多线程下载

Posted 2020-11-15 notfind

多线程下载

梨视频下载为例

要点：

1.获取所有视频的url,定义下载函数getdata

2.将下载数据保存到列表，定义数据保存函数savefunc

3.from multiprocessing.dummy import Pool

? pool = Pool(4)

? pool.get(getdata,url)

? pool.map(savefunc,datalist)

import requests,re
from multiprocessing.dummy import Pool
from lxml import etree
from random import randint

target_url = 'https://www.pearvideo.com'
headers = {
    'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (Khtml, like Gecko) Chrome/76.0.3809.132 Safari/537.36',
    }
response = requests.get(url = target_url,headers = headers)
ht = response.text
tree = etree.HTML(ht)
link_list = tree.xpath("//*[@id = 'actRecommendCont'][1]//a[@class = 'actcont-detail actplay']/@href")
videoUrl = []
for link in link_list:
    detail = 'https://www.pearvideo.com/' + link
    response1 = requests.get(url = detail,headers=headers)
    ht = response1.text
    reg = 'var contId.*?srcUrl="(.*?)"'
    link = re.findall(reg,ht,re.S)[0]
    videoUrl.append(link)
    
def getVideoDate(url):
    return requests.get(url = url,headers = headers).content
pool = Pool(4)
videoData_list = pool.map(getVideoDate,videoUrl)

def saveVideo(VD):
    fn = int(randint(1,999))
    with open('./%s.mp4' % fn,'wb') as f:
        f.write(VD)
        print('视频%s下载完毕' % fn)
pool.map(saveVideo,videoData_list)