02-12 单线程和多线程 代码实例

from multiprocessing.dummy import Pool
import requests
import time

start_time=time.time()
pool=Pool(5)

# 同步爬取代码
urls=['http://127.0.01:5000/bobo','http://127.0.01:5000/jay','http://127.0.01:5000/tom']
for url in urls:
    page_text=requests.get(url)
    print(page_text)
end_time=time.time()
t=end_time-start_time
print('同步抓取总耗时为:',t)

# 异步爬取代码
def get_request(url):
    return requests.get(url)
start_time2=time.time()
pool=Pool(3)

#多线程爬取
resp_list=pool.map(get_request,urls) #注意这resp是有map() 返回的一个lsit

#多线程解析
def parse(page_text)
    return '//p[@attrname="value"]'
resp=pool.map(parse,resp_list)

print('异步爬取耗时为',time.time()-start_time2)