9.4. Multiprocessing¶

IPC¶: Inter Process Communication
Process¶
Daemon¶
Dead Lock¶
Race Condition¶
Starvation¶

9.4.1. Problems¶

Zakleszczania
Race Condition

9.4.2. Cykl życia procesów¶

Tworzenie procesów
Zamykanie procesów
Multiprocesowość
Komunikacja między procesami

9.4.3. Rationale¶

IPC - Inter-Process Communication
Aby pickle mógł odtworzyć obiekt, musi posiadać jego definicję - klasę.

9.4.4. Define payload¶

iris.py:

from dataclasses import dataclass

@dataclass
class Iris:
    sepal_length: float
    sepal_width: float
    petal_length: float
    petal_width: float
    species: str

    def sepal_area(self):
        return self.sepal_length * self.sepal_width

    def petal_area(self):
        return self.petal_length * self.petal_width

    def total_area(self):
        return self.sepal_area() + self.petal_area()

9.4.5. Client¶

Obiekt wysyłający dane multiprocessing-client.py:

import pickle
from multiprocessing.connection import Client
from iris import Iris


flower = Iris(
    sepal_length=5.1,
    sepal_width=3.5,
    petal_length=1.4,
    petal_width=0.2,
    species='setosa'
)

payload = pickle.dumps(flower)


ADDRESS = ('localhost', 6000)
PASSWORD = b'My voice is my password, verify me.'

connection = Client(ADDRESS, authkey=PASSWORD)
connection.send(payload)
connection.send('close')
connection.close()

9.4.6. Listener¶

Obiekt nasłuchujący na połączenia multiprocessing-listener.py:

import pickle
from multiprocessing.connection import Listener
from iris import Iris


ADDRESS = ('localhost', 6000)
PASSWORD = b'My voice is my password, verify me.'

listener = Listener(ADDRESS, authkey=PASSWORD)
connection = listener.accept()

while True:
    payload = connection.recv()

    if payload == 'close':
        connection.close()
        break

    flower = pickle.loads(payload)
    area = flower.total_area()
    print(f'Area: {area}')

listener.close()