What is the best way to add cell neighbors - java

I have a grid, grid is and 2-demension array of Cell object.
public class Cell
{
int x;
int y;
ArrayList<Cell> nighbors=new ArrayList<Cell>();
public void addNeighbor(Cell cell)
{
this.neighbors.add(cell);
}
}
Every cell have 8 neighbors:
And also, one more, field is looped, like on the picture below:
So the neighbors for Cell(0,1) are also cells (5,0), (5,1), (5,2).
Now I fill neigbors like that:
public void addNeigbors(int x, int y)
{
Cell curentCell=grid[x][y];
if(x==0)
{
if(y==0)
{
curentCell.addNiegbor(this.cells[this.width-1][this.height-1]);
curentCell.addNiegbor(this.cells[x][this.height-1]);
curentCell.addNiegbor(this.cells[x+1][this.height-1]);
curentCell.addNiegbor(this.cells[x+1][y]);
curentCell.addNiegbor(this.cells[x+1][y+1]);
curentCell.addNiegbor(this.cells[x][y+1]);
curentCell.addNiegbor(this.cells[this.width-1][y+1]);
curentCell.addNiegbor(this.cells[this.width-1][y]);
}
else if(y==this.height-1)
{
// similar code
}
else
{
// and so on
}
}
// and so on
}
This code make my cry but I have no idea to make it better.
What can you advise me?

Storing references to each neighbor in Cell is a waste. IF cells need to access their neighbors, then put a reference to the grid array in each cell, and let cell calculate its neighbor indexes on the fly when necessary.
You could add a method like this:
Cell getNeighbor(int dx, int dy)
{
int w = grid.length;
int h = grid[x].length;
return grid[(x+w+dx)%w][(y+h+dy)%h];
}
If a cell needs to iterate through all of its neighbors, you can do it like this:
for (int dy=-1;dy<=1;++dy) {
for(int dx=-1;dx<=1;++dx) {
if (dx!=0 || dy!=0) {
processNeighbor(getNeighbor(dx,dy));
}
}
}

Something like this :
public void addNeigbors(int x, int y)
{
Cell curentCell=grid[x][y];
int xp1 = (x+1)%width;
int xm1 = (x-1+width)%width;
int yp1 = (y+1)%height;
int ym1 = (y-1+height)%height;
curentCell.addNiegbor(this.cells[xp1][y]);
curentCell.addNiegbor(this.cells[xp1][yp1]);
curentCell.addNiegbor(this.cells[xp1][ym1]);
curentCell.addNiegbor(this.cells[x][yp1]);
curentCell.addNiegbor(this.cells[x][ym1]);
curentCell.addNiegbor(this.cells[xm1][y]);
curentCell.addNiegbor(this.cells[xm1][yp1]);
curentCell.addNiegbor(this.cells[xm1][ym1]);
}

Create an (static) array that conains the x/y delta for all 8 neighbors so you can loop through it an add the delta to your cells x/y coordinates to get the neighbor coordinates and use mod -> % to deal with the borders:
public class Cell {
int x;
int y;
private static int delta[][] = {{-1,0},{-1,-1},{0,-1},{1,-1},{1,0},{1,1},{0,1},{-1,1}};
ArrayList < Cell > neighbors = new ArrayList < Cell > ();
public void addNeighbors() {
for (int i = 0; i < delta.length; i++) {
this.neighbors.add(this.cells
[Math.floorMod(this.x + delta[i][0], width)]
[Math.floorMod(this.y + delta[i][1], height)]
);
}
}
}

Loop over possible variations on x (dx) and y (dy), avoid the special case dx = dy = 0 and use mod to wrap around the grid.
public void addNeigbors(int x, int y) {
for(int dx = -1; dx <= 1; dx++) {
for(int dy = -1; dy <= 1; dy++) {
if(dx != 0 || dy != 0) {
int nx = Math.floorMod(x + dx, this.width);
int ny = Math.floorMod(y + dy, this.height);
this.cells[x][y].addNeighbors(this.cells[nx][ny]);
}
}
}
}

Related

How to convert the mouse position in pixels into the row and column on the grid?

