如何使用多处理python按顺序运行多个函数？

Question

I'm trying to run multiple functions with multiprocessing and running into a bit of wall. I want to run an initial function to completion on all processes/inputs and then run 2 or 3 other functions in parallel on the output of the first function. I've already got my search function. the code is there for the sake of explanation.

I'm not sure how to continue the code from here. I've put my initial attempt below. I want all instances of process1 to finish and then process2 and process3 to start in parallel.

Code is something like:

from multiprocessing import Pool


def init(*args):
    global working_dir
    [working_dir] = args

def process1(InFile):
    python.DoStuffWith.InFile
    Output.save.in(working_dir)

def process2(queue):
    inputfiles2 = []
    python.searchfunction.appendOutputof.process1.to.inputfiles2
    python.DoStuffWith.process1.Output
    python.Output

def process3(queue):
    inputfiles2 = []
    python.searchfunction.appendOutputof.process1.to.inputfiles2
    python.DoStuffWith.process1.Output
    python.Output

def MCprocess():
    working_dir = input("enter input: ")
    inputfiles1 = []
    python.searchfunction.appendfilesin.working_dir.to.inputfiles1
    with Pool(initializer=init, initargs=[working_dir], processes=16) as pool:
        pool.map(process1, inputfiles1)
        pool.close()
    
    #Editted Code
    queue = multiprocessing.Queue
    queue.put(working_dir)
    queue.put(working_dir)
    ProcessTwo = multiprocessing.Process(target=process2, args=(queue,))
    ProcessThree = multiprocessing.Process(target=process3, args=(queue,))
    ProcessTwo.start()
    ProcessThree.start()
     #OLD CODE
     #with Pool(initializer=init, initargs=[working_dir], processes=16) as pool:
        #pool.map_async(process2)
        #pool.map_async(process3)
    


if __name__ == '__main__':
    MCprocess()

Answer 1

Your best bet is to use an Event. The first process calls event.set() when it is done to indicate that the event has happened. The waiting processes use event.wait() or one of its variants to wait to be awoken that the event has been set.