'll provide a distance-measure function for triangular grids just as soon as we figure out
* what the heck one looks like. :-)
*/
public interface Grid2D extends java.io.Serializable
{
/** Returns the width of the field. */
public int getWidth();
/** Returns the width of the field. */
public int getHeight();
/** Toroidal x. The following definition:
final int length = this.length;
if (z >= 0) return (z % length);
final int length2 = (z % length) + length;
if (length2 < length) return length2;
return 0;
... produces the correct code and is 27 bytes, so it's likely to be inlined in Hotspot for 1.4.1.
*/
public int tx(final int x);
/** Toroidal y. The following definition:
final int length = this.length;
if (z >= 0) return (z % length);
final int length2 = (z % length) + length;
if (length2 < length) return length2;
return 0;
... produces the correct code and is 27 bytes, so it's likely to be inlined in Hotspot for 1.4.1.
*/
public int ty(final int y);
/** Simple [and fast] toroidal x. Use this if the values you'd pass in never stray
beyond (-width ... width * 2) not inclusive. It's a bit faster than the full
toroidal computation as it uses if statements rather than two modulos.
The following definition:
{ int width = this.width; if (x >= 0) { if (x < width) return x; return x - width; } return x + width; }
...produces the shortest code (24 bytes) and is inlined in Hotspot for 1.4.1. However
in most cases removing the int width = this.width; is likely to be a little faster if most
objects are usually within the toroidal region. */
public int stx(final int x);
/** Simple [and fast] toroidal y. Use this if the values you'd pass in never stray
beyond (-height ... height * 2) not inclusive. It's a bit faster than the full
toroidal computation as it uses if statements rather than two modulos.
The following definition:
{ int height = this.height; if (y >= 0) { if (y < height) return y ; return y - height; } return y + height; }
...produces the shortest code (24 bytes) and is inlined in Hotspot for 1.4.1. However
in most cases removing the int height = this.height; is likely to be a little faster if most
objects are usually within the toroidal region. */
public int sty(final int y);
/** Hex upleft x. */
public int ulx(final int x, final int y);
/** Hex upleft y. */
public int uly(final int x, final int y);
/** Hex upright x.*/
public int urx(final int x, final int y);
/** Hex upright y.*/
public int ury(final int x, final int y);
/** Hex downleft x.*/
public int dlx(final int x, final int y);
/** Hex downleft y.*/
public int dly(final int x, final int y);
/** Hex downright x. */
public int drx(final int x, final int y);
/** Hex downright y. */
public int dry(final int x, final int y);
/** Hex up x. */
public int upx(final int x, final int y);
/** Hex up y. */
public int upy(final int x, final int y);
/** Hex down x. */
public int downx(final int x, final int y);
/** Hex down y. */
public int downy(final int x, final int y);
/** Horizontal edge is on the bottom for triangle. True if x + y is odd.
One definition of this is return ((x + y) & 1) == 1;*/
public boolean trb(final int x, final int y);
/** Horizontal edge is on the top for triangle. True if x + y is even.
One definition of this is return ((x + y) & 1) == 0;*/
public boolean trt(final int x, final int y);
/** Bounded Mode for neighborhood lookup. Indicates that the Grid2D in question
is being used in a way that assumes that it
has no valid locations outside of the rectangle starting at (0,0) and
ending at (width-1, height-1) inclusive. */
public static int BOUNDED = 0;
/** Bounded Mode for neighborhood lookup. Indicates that the Grid2D in question
is being used in a way that assumes that any numerical location is a valid
location. Note that Grid2D subclasses based on arrays, such as DoubleGrid2D,
IntGrid2D, ObjectGrid2D, and DenseGrid2D, cannot be used in an unbounded
fashion. */
public static int UNBOUNDED = 1;
/** Bounded Mode for toroidal lookup. Indicates that the Grid2D in question
is being used in a way that assumes that it is bounded, but wrap-around: for
example, (0,0) is located one away diagonally from (width-1, height-1).