I am basically making a battleship guessing game where you have to the position of a ship by the click of your mouse. When a position of the ship is guessed correctly it deletes that ship cell from the array and when every cell is guessed correctly, the game is over.
What I am now struggling on is to
keep the ship cells within the canvas
convert the mouse position in pixels into the row and column on the grid
if the guess is correct, add the guess to the hit array and if missed adding it to the miss array.
when a guess is made, in addition to colouring the cell, print either “Hit!” or “Miss!” on the cell
sinking the ship when all cells have been hit
In your code you've mixed rows and columns. The x coordinate goes from the left to the right, this are the columns. The y axis goes from the top to the bottom and corresponds to the rows.
Don't store column, row, hit and miss in arrays. But use 2-dimensional arrays to store the position of the ship and the positions of mouse clicks:
boolean [][] ship;
boolean [][] click;
keep the ship cells within the canvas
If the direction is horizontal, then the x start position of the ship has to be less than NUM_COLS - shipLength:
randomX = (int)random(NUM_COLS - shipLength);
randomY = (int)random(NUM_ROWS);
If the direction is horizontal, then the y start position of the ship has to be less than NUM_ROWS - shipLength:
randomX = (int)random(NUM_COLS);
randomY = (int)random(NUM_ROWS - shipLength);
Call randomShip in setup rather than draw:
void setup() {
size(600, 500);
randomShip();
println(store);
}
void draw() {
// randomShip(); <---- delete
drawCells (row, column, shipLength, (255) );
}
Generate the random position and size of the ship in randomShip;
void randomShip () {
ship = new boolean[NUM_COLS][NUM_ROWS];
click = new boolean[NUM_COLS][NUM_ROWS];
shipLength = (int)random (3, 8);
int store = (int)random(vert, horz);
if (store >= 0) {
int randomX = (int)random(NUM_COLS - shipLength);
int randomY = (int)random(NUM_ROWS);
for (int i = 0; i < shipLength; i++ ) {
ship[randomX + i][randomY] = true;
}
} else {
int randomX = (int)random(NUM_COLS);
int randomY = (int)random(NUM_ROWS - shipLength);
for (int i = 0; i < shipLength; i++ ) {
ship[randomX][randomY+1] = true;
}
}
println(shipLength);
}
convert the mouse position in pixels into the row and column on the grid
if the guess is correct, add the guess to the hit array and if missed adding it to the miss array.
The cell which was clicked can be get by the dividing the mouse coordinates mouseX and mouseY by CELLSIZE
int cell_x = mouseX / CELLSIZE;
int cell_y = mouseY / CELLSIZE;
Store mark the clicked cells and count the hits and miss in mouseClicked:
void mouseClicked () {
int cell_x = mouseX / CELLSIZE;
int cell_y = mouseY / CELLSIZE;
if (!click[cell_x][cell_y]) {
click[cell_x][cell_y] = true;
if ( ship[cell_x][cell_y] ) {
hitCount ++;
} else {
missCount ++;
}
}
}
when a guess is made, in addition to colouring the cell, print either “Hit!” or “Miss!” on the cell
Evaluate the ship position (ship[][]) and clicked positions (click[][]) in drawCells. Draw the cells and the text dependent on the states in 2 nested loops:
void drawCells(int colour) {
for (int i = 0; i < NUM_COLS; i++) {
for (int j = 0; j < NUM_ROWS; j++) {
float x = i * CELLSIZE;
float y = j * CELLSIZE;
if (ship[i][j]) {
fill (colour);
rect(x, y, CELLSIZE, CELLSIZE);
}
if (click[i][j]) {
fill(255, 0, 0);
textSize(15);
text(ship[i][j] ? "hit" : "miss", x+10, y+30);
}
}
}
}
sinking the ship when all cells have been hit
Handle the end of the game in draw:
e.g.
void draw() {
drawCells(255);
if (hitCount == shipLength ) {
// [...]
}
}
Full code listing:
final int CELLSIZE = 50;
final int NUM_ROWS = 10;
final int NUM_COLS = 12;
int horz = (int)random(50);
int vert = (int)random(-50);
int store;
int shipLength;
boolean [][] ship;
boolean [][] click;
int hitCount = 0;
int missCount = 0;
void setup() {
size(600, 500);
randomShip();
println(store);
}
void draw() {
drawCells(255);
if (hitCount == shipLength ) {
// [...]
}
}
void drawCells(int colour) {
for (int i = 0; i < NUM_COLS; i++) {
for (int j = 0; j < NUM_ROWS; j++) {
float x = i * CELLSIZE;
float y = j * CELLSIZE;
if (ship[i][j]) {
fill (colour);
rect(x, y, CELLSIZE, CELLSIZE);
}
if (click[i][j]) {
fill(255, 0, 0);
textSize(15);
text(ship[i][j] ? "hit" : "miss", x+10, y+30);
}
}
}
}
void randomShip () {
ship = new boolean[NUM_COLS][NUM_ROWS];
click = new boolean[NUM_COLS][NUM_ROWS];
hitCount = 0;
missCount = 0;
shipLength = (int)random (3, 8);
int store = (int)random(vert, horz);
if (store >= 0) {
int randomX = (int)random(NUM_COLS - shipLength);
int randomY = (int)random(NUM_ROWS);
for (int i = 0; i < shipLength; i++ ) {
ship[randomX + i][randomY] = true;
}
} else {
int randomX = (int)random(NUM_COLS);
int randomY = (int)random(NUM_ROWS - shipLength);
for (int i = 0; i < shipLength; i++ ) {
ship[randomX][randomY+1] = true;
}
}
println(shipLength);
}
void mouseClicked () {
int cell_x = mouseX / CELLSIZE;
int cell_y = mouseY / CELLSIZE;
if (!click[cell_x][cell_y]) {
click[cell_x][cell_y] = true;
if ( ship[cell_x][cell_y] ) {
hitCount ++;
} else {
missCount ++;
}
}
}

