import time
COUNT = 100000000
start_time = time.time()
def count(num):
counter = 0
while counter < num:
counter += 1
count(COUNT)
return "time: %s seconds" % (time.time() - start_time)import threading
import time
start_time = time.time()
COUNT = 100000000
def count(num):
counter = 0
while counter < num:
counter += 1
t1 = threading.Thread(target=count, args=(COUNT//2,))
t2 = threading.Thread(target=count, args=(COUNT//2,))
t1.start()
t2.start()
t1.join()
t2.join()
return "time: %s seconds" % (time.time() - start_time)import time
start_time = time.time()
def nap(duration):
time.sleep(duration)
print("Nap duration: {}".format(duration))
nap(2)
nap(2)
nap(2)
return "time: %s seconds" % (time.time() - start_time)import time
import threading
start_time = time.time()
def nap(duration):
time.sleep(duration)
print("Nap duration: {}".format(duration))
t1 = threading.Thread(target=nap, args=(2,))
t2 = threading.Thread(target=nap, args=(2,))
t3 = threading.Thread(target=nap, args=(2,))
t1.start()
t2.start()
t3.start()
t1.join()
t2.join()
t3.join()
return "time: %s seconds" % (time.time() - start_time)Iterators provide a way to fetch items lazily when scanning sequences that don’t fit in memory.
They need to provide next and iter methods.
class Ip:
def __init__(self, ip):
self.ip = ip
self.octets = ip.split('.')
def __repr__(self):
return 'IP(%s)' % self.ip
def __iter__(self):
return IpIterator(self.octets)
class IpIterator:
def __init__(self, octets):
self.octets = octets
self.index = 0
def __next__(self):
try:
octet = self.octets[self.index]
except IndexError:
raise StopIteration()
self.index += 1
return octet
def __iter__(self):
return self
def __len__(self):
return 4
ip = Ip('192.168.12.4')
for octet in ip:
print(octet)
it = iter(Ip('192.167.1.3'))
for i in range(len(it)):
octet = next(it) # Same than it.__next__()
print(octet)192 168 12 4 192 167 1 3
Generators return an intermediate result, but maintain the function’s local state so it can be resumed where it left off.
def geo_prog(begin, ratio, limit=None):
while limit is None or begin < limit:
yield begin
begin *= ratio
prog = geo_prog(1, 3, 100)
print(prog)
for i in prog:
print(i)
# Generators share the same interface than iterators
del prog
prog = geo_prog(1, 4)
for i in range(5):
print(prog.__next__())<generator object geo_prog at 0x10c4b5728> 1 3 9 27 81 1 4 16 64 256
Coroutines share interface with generators but add the .send method.
def average():
total = 0.0
count = 0
avg = None
while True:
element = yield avg
total += element
count += 1
avg = total/count
avg = average()
next(avg)
for n in range(1, 11):
print(avg.send(n))1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
from collections import namedtuple
Event = namedtuple('Event', 'time team description')
def dev_process(team, tickets, start_time=0):
time = yield Event(start_time, team, 'TO DO')
for i in range(tickets):
time = yield Event(time, team, 'IN PROGRESS')
time = yield Event(time, team, 'IN QA')
time = yield Event(time, team, 'DONE')
yield Event(time, team, 'ALL DONE')
dev = dev_process('pod 1', 2, 0)
todo = next(dev)
in_progress = dev.send(todo.time + 1)
in_qa = dev.send(in_progress.time + 7)
done = dev.send(in_qa.time + 4)
print(todo, in_progress, in_qa, done, sep='\n')Event(time=0, team='pod 1', description='TO DO') Event(time=1, team='pod 1', description='IN PROGRESS') Event(time=8, team='pod 1', description='IN QA') Event(time=12, team='pod 1', description='DONE')
from queue import PriorityQueue
from random import randint
from collections import namedtuple
Event = namedtuple('Event', 'time team description')
def dev_process(team, tickets, start_time=0):
time = yield Event(start_time, team, 'TO DO')
for i in range(tickets):
time = yield Event(time, team, 'IN PROGRESS')
time = yield Event(time, team, 'IN QA')
time = yield Event(time, team, 'DONE')
yield Event(time, team, 'ALL DONE')
