sim.field.grid
Interface Grid2D

All Superinterfaces:
java.io.Serializable
All Known Implementing Classes:
AbstractGrid2D, DenseGrid2D, DoubleGrid2D, IntGrid2D, ObjectGrid2D, SparseGrid2D

public interface Grid2D
extends java.io.Serializable

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. :-)


Field Summary
static int ALL
          "All" measurement rule for raidal neighborhood lookup.
static int ANY
          "Any" measurement rule for raidal neighborhood lookup.
static int ANY_SIZE
          Pass this into buildMap to indicate that it should make a map of any size it likes.
static int BOUNDED
          Bounded Mode for neighborhood lookup.
static int CENTER
          Center measurement rule for raidal neighborhood lookup.
static int TOROIDAL
          Bounded Mode for toroidal lookup.
static int UNBOUNDED
          Bounded Mode for neighborhood lookup.
 
Method Summary
 java.util.Map buildMap(int size)
          Creates a map of the provided size (or any size it likes if ANY_SIZE is passed in).
 java.util.Map buildMap(java.util.Map other)
          Creates a Map which is a copy of another.
 int dlx(int x, int y)
          Hex downleft x.
 int dly(int x, int y)
          Hex downleft y.
 int downx(int x, int y)
          Hex down x.
 int downy(int x, int y)
          Hex down y.
 int drx(int x, int y)
          Hex downright x.
 int dry(int x, int y)
          Hex downright y.
 int getHeight()
          Returns the width of the field.
 void getHexagonalLocations(int x, int y, 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.
 void getMooreLocations(int x, int y, int dist, int mode, boolean includeOrigin, IntBag xPos, IntBag yPos)
          Gets all neighbors of a location that satisfy max( abs(x-X) , abs(y-Y) ) <= dist.
 void getNeighborsHamiltonianDistance(int x, int y, int dist, boolean toroidal, IntBag xPos, IntBag yPos)
          Deprecated.  
 void getNeighborsHexagonalDistance(int x, int y, int dist, boolean toroidal, IntBag xPos, IntBag yPos)
          Deprecated.  
 void getNeighborsMaxDistance(int x, int y, int dist, boolean toroidal, IntBag xPos, IntBag yPos)
          Deprecated.  
 void getRadialLocations(int x, int y, 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.
 void getRadialLocations(int x, int y, double dist, int mode, boolean includeOrigin, int measurementRule, boolean closed, IntBag xPos, IntBag yPos)
          Gets all neighbors overlapping with a circular region centered at (X,Y) and with a radius of dist.
 void getVonNeumannLocations(int x, int y, 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.
 int getWidth()
          Returns the width of the field.
 int stx(int x)
          Simple [and fast] toroidal x.
 int sty(int y)
          Simple [and fast] toroidal y.
 boolean trb(int x, int y)
          Horizontal edge is on the bottom for triangle.
 boolean trt(int x, int y)
          Horizontal edge is on the top for triangle.
 int tx(int x)
          Toroidal x.
 int ty(int y)
          Toroidal y.
 int ulx(int x, int y)
          Hex upleft x.
 int uly(int x, int y)
          Hex upleft y.
 int upx(int x, int y)
          Hex up x.
 int upy(int x, int y)
          Hex up y.
 int urx(int x, int y)
          Hex upright x.
 int ury(int x, int y)
          Hex upright y.
 

Field Detail

BOUNDED

static final int BOUNDED
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.

See Also:
Constant Field Values

UNBOUNDED

static final int UNBOUNDED
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.

See Also:
Constant Field Values

TOROIDAL

static final int TOROIDAL
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).

See Also:
Constant Field Values

CENTER

static final int CENTER
Center measurement rule for raidal neighborhood lookup. Indicates that radial lookup will include locations whose grid cell centers overlap with the neighborhood region.

See Also:
Constant Field Values

ALL

static final int ALL
"All" measurement rule for raidal neighborhood lookup. Indicates that radial lookup will include locations whose grid cells are entirely within the neighborhood region.

See Also:
Constant Field Values

ANY

static final int ANY
"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.

See Also:
Constant Field Values

ANY_SIZE

static final int ANY_SIZE
Pass this into buildMap to indicate that it should make a map of any size it likes.

See Also:
Constant Field Values
Method Detail

getWidth

int getWidth()
Returns the width of the field.


getHeight

int getHeight()
Returns the width of the field.


tx

int tx(int x)
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.


ty

int ty(int y)
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.


stx

int stx(int x)
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.


sty

int sty(int y)
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.


ulx

int ulx(int x,
        int y)
Hex upleft x.


uly

int uly(int x,
        int y)
Hex upleft y.


urx

int urx(int x,
        int y)
Hex upright x.


ury

int ury(int x,
        int y)
Hex upright y.


dlx

int dlx(int x,
        int y)
Hex downleft x.


dly

int dly(int x,
        int y)
Hex downleft y.


drx

int drx(int x,
        int y)
Hex downright x.


dry

int dry(int x,
        int y)
Hex downright y.


upx

int upx(int x,
        int y)
Hex up x.


upy

int upy(int x,
        int y)
Hex up y.


downx

int downx(int x,
          int y)
Hex down x.


downy

int downy(int x,
          int y)
Hex down y.


trb

boolean trb(int x,
            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;


trt

boolean trt(int x,
            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;


getNeighborsMaxDistance

void getNeighborsMaxDistance(int x,
                             int y,
                             int dist,
                             boolean toroidal,
                             IntBag xPos,
                             IntBag yPos)
Deprecated. 

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);


getMooreLocations

void getMooreLocations(int x,
                       int y,
                       int dist,
                       int mode,
                       boolean includeOrigin,
                       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.


getNeighborsHamiltonianDistance

void getNeighborsHamiltonianDistance(int x,
                                     int y,
                                     int dist,
                                     boolean toroidal,
                                     IntBag xPos,
                                     IntBag yPos)
Deprecated. 

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);


getVonNeumannLocations

void getVonNeumannLocations(int x,
                            int y,
                            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 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.


getNeighborsHexagonalDistance

void getNeighborsHexagonalDistance(int x,
                                   int y,
                                   int dist,
                                   boolean toroidal,
                                   IntBag xPos,
                                   IntBag yPos)
Deprecated. 

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);


getHexagonalLocations

void getHexagonalLocations(int x,
                           int y,
                           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 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.


getRadialLocations

void getRadialLocations(int x,
                        int y,
                        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. 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.


getRadialLocations

void getRadialLocations(int x,
                        int y,
                        double dist,
                        int mode,
                        boolean includeOrigin,
                        int measurementRule,
                        boolean closed,
                        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.


buildMap

java.util.Map buildMap(java.util.Map other)
Creates a Map which is a copy of another. By default, HashMap is used.


buildMap

java.util.Map buildMap(int size)
Creates a map of the provided size (or any size it likes if ANY_SIZE is passed in). By default, HashMap is used.