I am having trouble understanding constructor parameter values relating to coordinate detection

package tictactoe;
import processing.core.PApplet;
public class TicTacToe extends PApplet {
int cols = 3;
int rows = 3;
int h;
int w;
public static void main(String[] args) {
PApplet.main("tictactoe.TicTacToe");
}
public void setup() {
//Organization for size of columns and rows
w = width / cols;
h = height / rows;
}
public void settings() {
size(300, 300);
}
public void draw() {
//Draw a tic-tac-toe grid
background(255);
for (int i = 0; i < rows; i++) {
for (int j = 0; j < cols; j++) {
stroke (0);
noFill();
rect(i*w, j*h, w, h);
}
}
}
public class GridSquare{
//Game state (x's and o's), variables, etc.
public float x;
public float y;
public float w;
public float h;
public int state;
public void drawTurn() {
if (state == 0) {
ellipse(x+w/2,y+h/2,w,h);
}
if (state == 1) {
line(x,y,x+w,y+h);
line(x+w,y,x,y+h);
}
}
//Trouble understanding Grid Detection
void onClick(int clickedX, int clickedY, int turn) {
if (clickedX > x && clickedX < x + w && clickedY > y && clickedY < y + h) {
}
}
}
}
I am having trouble understanding the grid detection if statement. Since the parameters x, y, w, and h all had a semi colon when being defined in the GridSquare class, wouldn't that mean that those variables all equal 0? In that case, wouldn't the if statement be checking if the click x integer is greater than 0, and clicked x is less than 0 (x) + 0 (w)? Or am I wrong, and are the variables representing integers previously defined in the code?
It might be more easily understood if you think about having multiple games (GridSquares) going on at the same time, each with a border around it, and each being drawn on the same panel. The check is simply determining if the click in the panel is valid for that game (GridSquare). With multiple games at once, each one would have a different value for x and y; imagining two games, side by side, the first would probably have (x=0, y=0), but the second game would be at least 300 units (pixels, perhaps) to the right, giving values (x=300, y=0).

