2

I want to update a progress bar from inside a spawned process as follows:

import multiprocessing as mp
import random
import time
from tqdm import tqdm

def test(queue, pbar, lock):
    while True:
        x = queue.get()
        if x is None:
            break
        for i in range(x):
            time.sleep(1)
            lock.acquire()
            pbar.update(1)
            lock.release()

queue = mp.Queue()
lock = mp.Lock()
processes = []
pbar = tqdm(total=5050)
for rank in range(4):
    p = mp.Process(target=test, args=(queue, pbar, lock))
    p.start()
    processes.append(p)
pbar.close()

for idx in range(100):
    queue.put(idx)

for _ in range(4):
    queue.put(None)  # sentinel values to signal subprocesses to exit

for p in processes:
        p.join()  # wait for all subprocesses to finish

The above gives inconsistent updates (progess goes up and down). I found this answer, but none of them work for me because I want to update the progress bar inside the test function. How can I do this?

Priyatham
  • 2,821
  • 1
  • 19
  • 33

1 Answers1

2

I'd slightly restructure the program:

1.) Create update_bar process that creates a progress bar and reads from another queue values and updates the bar with these values

2.) This update process has daemon=True parameter, so it won't block upon exit

3.) The test processes receives upon start the bar_queue and put values there if they want to update the progress bar.

import time
from tqdm import tqdm
import multiprocessing as mp


def test(queue, bar_queue):
    while True:
        x = queue.get()
        if x is None:
            break
        for _ in range(x):
            time.sleep(0.05)
            bar_queue.put_nowait(1)


def update_bar(q):
    pbar = tqdm(total=188)

    while True:
        x = q.get()
        pbar.update(x)


if __name__ == "__main__":
    queue = mp.Queue()
    bar_queue = mp.Queue()

    processes = [
        mp.Process(target=test, args=(queue, bar_queue)) for _ in range(4)
    ]

    # start update progress bar process
    # daemon= parameter is set to True so this process won't block us upon exit
    bar_process = mp.Process(target=update_bar, args=(bar_queue,), daemon=True)
    bar_process.start()

    for p in processes:
        p.start()

    for idx in range(20):
        queue.put(idx)

    for _ in range(4):
        queue.put(None)  # sentinel values to signal subprocesses to exit

    for p in processes:
        p.join()  # wait for all subprocesses to finish

    time.sleep(0.5)
Andrej Kesely
  • 168,389
  • 15
  • 48
  • 91