getimg

来源:互联网 发布:伤感的网络歌曲大全 编辑:程序博客网 时间:2024/06/16 02:55
#coding=utf-8
import re
import urllib.request


def getHtml(url):
    page = urllib.request.urlopen(url)
    html = page.read()
    return html


def getImg(html):
    html = html.decode('utf-8')
    reg = r'src="(.+?\.jpg)" pic_ext'
    imgre = re.compile(reg)
    imglist = re.findall(imgre,html)
    x = 0
    for imgurl in imglist:
        urllib.request.urlretrieve(imgurl,'%s.jpg' % x)
        x += 1
    return imglist  


html = getHtml("http://tieba.baidu.com/p/3620213846")


print (getImg(html))
0 0