Java - printing a 2D array for Conways Game of Life

I started today to program Conways Game of Life. In a first step, I just want the user to input the length of the (squadratic) field which is then displayed on the screen. But I'm getting a NullPointerException in the printGrid() method. Here are the necessary code examples:
public class Grid {
private Cell[][]grid;
public Grid (int feldlänge) {
grid = new Cell[feldlänge][feldlänge];
int x, y;
for (y = 0; y < feldlänge; y = y + 1) {
for (x = 0; x < feldlänge; x = x + 1) {
Cell cell;
cell = new Cell(x,y);
cell.setLife(false);
} // for
} // for
} // Konstruktor Grid
public String printGrid () {
String ausgabe = "";
int x, y;
for (y = 0; y < grid.length; y = y + 1) {
for (x = 0; x < grid.length; x = x + 1) {
if (grid[x][y].isAlive()) { // Here's the NullPointerException
ausgabe = ausgabe + "■";
}
if (!grid[x][y].isAlive()) {
ausgabe = ausgabe + "□";
}
}
ausgabe = ausgabe + "\n";
}
return ausgabe;
}
public class Cell {
private int x, y;
private boolean isAlive;
public Cell (int pX, int pY) {
x = pX;
y = pY;
} // Konstruktor Cell
public void setLife (boolean pLife) {
isAlive = pLife;
} // Methode setLife
public int getX () {
return x;
} // Methode getX
public int getY () {
return y;
} // Methode getY
public boolean isAlive () {
return isAlive;
}
}
It's kind of embarrassing I can't find the mistake by myself. I guess I'm overlooking something simple.
Already thanks a lot for any help!
Update: Already solved!
I just didn't add the cell to the array. It works now.
You don't seem to add the cell into your grid array.
public Grid (int feldlänge) {
grid = new Cell[feldlänge][feldlänge];
int x, y;
for (y = 0; y < feldlänge; y = y + 1) {
for (x = 0; x < feldlänge; x = x + 1) {
Cell cell;
cell = new Cell(x,y);
cell.setLife(false);
grid[x][y] = cell; //put the cell in the grid.
} // for
} // for
} // Konstruktor Grid
You have to add cell to your array. (german field = english array)
Also: instead of
if( someBoolean){}
if( !someBoolean){}
you should use
if( someBoolean){}
else {}
This makes it more clear what the code does