*/
public static int TOROIDAL = 2;
/** Center measurement rule for raidal neighborhood lookup. Indicates that
radial lookup will include locations whose grid cell centers overlap with
the neighborhood region. */
public static int CENTER = 1024;
/** "All" measurement rule for raidal neighborhood lookup. Indicates that
radial lookup will include locations whose grid cells are entirely within
the neighborhood region. */
public static int ALL = 1025;
/** "Any" measurement rule for raidal neighborhood lookup. Indicates that
radial lookup will include locations whose grid cells have any overlap
at all with the neighborhood region. */
public static int ANY = 1026;
/**
* Gets all neighbors of a location that satisfy max( abs(x-X) , abs(y-Y) ) <= dist. This region forms a
* square 2*dist+1 cells across, centered at (X,Y). If dist==1, this
* is equivalent to the so-called "Moore Neighborhood" (the eight neighbors surrounding (X,Y)), plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may only run in two modes: toroidal or bounded. Unbounded lookup is not permitted, and so
* this function is deprecated: instead you should use the other version of this function which has more functionality.
* If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
The origin -- that is, the (x,y) point at the center of the neighborhood -- is always included in the results.
*
*
This function is equivalent to: getNeighborsMaxDistance(x,y,dist,toroidal ? Grid2D.TOROIDAL : Grid2D.BOUNDED, true, xPos, yPos);
*
* @deprecated
*/
public void getNeighborsMaxDistance( final int x, final int y, final int dist, final boolean toroidal, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors of a location that satisfy max( abs(x-X) , abs(y-Y) ) <= dist. This region forms a
* square 2*dist+1 cells across, centered at (X,Y). If dist==1, this
* is equivalent to the so-called "Moore Neighborhood" (the eight neighbors surrounding (X,Y)), plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may be run in one of three modes: Grid2D.BOUNDED, Grid2D.UNBOUNDED, and Grid2D.TOROIDAL. If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. If "unbounded", then the neighbors are not so restricted. Note that unbounded
* neighborhood lookup only makes sense if your grid allows locations to actually be outside this box. For example,
* SparseGrid2D permits this but ObjectGrid2D and DoubleGrid2D and IntGrid2D and DenseGrid2D do not. Finally if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
You can also opt to include the origin -- that is, the (x,y) point at the center of the neighborhood -- in the neighborhood results.
*/
public void getMooreLocations( final int x, final int y, final int dist, int mode, boolean includeOrigin, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors of a location that satisfy abs(x-X) + abs(y-Y) <= dist. This region forms a diamond
* 2*dist+1 cells from point to opposite point inclusive, centered at (X,Y). If dist==1 this is
* equivalent to the so-called "Von-Neumann Neighborhood" (the four neighbors above, below, left, and right of (X,Y)),
* plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may only run in two modes: toroidal or bounded. Unbounded lookup is not permitted, and so
* this function is deprecated: instead you should use the other version of this function which has more functionality.
* If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
The origin -- that is, the (x,y) point at the center of the neighborhood -- is always included in the results.
*
*
This function is equivalent to: getNeighborsHamiltonianDistance(x,y,dist,toroidal ? Grid2D.TOROIDAL : Grid2D.BOUNDED, true, xPos, yPos);
*
* @deprecated
*/
public void getNeighborsHamiltonianDistance( final int x, final int y, final int dist, final boolean toroidal, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors of a location that satisfy abs(x-X) + abs(y-Y) <= dist. This region forms a diamond
* 2*dist+1 cells from point to opposite point inclusive, centered at (X,Y). If dist==1 this is
* equivalent to the so-called "Von-Neumann Neighborhood" (the four neighbors above, below, left, and right of (X,Y)),
* plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may be run in one of three modes: Grid2D.BOUNDED, Grid2D.UNBOUNDED, and Grid2D.TOROIDAL. If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. If "unbounded", then the neighbors are not so restricted. Note that unbounded
* neighborhood lookup only makes sense if your grid allows locations to actually be outside this box. For example,
* SparseGrid2D permits this but ObjectGrid2D and DoubleGrid2D and IntGrid2D and DenseGrid2D do not. Finally if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
You can also opt to include the origin -- that is, the (x,y) point at the center of the neighborhood -- in the neighborhood results.
*/
public void getVonNeumannLocations( final int x, final int y, final int dist, int mode, boolean includeOrigin, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors located within the hexagon centered at (X,Y) and 2*dist+1 cells from point to opposite point
* inclusive.
* If dist==1, this is equivalent to the six neighbors immediately surrounding (X,Y),
* plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may only run in two modes: toroidal or bounded. Unbounded lookup is not permitted, and so
* this function is deprecated: instead you should use the other version of this function which has more functionality.
* If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
The origin -- that is, the (x,y) point at the center of the neighborhood -- is always included in the results.
