脚本专栏 
首页 > 脚本专栏 > 浏览文章

Python并发请求下限制QPS(每秒查询率)的实现代码

(编辑:jimmy 日期: 2024/11/18 浏览:3 次 )

"htmlcode">

import grequests

urls = [
 "https://www.baidu.com",
 "https://www.google.com"
]
requests = [
 grequests.get(url)
 for url in urls
] * 1000

rate = 20 # 表示 20 请求/秒

time.sleep(1)

"normal">请++time.sleep(1)Time = 请求准备时延 + 请求发送时延 + time.sleep(1)Time=请求准备时延+请求发送时延+time.sleep(1) "htmlcode">

from time import sleep

req_groups = [
 requests[i: i+rate]
 for i in range(0, len(requests), rate)
]

ret = []
for req_group in req_groups:
 ret += grequests.map(req_group)
 sleep(1)

print(ret)

令牌桶(token bucket)方法

"normal">请++延Time = 请求准备时延 + 请求发送时延 + 令牌桶阻塞时延Time=请求准备时延+请求发送时延+令牌桶阻塞时延 1"normal">请+延令牌桶阻塞时延 ≈ 1 - 请求准备时延 + 请求发送时延令牌桶阻塞时延≈1"htmlcode">

from time import time

class Throttle:
 def __init__(self, rate):
  self.rate = rate
  self.tokens = 0
  self.last = 0
 
 def consume(self, amount=1):
  now = time()
  
  if self.last == 0:
   self.last = now
  
  elapsed = now - self.last

  if int(elapsed * self.rate):
   self.tokens += int(elapsed * self.rate)
   self.last = now
  
  self.tokens = (
   self.rate
   if self.tokens > self.rate
   else self.tokens
  )
  
  if self.tokens >= amount:
   self.tokens -= amount
  else:
   amount = 0
  
  return amount

throttle = Throttle(rate)

req_groups = [
 requests[i: i+rate]
 for i in range(0, len(requests), rate)
]

ret = []
for req_group in req_groups:
 ret += grequests.map(req_group)
 while throttle.consume():
  pass # 阻塞

print(ret)

GRequests-Throttle

"htmlcode">

pip install grequests-throttle
import grequests_throttle as gt

ret = gt.map(requests, rate=rate)
print(ret)

总结

  如果并发请求数量较小,可以考虑使用time.sleep(1)简单快捷;当并发请求数量较大时,使用令牌桶(token bucket)方法能最大化利用每一秒;如果不想写太多代码,可以使用GRequests-Throttle包进行请求流量控制。

上一篇:Python爬虫爬取百度搜索内容代码实例
下一篇:python3读取autocad图形文件.py实例
友情链接:杰晶网络 DDR爱好者之家 南强小屋 黑松山资源网 白云城资源网 SiteMap