I need some help understanding arrays of pixels

I'm using Processing to divide a large image into a series of smaller, rectangular nodes.
Processing stores the color value for the pixels of a PImage in a pixels array, which I am accessing to break up the image into smaller parts. For some reason, I am getting this output, when my intent was for the entire image to be displayed when the nodes are arranged in draw().
Here is my main class:
ArrayList node = new ArrayList();
PImage grid;
PVector nodeDimensions = new PVector(210, 185);
PVector gridDimensions = new PVector(2549, 3300);
String name = "gridscan.jpeg";
void setup() {
size(500, 500);
grid = loadImage(name);
grid.loadPixels();
fillPixels();
noLoop();
}
void fillPixels() {
int nodeNum = 0;
for (int startX = 0; startX < 2549 - nodeDimensions.x; startX += nodeDimensions.x) {
for (int startY = 0; startY < 3300 - nodeDimensions.y; startY += nodeDimensions.y) {
node.add(new Node());
sendPixels(new PVector(startX, startY), nodeNum);
nodeNum++;
}
}
}
void sendPixels(PVector start, int nodeNum) {
for (int x = int(start.x); x < start.x + nodeDimensions.x; x++) {
for (int y = int(start.y); y < start.x + nodeDimensions.y; y++) {
Node _node = (Node) node.get(node.size() - 1);
_node.fillPixel(new PVector(x, y), grid.pixels[int(y*gridDimensions.x+x)]);
}
}
}
void draw() {
drawNodes();
}
void drawNodes() {
int nodeNum = 0;
for (int x = 0; x < width; x += nodeDimensions.x) {
for (int y = 0; y < height; y += nodeDimensions.y) {
Node _node = (Node) node.get(nodeNum);
_node.drawMe(new PVector(x - (nodeDimensions.x/2), y - (nodeDimensions.y/2)));
nodeNum++;
}
}
}
And here is the Node class:
class Node {
color[] pixel;
Node() {
pixel = new color[int(nodeDimensions.x * nodeDimensions.y)];
}
void fillPixel(PVector pos, color pixelValue) {
if(int(pos.y * nodeDimensions.y + pos.x) < 38850) pixel[int(pos.y * nodeDimensions.y + pos.x)] = pixelValue;
}
void drawMe(PVector centerPos) {
pushMatrix();
translate(centerPos.x, centerPos.y);
for(int x = 0; x < nodeDimensions.x; x++) {
for(int y = 0; y < nodeDimensions.y; y++) {
stroke(getPixelColor(new PVector(x, y)));
point(x,y);
}
}
popMatrix();
}
color getPixelColor(PVector pos) {
return pixel[int(pos.y * nodeDimensions.x + pos.x)];
}
}
Hopefully my code makes sense. I suspect the issue is in the sendPixels() method of the main class.
I used this this page from the Processing reference as a guide for creating that function, and I'm not sure where my logic is wrong.
Any help would be appreciated, and please let me know if I can clarify something.
According to getPixelColor(), it seems that it uses rows.
So if you have a 5x5 square image then 2x2 would be 7.
To get the index you use this formula:
index = (y - 1) * width + x
Explained this way it's look pretty simple, doesn't it?
Alternatively, you may be able to use getSubimage() on the BufferedImage returned by the getImage method of PImage. There's a related example here.

How can I simplify this code? (Chess game obstruction testing)

