如何在 tkinter 旁边运行循环？ - Python 二维码阅读器图形用户界面
Posted
技术标签:
【中文标题】如何在 tkinter 旁边运行循环？ - Python 二维码阅读器图形用户界面【英文标题】：How can you run loops alongside tkinter? - Python QR code reader GUI 【发布时间】：2021-10-27 19:00:39 【问题描述】：我正在尝试编写一个也可以处理二维码的 Tkinter 应用程序。为此，我需要循环检查 QR 码是否有效并且我需要发出发布请求。我完全意识到我编码的方式非常低效。有一个更好的方法吗？这是我目前所拥有的：
import cv2
import tkinter as tk
from PIL import Image, ImageTk
import sys
import os
import pyzbar.pyzbar as zbar
import threading
import requests
import queue
result = []
decodedCode = ""
logo = "logo.png"
settings = "settings.png"
if os.environ.get('DISPLAY','') == '':
print('no display found. Using :0.0')
os.environ.__setitem__('DISPLAY', ':0.0')
#create main window
master = tk.Tk()
master.title("tester")
master.geometry("480x800")
master.configure(bg='white')
ttelogo = tk.PhotoImage(file = logo)
settingslogo = tk.PhotoImage(file = settings)
#settings button
settings_frame = tk.Frame(master,width=50,height=50,bg="white")
settings_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
settingsBtn = tk.Button(settings_frame, image=settingslogo).pack()
settings_frame.place(x=430,y=0)
#logo
img = tk.Label(master, image=ttelogo, bg='white')
img.image = ttelogo
img.place(x=176.5,y=10)
#Name Label
label_frame = tk.Frame(master,width=400,height=100,bg="white")
label_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
tk.Label(label_frame,bg="white",fg="black",text="John Smith Smithington III",font=("Calibri",22)).pack()
label_frame.place(x=40,y=140)
#Instructions Label
instructions_frame = tk.Frame(master,width=440,height=100,bg="white")
instructions_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
tk.Label(instructions_frame,bg="white",fg="black",text="Place your pass under the scanner below.",font=("Calibri",10)).pack()
instructions_frame.place(x=20,y=210)
#Camera Window
cameraFrame = tk.Frame(master, width=440, height=480)
cameraFrame.place(x=20, y=260)
#Camera Feed
lmain = tk.Label(cameraFrame)
lmain.place(x=0, y=0)
cap = cv2.VideoCapture(0)
def startScanning():
global cap
_, frame = cap.read()
frame = cv2.flip(frame, 1)
cv2image = cv2.cvtColor(frame, cv2.COLOR_BGR2RGBA)
img = Image.fromarray(cv2image)
imgtk = ImageTk.PhotoImage(image=img)
lmain.imgtk = imgtk
lmain.configure(image=imgtk)
lmain.after(10, startScanning)
def processScan():
global decodedCode
stopped = False
delay = 1
while(True):
ret = cv2.waitKey(delay) & 0xFF
if ret == ord('c'): # continue
stopped = False
delay = 1
if ret == ord('q'):
break
if stopped or (ret == ord('s')): # stop
stopped = True
delay = 30
continue
# Capture frame-by-frame
ret, frame = cap.read()
decodedObjects = zbar.decode(frame)
if len(decodedObjects) > 0:
stopped = True
for code in decodedObjects:
#print("Data", obj.data)
#API Calls
decodedCode = code.data.decode('utf-8')
# When everything done, release the capture
cap.release()
cv2.destroyAllWindows()
def checkCode():
global decodedCode
while True:
if decodedCode != "":
print (decodedCode)
result = requests.post("https://example.com/login/index.php", data='action': 'validate_scan', 'uuid': decodedCode).text
print(result)
decodedCode = ""
startScanning() #Display 2
threading.Thread(name='background', target=processScan).start()
threading.Thread(name='background2', target=checkCode).start()
master.mainloop() #Starts GUI
编辑：队列形式的新版本：
# import all the necessary modules
from tkinter import *
from multiprocessing import Process, Queue
from queue import Empty # for excepting a specific error
import numpy as np
import cv2
from PIL import Image, ImageTk
import sys
import os
import pyzbar.pyzbar as zbar
import threading
import requests
# this is the function that will be run in a child process
def processScan(queue_): # pass the queue as an argument
stopped = False
delay = 1
while(True):
ret = cv2.waitKey(delay) & 0xFF
if ret == ord('c'): # continue
stopped = False
delay = 1
if ret == ord('q'):
break
if stopped or (ret == ord('s')): # stop
stopped = True
delay = 30
continue
# Capture frame-by-frame
ret, frame = cap.read()