teams = {1: dev_process(team=1, tickets=2, start_time=0),
2: dev_process(team=2, tickets=3, start_time=3),
3: dev_process(team=3, tickets=1, start_time=4)}
class Simulator:
def __init__(self, procs):
self.events = PriorityQueue()
self.procs = procs
def run(self, end_time):
for key, proc in sorted(self.procs.items()):
event = next(proc)
self.events.put(event)
time = 0
while time < end_time:
if self.events.empty():
break
event = self.events.get()
time, proc_id, past_action = event
print('team: ', proc_id, proc_id * ' ', event)
active_proc = self.procs[proc_id]
next_time = time + randint(1, 5)
try:
next_event = active_proc.send(next_time)
except StopIteration:
del self.procs[proc_id]
else:
self.events.put(next_event)
else:
print('%s events pending' % self.events.qsize())
sim = Simulator(teams)
sim.run(20)team: 1 Event(time=0, team=1, description='TO DO') team: 1 Event(time=1, team=1, description='IN PROGRESS') team: 2 Event(time=3, team=2, description='TO DO') team: 1 Event(time=4, team=1, description='IN QA') team: 3 Event(time=4, team=3, description='TO DO') team: 1 Event(time=5, team=1, description='DONE') team: 2 Event(time=6, team=2, description='IN PROGRESS') team: 1 Event(time=8, team=1, description='IN PROGRESS') team: 2 Event(time=9, team=2, description='IN QA') team: 3 Event(time=9, team=3, description='IN PROGRESS') team: 1 Event(time=10, team=1, description='IN QA') team: 3 Event(time=11, team=3, description='IN QA') team: 2 Event(time=13, team=2, description='DONE') team: 1 Event(time=14, team=1, description='DONE') team: 2 Event(time=15, team=2, description='IN PROGRESS') team: 3 Event(time=16, team=3, description='DONE') team: 1 Event(time=18, team=1, description='ALL DONE') team: 2 Event(time=18, team=2, description='IN QA') team: 3 Event(time=18, team=3, description='ALL DONE') team: 2 Event(time=21, team=2, description='DONE') 1 events pending
asyncio implements concurrency using generators coordinated by an event loop.
import json
import asyncio
import aiohttp
@asyncio.coroutine
def get_pokemon(number):
url = 'http://pokeapi.co/api/v2/pokemon/{}'.format(number)
resp = yield from aiohttp.request('GET', url)
data = yield from resp.read()
return data
@asyncio.coroutine
def download_pokemon(number):
pokemon = yield from get_pokemon(number)
poke_json = json.loads(pokemon)
return poke_json
def download_pokemons(pokemon_numbers):
loop = asyncio.get_event_loop()
downloads = [download_pokemon(number) for number in sorted(pokemon_numbers)]
wait = asyncio.wait(downloads)
response, _ = loop.run_until_complete(wait)
loop.close()
return response
poke_tasks = download_pokemons([4, 5, 6])
for task in poke_tasks:
pokemon = task.result()
print(pokemon['name'])charmeleon charizard charmander
async and await are keywords added to…
This proposal aims to answer that need by making writing explicitly asynchronous, concurrent Python code easier and more Pythonic.
It is proposed to make coroutines a proper standalone concept in Python, and introduce new supporting syntax. The ultimate goal is to help establish a common, easily approachable, mental model of asynchronous programming in Python and make it as close to synchronous programming as possible.
import json
import aiohttp
async def get_pokemon(number):
url = 'http://pokeapi.co/api/v2/pokemon/{}'.format(number)
resp = await aiohttp.request('GET', url)
data = await resp.read()
return data
async def download_pokemon(number):
pokemon = await get_pokemon(number)
poke_json = json.loads(pokemon)
return poke_json
def download_pokemons(pokemon_numbers):
loop = asyncio.get_event_loop()
downloads = [download_pokemon(number) for number in sorted(pokemon_numbers)]
wait = asyncio.wait(downloads)
response, _ = loop.run_until_complete(wait)
loop.close()
return response
poke_tasks = download_pokemons([4, 5, 6])
for task in poke_tasks:
pokemon = task.result()
print(pokemon['name'])charizard charmander charmeleon