Pyqt UI non responsive on Multithreading with python

[JonB]

@Albitroz
You can use a queue, that in itself is fine. You should also be careful to mutex, either code yourself or it must be built into the queue.

The signal/slot would be needed at least to let the other side know (in your case, the main thread) that new data has been queued. How else would you intend to do that bit?

[Albitroz]

@JonB i was planing to check the queue size at the other end at a pre defined interval, if queue size is greater than a pre defined size, i just dequeue my queue and get the data. thats my plan is that possible?

[Albitroz]

@JonB in the same code, im facing issues in stopping my thread,
im calling a function from my main code

def stop_task():
    running=False

which will set the running variable to False, and in my thread, im checking for that variable

def ai_read():
    print("Inside Thread")
    task_in.start()
    while(running):
            data=task_in.read(number_of_samples_per_channel=100)
            #print(data)
            time.sleep(.01)

I have verified the value of running with printf inside the thread, its not getting updated, is there any thread safe variable?

[JonB]

@Albitroz said in Pyqt UI non responsive on Multithreading with python:

@JonB i was planing to check the queue size at the other end at a pre defined interval, if queue size is greater than a pre defined size, i just dequeue my queue and get the data. thats my plan is that possible?

It will work, bit it's not as efficient as it could be.

To do that, your main thread will have to have that interval timer and check each time. Plus often(?) the queue will not be of that size, and it will be a waste of time. Wouldn't you like to know when the thread actually pushes another item to the queue, and only check then/when the count is reached? That is what having the thread send a signal (when it queues), and the main thread have a slot attached to the signal, would achieve.

The only time where "timed-polling" is (almost) just as good is if you only want the receiver to pull from the queue at those interval points, not as soon as the data arrives. All in all it's hard to see why signalling would not be a preferable solution.

Sooner or later you will have to use signals & slots in Qt programming, so why not now?

[JonB]

@Albitroz said in Pyqt UI non responsive on Multithreading with python:

I have verified the value of running with printf inside the thread, its not getting updated, is there any thread safe variable?

All your threading stuff is being done with Python threading. So your questions are best asked in a Python forum. I don't know when stop_task() gets called while ai_read() is in a loop.

Up to you, but I would use Qt signals & slots for everything like this.

[Albitroz]

@JonB ok i will use signals and slots as a callback to get the data, but im facing issues stopping the thread

[SGaist]

Hi,

What issues are you having ?

Note that depending on what you do in your loop you should put break statement at regular intervals to allow for an easier way out of it.

[Albitroz]

@SGaist HI
Currently im facing issues, in transferring data from different threads

Im splitting my entire project to different ".py " file for easy debugging, in this case how can i transfer data from one file to other

to be pressie,

I have my main ".py file which runs the pyqt event loop, i have another file which generates some data through a thread". How can i pass data from this thread to my main python file, can i use signals and slots to transfer data to the main file which holds the UI

my main intention is to send data from different threads to the main file which runs QT event and UI so that i can display data from these threads

which is better, implementing a queue or signals and slots?

[SGaist]

The files play no role here.

If you are using QThread then you can easily use signals and slots to transfert data.

Otherwise are you using Python's multithreading or multiprocessing ? If you really want parallel processing with Python, you shall use the later.

[Albitroz]

@SGaist
I will be handling around 20 threads of different functionalities, and all are time critical activities, so stack overflow suggests using Multiprocessing library which one do you prefer?

if i go with multiprocessing library, how can i transfer data between files , my main action is to display these processed data on UI and thats the reason i use PyQt

[SGaist]

You can check this article. At the there's a link for a mix between PyQt and multiprocessing.