decodedObjects = zbar.decode(frame)
if len(decodedObjects) > 0:
stopped = True
for code in decodedObjects:
#print("Data", obj.data)
#API Calls
queue_.put(code.data.decode('utf-8'))
# When everything done, release the capture
cap.release()
cv2.destroyAllWindows()
#master.after(2000, processScan)
#return r
# just a function to not write a `lambda`, just easier
# to read code
def startScanning():
global cap
_, frame = cap.read()
frame = cv2.flip(frame, 1)
cv2image = cv2.cvtColor(frame, cv2.COLOR_BGR2RGBA)
img = Image.fromarray(cv2image)
imgtk = ImageTk.PhotoImage(image=img)
lmain.imgtk = imgtk
lmain.configure(image=imgtk)
lmain.after(10, startScanning)
#processScan()
#threading.Thread(name='background', target=processScan).start()
# set process `daemon = True` so that it gets terminated with the
# main process, this approach is not suggested if you write to file
# but otherwise it shouldn't cause any issues (maybe an error but
# that probably can be handled with `try/except`)
Process(target=processScan, args=(queue, ), daemon=True).start()
# here is the loop for updating the label
# basically get the info from the queue
def update_label():
try:
# try getting data but since it is `block=False`
# if there is nothing in the queue it will not block
# this process waiting for data to appear in the queue
# but it will raise the Empty error
data = queue.get(block=False)
except Empty:
pass
else:
# if no error was raised just config
# label to new data
labelFrame.config(text=data)
finally:
# and simply schedule this function to run again in
# 100 milliseconds (this btw is not recursion)
master.after(100, update_label)
# crucial part is to use this if statement because
# child processes run this whole script again
if __name__ == '__main__':
master = Tk()
logo = "logo.png"
settings = "settings.png"
if os.environ.get('DISPLAY','') == '':
print('no display found. Using :0.0')
os.environ.__setitem__('DISPLAY', ':0.0')
#create main window
master.attributes("-fullscreen", True)
master.title("Title")
master.geometry("480x800")
master.configure(bg='white')
ttelogo = PhotoImage(file = logo)
settingslogo = PhotoImage(file = settings)
#settings button
settings_frame = Frame(master,width=50,height=50,bg="white")
settings_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
settingsBtn = Button(settings_frame, image=settingslogo).pack()
settings_frame.place(x=430,y=0)
#logo
img = Label(master, image=ttelogo, bg='white')
img.image = ttelogo
img.place(x=176.5,y=10)
#Name Label
label_frame = Frame(master,width=400,height=100,bg="white")
label_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
Label(label_frame,bg="white",fg="black",text="John Smith Smithington III",font=("Calibri",22)).pack()
label_frame.place(x=40,y=140)
#Instructions Label
instructions_frame = Frame(master,width=440,height=100,bg="white")
instructions_frame.pack_propagate(0) # Stops child widgets of label_frame from resizing it
Label(instructions_frame,bg="white",fg="black",text="Place your pass under the scanner below.",font=("Calibri",15)).pack()
instructions_frame.place(x=20,y=210)
#Camera Window
cameraFrame = Frame(master, width=440, height=480)
cameraFrame.place(x=20, y=260)
#Camera Feed
lmain = Label(cameraFrame)
lmain.place(x=0, y=0)
cap = cv2.VideoCapture(0)
# define queue (since it is a global variable now
# it can be easily used in the functions)
queue = Queue()
#label = Label(root)
#label.pack()
# initially start the update function
update_label()
# just a button for starting the process, but you can also simply
# call the function
#Button(root, text='Start', command=start_process).pack()
startScanning()
master.mainloop()
仍然遇到错误。也不确定这是否是正确的队列语法。相机源不是实时的。目前只显示一张静态图片。
【问题讨论】：
一种优化是使用queue 将代码从processScan 发送到checkCode。这样，checkCode 可以在无事可做时阻塞，而不是在紧密循环中占用 100% 的 CPU。
您可以提供@TimRoberts 示例的任何更改吗？我之前尝试过排队，但运气不佳。谢谢！
我还注意到，在运行这两个线程时，CPU 几乎处于 200%……目前效率非常低。我只是不熟悉如何与 tkinter 一起运行。我尝试使用“master.after()”方法，但这也不起作用。