*
*
This function is equivalent to: getNeighborsHexagonalDistance(x,y,dist,toroidal ? Grid2D.TOROIDAL : Grid2D.BOUNDED, true, xPos, yPos);
*
* @deprecated
*/
public void getNeighborsHexagonalDistance( final int x, final int y, final int dist, final boolean toroidal, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors located within the hexagon centered at (X,Y) and 2*dist+1 cells from point to opposite point
* inclusive.
* If dist==1, this is equivalent to the six neighbors immediately surrounding (X,Y),
* plus (X,Y) itself.
* Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may be run in one of three modes: Grid2D.BOUNDED, Grid2D.UNBOUNDED, and Grid2D.TOROIDAL. If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. If "unbounded", then the neighbors are not so restricted. Note that unbounded
* neighborhood lookup only makes sense if your grid allows locations to actually be outside this box. For example,
* SparseGrid2D permits this but ObjectGrid2D and DoubleGrid2D and IntGrid2D and DenseGrid2D do not. Finally if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
You can also opt to include the origin -- that is, the (x,y) point at the center of the neighborhood -- in the neighborhood results.
*/
public void getHexagonalLocations( final int x, final int y, final int dist, int mode, boolean includeOrigin, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors overlapping with a circular region centered at (X,Y) and with a radius of dist.
* The measurement rule is Grid2D.ANY, meaning those cells which overlap at all with the region.
* The region is closed, meaning that that points which touch on the outer surface of the circle will be
* considered members of the region.
*
*
Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may be run in one of three modes: Grid2D.BOUNDED, Grid2D.UNBOUNDED, and Grid2D.TOROIDAL. If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. If "unbounded", then the neighbors are not so restricted. Note that unbounded
* neighborhood lookup only makes sense if your grid allows locations to actually be outside this box. For example,
* SparseGrid2D permits this but ObjectGrid2D and DoubleGrid2D and IntGrid2D and DenseGrid2D do not. Finally if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
You can also opt to include the origin -- that is, the (x,y) point at the center of the neighborhood -- in the neighborhood results.
*/
public void getRadialLocations( final int x, final int y, final double dist, int mode, boolean includeOrigin, IntBag xPos, IntBag yPos );
/**
* Gets all neighbors overlapping with a circular region centered at (X,Y) and with a radius of dist.
* If measurementRule is Grid2D.CENTER, then the measurement rule will be those cells whose centers
* overlap with the region. If measurementRule is Grid2D.ALL, then the measurement rule will be those
* cells which entirely overlap with the region. If measurementrule is Grid2D.ANY, then the measurement
* rule will be those cells which overlap at all with the region. If closed is true, then the region will
* be considered "closed", that is, that points which touch on the outer surface of the circle will be
* considered members of the region. If closed is open, then the region will be considered "open", that is,
* that points which touch on the outer surface of the circle will NOT be considered members of the region.
*
*
Places each x and y value of these locations in the provided IntBags xPos and yPos, clearing the bags first.
*
*
This function may be run in one of three modes: Grid2D.BOUNDED, Grid2D.UNBOUNDED, and Grid2D.TOROIDAL. If "bounded",
* then the neighbors are restricted to be only those which lie within the box ranging from (0,0) to (width, height),
* that is, the width and height of the grid. If "unbounded", then the neighbors are not so restricted. Note that unbounded
* neighborhood lookup only makes sense if your grid allows locations to actually be outside this box. For example,
* SparseGrid2D permits this but ObjectGrid2D and DoubleGrid2D and IntGrid2D and DenseGrid2D do not. Finally if "toroidal",
* then the environment is assumed to be toroidal, that is, wrap-around, and neighbors are computed in this fashion. Toroidal
* locations will not appear multiple times: specifically, if the neighborhood distance is so large that it wraps completely around
* the width or height of the box, neighbors will not be counted multiple times. Note that to ensure this, subclasses may need to
* resort to expensive duplicate removal, so it's not suggested you use so unreasonably large distances.
*
*
You can also opt to include the origin -- that is, the (x,y) point at the center of the neighborhood -- in the neighborhood results.
*/
public void getRadialLocations( final int x, final int y, final double dist, int mode, boolean includeOrigin, int measurementRule, boolean closed, IntBag xPos, IntBag yPos );
/** Pass this into buildMap to indicate that it should make a map of any size it likes. */
public static final int ANY_SIZE = 0;
/** Creates a Map which is a copy of another. By default, HashMap is used. */
public Map buildMap(Map other);
/** Creates a map of the provided size (or any size it likes if ANY_SIZE is passed in). By default, HashMap is used. */
public Map buildMap(int size);
}