I'm making a chess game in Java, and testing to make sure there are no pieces blocking the path of the piece being moved. The piece moves from (srcX,srcY) to (dstX,dstY).
I've written this code which checks if there are any obstructions for a rook:
if(dstY == srcY) {
// No change on Y axis, so moving east or west
if(dstX > srcX) {
// Moving east
// Test every cell the piece will pass over
for(int x = srcX+1; x < dstX; x++) {
// Is the cell set?
if(isPiece(x, srcY)) {
return true;
}
}
} else {
// Must be moving west
// Test every cell the piece will pass over
for(int x = srcX-1; x > dstX; x--) {
// Is the cell set?
if(isPiece(x, srcY)) {
return true;
}
}
}
} else if(dstX == srcX) {
// No change on X axis, so moving north or south
if(dstY > srcY) {
// Moving north
// Test every cell the piece will pass over
for(int y = srcY+1; y < dstY; y++) {
// Is the cell set?
if(isPiece(srcX, y)) {
return true;
}
}
} else {
// Must be moving south
// Test every cell the piece will pass over
for(int y = srcY-1; y > dstY; y--) {
// Is the cell set?
if(isPiece(srcX, y)) {
return true;
}
}
}
}
but it's a bit big and I'm sure it can be simplied.. any ideas?
ps, this is ONLY obstruction testing. I've already validated everything else.
Once you've tested for direction, you can set dx, dy values (e.g. dx=1, dy=0 for east). Then you can have a single for loop for all cases and just increment x and y by dx and dy respectively at each iteration.
You can then simplify the direction checking into the following:
if dstY == srcY: dy = 0
else: dy = (dstY - srcY) / abs(dstY - srcY)
if dstX == srcX: dx = 0
else: dx = (dstX - srcX) / abs(dstX - srcX)
Code:
int dx, dy;
if (dstY == srcY) dy = 0;
else dy = (dstY - srcY) / Math.abs(dstY - srcY);
if (dstX == srcX) dx = 0;
else dx = (dstX - srcX) / Math.abs(dstX - srcX);
while (srcX != dstX || srcY != dstY) {
srcX += dx; srcY += dy;
if (isPiece(srcX, srcY))
return true;
}
return false;
Also beware that this code (and yours) will fail if the move is not horizontal, vertical or diagonal.
You could do something along these lines (untested as I don't have a compiler to hand):
int dx = 0;
int dy = 0;
if (dstX != srcX) {
dx = (dstX > srcX) ? 1 : -1;
} else if (dstY != srcY) {
dy = (dstY > srcY) ? 1 : -1;
}
int x = srcX + dx;
int y = srcY + dy;
while (x != dstX || y != dstY) {
if (isPiece(x, y)) return true;
x += dx;
y += dy;
}
First, write tests. Lots and lots of tests. That way you can be confident that you're simplifying without changing the meaning of the code.
Refactoring without unit tests is like walking a high wire without a safety net.
Nearly the same, but with for loops:
// move along x axis
for (int x = 1; x < Math.abs(srcX - dstX); x++) {
int curX = (srcX - dstX) < 0 ? srcX - x : srcX + x;
if (isPiece(curX, srcY))
return true;
}
// move along y axis
for (int y = 1; y <= Math.abs(srcY - dstY); y++) {
int curY = (srcY - dstY) < 0 ? srcY - y : srcY + y;
if (isPiece(srcX, curY))
return true;
}
My soultion would be: introduce a direction class, and then do the check in this style:
isBlocked(startPossition, direction, numberOfFields)
I have done a little example, using 3 Classes.
Direction - an enum to represent the 8 directions (2 horizontal, 2 vertical, 4 diagonal)
Position - the x and y value of an position
LinarMove - represent one linear Move(startPossition, direction, numberOfFields) and contains the isBlockedMethod
The Enum:
public enum Direction {
UP(0, 1),
DOWN(0, -1),
LEFT(1, 0),
RIGHT(-1, 0),
UP_LEFT(UP, LEFT),
UP_RIGTH(UP, RIGHT),
DOWN_LEFT(DOWN, LEFT),
DOWN_RIGHT(
DOWN, RIGHT);
private final int incrementX;
private final int incrementY;
private Direction(int incrementX, int incrementY) {
this.incrementX = incrementX;
this.incrementY = incrementY;
}
private Direction(Direction sub1, Direction sub2) {
this.incrementX = sub1.incrementX + sub2.incrementX;
this.incrementY = sub1.incrementY + sub2.incrementY;
}
public Position oneField(Position start) {
return new Position(start.getX() + this.incrementX, start.getY()
+ this.incrementY);
}
}
The purpuse of second constructor is only that it alowes to write the diagonal moves in a more readable way.
public class Position {
private final int x;
private final int y;
public Position(int x, int y) {
super();
this.x = x;
this.y = y;
}
public int getX() {
return x;
}
public int getY() {
return y;
}
#Override
public String toString() {
return "x:" + x + ", y:" + y;
}
}
The Move contains the isBlocked Method -- you can see how small it get, and how readable it become. At least there is no single direction related if statement left.
The name LinareMove sugget that there is possible an other kind of move for the knight.
public class LinearMove {
private final Position start;
private final Direction direction;
/** Length of the move. */
private final int numberOfFields;
public LinearMove(Position start, Direction direction, int numberOfFields) {
super();
this.start = start;
this.direction = direction;
this.numberOfFields = numberOfFields;
}
boolean isBlocked() {
Position pos = this.start;
for (int i = 0; i < (this.numberOfFields - 1); i++) {
pos = this.direction.oneField(pos);
if (isPiece(pos)) {
return true;
}
}
return false;
}
boolean isPiece(Position pos) {
//Dummy;
System.out.println(pos);
return false;
}
}
And this is how it works:
public static void main(String[] args) {
new LinearMove(new Position(1, 1), Direction.RIGHT, 3).isBlocked();
}
You maybe noticed, that the knights move is some kind of probem. With this soultion you could model it in two ways:
- 4 special Directions
- an other kind of move class (this is the more cleaner way, because you could always return true, in the isBockedMethod)

Categories

Resources