python爬取贴吧图片

 1 #!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time    : 2017/7/19 16:21
# @Author  : wqj
# @Contact : wqjhky@gmail.com
# @Site    : 
# @File    : img.py
# @Software: PyCharm Community Edition
import re
import urllib
import sys
import os

reload(sys)
sys.setdefaultencoding('utf8')


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


def getImg(html, x):
    reg = r'src="(http://img.*?.jpg)"'
    imgre = re.compile(reg)
    imList = re.findall(reg, html)

    print(imList)
    for i in imList:
        print(i)
        print x
        urllib.urlretrieve(i, '%s.jpg' % x)
        x += 1
    return x


x = 1
url = "http://tieba.baidu.com/f?ie=utf-8&kw=%E6%B2%B3%E5%8D%97%E7%A7%91%E6%8A%80%E5%AD%A6%E9%99%A2&fr=search?pn="
for k in range(1, 28):
    ul = url + str(k)
    print ul
    html = getHtml(ul)
    # print html
    x = getImg(html, x)
    dirpath = r'D:img'
    filename = os.path.join(dirpath, str(x) + ".jpg")

 以上是代码，程序较为简单