为什么我的 Tkinter 生命游戏实施会逐渐变慢？

Question

编辑：

对于以后遇到此问题的任何人：

请检查你是否有一段代码创建对象实例，并在询问类似问题之前先删除它。

** 问题开头：**

我想用 Tkinter 实现 Conway 的生命游戏。我的代码似乎确实有效，但每次我启动该应用程序时，它都会变慢，直到 canvas 每秒只更新一次。

我已经研究过类似的问题，但这里的问题是，在每一个问题中，人们不断地创建矩形，这显然会导致帧速率问题，但我在我的实现中改变了它（我初始化 canvas ROWS*COLS 矩形，然后只改变它们的颜色）。

负责更新 canvas 的主要函数是“startConway()”函数。如您所见，我尝试测量时间 (t0,t1,t2) 以查看一个 for 循环是否导致问题，但即使 canvas 变慢，时间几乎保持不变并且始终低于 20女士

关于 Tkinter 有什么我不了解的东西阻碍了我的进步吗？还是我的方法太慢了？哪里可以改进？

如果您想尝试一下，只需复制代码，然后运行即可。在 canvas 上，点击一些像素，然后最后开始 Conway 游戏，点击一个黑色像素。

提前致谢。

代码如下：

import tkinter as tk
import numpy as np
import time
import random
import copy
import math

# Set number of rows and columns
ROWS = 10
COLS = 10
ANIMATION_DELAY = 50
NEIGHBOUR_SEARCH_DELAY = 0
BORDER_PROB = 0.2
ITERATIONS_CONWAY = 1000

# Create a grid of None to store the references to the tiles
cells = [[None for _ in range(ROWS)] for _ in range(COLS)]

class Cell:

    def __init__(self, col, row):
        self.col = col
        self.row = row
        self.rect = None
        self.text = None
        createRectangle(self,"","","")
        # extras for A_Star:
        self.cost_g = 0
        self.cost = 0
        self.parent = None
        self.dist_to_neighbours = []
    
    def equals(self,cell):
        return (self.col == cell.col and self.row == cell.row)
    
    def changeText(self,txt,color):
        c.itemconfig(self.text,text=txt,fill=color)
        
    def changeRect(self,color):
        c.itemconfig(self.rect, fill=color)
    
    def toString(self):
        return str(self.col) + "|" + str(self.row)
    
    # extras for A_Star:
    def getDist(self,cell):
        return 1
    
def callback2(event):
    # Calculate column and row number
    col = int(event.x/col_width)
    row = int(event.y/row_height)
    clicked_cell = cells[col][row]
    # If the tile is not filled, create a rectangle
    if not c.itemcget(clicked_cell.rect,"fill") == "black":
        clicked_cell.changeRect("black")
    else:
            startConway(ITERATIONS_CONWAY)

def startConway(iteration):
    
    if iteration == 0:
        return
    
    alive = np.zeros([COLS,ROWS])
    
    t0 = time.time()
    
    for row in range(ROWS):
        for col in range(COLS):
            current_cell = cells[col][row]
            neighbours = getAliveNeighbours(current_cell)
            isAlive = c.itemcget(current_cell.rect,"fill") == "black"
            
            willSurvive = isAlive and (neighbours == 2 or neighbours == 3)
            willResurrect = not isAlive and neighbours == 3
            
            if willSurvive or willResurrect:
                alive[col][row] = 1
            else:
                alive[col][row] = 0
            
    t1 = time.time()
    
    for row in range(ROWS):
        for col in range(COLS):
            current_cell = cells[col][row]
            isAlive = alive[col][row]
            
            if isAlive:
                current_cell.changeRect("black")
            else:
                current_cell.changeRect("white")
                
    t2 = time.time()
    
    root.after(ANIMATION_DELAY,startConway,iteration-1)
       
    total = t1-t0
    
    print("neighbour Loop: ", t1-t0)
    print("draw Loop: ", t2-t1)
    print("----------------------")
    
def getAliveNeighbours(coord):
    count = 0
    
    upper_left = Cell(coord.col - 1,coord.row - 1) 
    neighbours = []
    for row in range(3):
        for col in range(3):
            new_col = (upper_left.col + col) % COLS
            new_row = (upper_left.row + row) % ROWS
            
            isStartingCell = (row == 1 and col == 1)
            isAlive = c.itemcget(cells[new_col][new_row].rect,"fill") == "black"
            
            if not isStartingCell and isAlive:
                count = count + 1
                
    return count

def createRectangle(cell,rect_color,txt,text_color):
        cell.rect = c.create_rectangle(cell.col*col_width, cell.row*row_height, (cell.col+1)*col_width, (cell.row+1)*row_height, fill=rect_color)
        cell.text = c.create_text(((cell.col*col_width + (cell.col+1)*col_width)/2, (cell.row*row_height + (cell.row+1)*row_height)/2), text=txt, fill=text_color)

def initializeCanvas(hasBorder):
    for row in range(ROWS):
        for col in range(COLS):
            cells[col][row] = Cell(col,row)
            if hasBorder and (row == 0 or col == 0 or col == COLS-1 or row == ROWS-1):
                cells[col][row].changeRect("black")
                
root = tk.Tk()

c = tk.Canvas(root, width=500, height=500, borderwidth=5, background='white')
c.pack()
c.bind("<Button-1>", callback2)
c.update()

col_width = c.winfo_width()/COLS
row_height = c.winfo_height()/ROWS

initializeCanvas(False)
            
root.mainloop()

Answer 1

每次调用 getAliveNeighbors 都会创建一个新的 Cell 实例。这会在 canvas 上创建一个新矩形。您每 50 毫秒调用此函数 100 次（10 行乘以 10 列）。因此，您每秒都在创建 2,000 个新矩形。

一旦您在 canvas 上获得数万个对象，canvas 就会出现已知的性能问题，这种情况会以每秒 2,000 个矩形的速度发生。

为什么我的 Tkinter 生命游戏实施会逐渐变慢？

Why does my Tkinter Game of Life Implementation slow down progressively?

python

tkinter

python-3.x

tkinter-canvas