对于 cpu 繁重的任务我听说multiprocessing 更好，因为它跨越多个内核的进程，但是tkinter 本身不能在任何子进程上运行，只能在计算部分甚至那么只有可以腌制的对象，也适用于队列，那么你应该使用multiprocessing.Queue而不是queue.Queue
@Matiiss 你能给我举个例子吗？我之前尝试过使用队列，但运气不佳。
【参考方案1】：
这和multiprocessing 一样简单（代码 cmets 中的解释（这就是为什么如果删除 cmets 代码可能看起来有很多事情发生，它大约只有 40 行代码）对于这个例子））：
# import all the necessary modules
from tkinter import Tk, Label, Button
from multiprocessing import Process, Queue
from queue import Empty # for excepting a specific error
# this is the function that will be run in a child process
def count(queue_): # pass the queue as an argument
# these modules are just to show the functionality
# but time might be necessary otherwise
# if theses are imported in the global namespace
# then you don't need to import them here again
import itertools
import time
for i in itertools.count():
# do the main looping
# put data in the queue
queue_.put(i)
# you may not need to use sleep here depending
# on how long it takes for the
# loop to finish the iteration
# the issue is that the loop may finish
# way faster than the update function runs
# and that will fill up the queue
# I think the queue size can be limited
# but putting sleep like this to allow
# the update loop to update will work too
time.sleep(0.2)
# just a function to not write a `lambda`, just easier
# to read code
def start_process():
# set process `daemon = True` so that it gets terminated with the
# main process, this approach is not suggested if you write to file
# but otherwise it shouldn't cause any issues (maybe an error but
# that probably can be handled with `try/except`)
Process(target=count, args=(queue, ), daemon=True).start()
# here is the loop for updating the label
# basically get the info from the queue
def update_label():
try:
# try getting data but since it is `block=False`
# if there is nothing in the queue it will not block
# this process waiting for data to appear in the queue
# but it will raise the Empty error
data = queue.get(block=False)
except Empty:
pass
else:
# if no error was raised just config
# label to new data
label.config(text=data)
finally:
# and simply schedule this function to run again in
# 100 milliseconds (this btw is not recursion)
root.after(100, update_label)
# crucial part is to use this if statement because
# child processes run this whole script again
if __name__ == '__main__':
root = Tk()
# define queue (since it is a global variable now
# it can be easily used in the functions)
queue = Queue()
label = Label(root)
label.pack()
# initially start the update function
update_label()
# just a button for starting the process, but you can also simply
# call the function
Button(root, text='Start', command=start_process).pack()
root.mainloop()
我对@987654323@ 不太熟悉，所以它可能在子进程中不起作用（我想我已经这样做了），在这种情况下，您将不得不使用threading 或使用subprocess 的两种替代方法：将opencv 部分放在另一个文件中，并使用subprocess.Popen 中的python 解释器运行该文件（不适用于未安装Python 的计算机）或转换其他python将带有opencv 部分的文件添加到.exe 文件或某事并在subprocess.Popen 中运行该可执行文件（这两种方法都可能是不必要的，您可能会使用multiprocessing 或threading）
要改用threading，请更改这些行：
# from multiprocessing import Process, Queue
from threading import Thread
from queue import Empty, Queue
和
# Process(target=count, args=(queue, ), daemon=True).start()
Thread(target=count, args=(queue, ), daemon=True).start()
【讨论】：
多线程在这里应该可以正常工作。不需要多处理。应避免显式子流程。 @ChristophRackwitz OP 说 CPU 使用率高，multiprocessing 应该限制的东西，不是吗？无论如何，如果使用队列，那么唯一的区别是使用来自queue 模块的队列并将Process 替换为来自threading 库的Thread，如果你想使用Thread，基本上改变为3 行。否则，除了Process 会使用自己的资源并且不会减慢tkinter 的速度（出于好奇）之外，这种情况是否真的有区别？你也是说不应该以这种方式使用subprocess.Popen（如我的回答中所述）吗？
嗯，你需要了解你的工作量。 multiprocessing 不会以任何方式降低 CPU 利用率。从技术上讲，它实际上增加了开销。它所做的是将处理需求分散到其他 CPU。扫描 QR 码的图像是一项耗时的任务。如果要降低 CPU 负载，则需要降低每秒帧率。
@TimRoberts 好的，我想我终于明白了。 multiprocessing 适用于 CPU 绑定任务，而 threading - 适用于 IO 绑定任务。据我了解，检查 QR 码的任务将是 IO 绑定的，但是这两个模块都不是为了减少 CPU 使用率？它们只允许同时运行代码（在使用 tkinter 之类的东西时是必需的（除非循环不需要太多时间））并且可以减少执行任务所需的时间，但是为了减少 CPU 使用率，代码应该被优化或对于这种情况，运行次数更少？
@Matiss tkinter 不需要 CPU。 tkinter 进程花费时间等待点击和击键。以上是关于如何在 tkinter 旁边运行循环？ - Python 二维码阅读器图形用户界面的主要内容，如果未能解决你的问题，请参考以下文章
如何在另一个任务旁边运行 tkinter 的 after event？