|
楼主 |
发表于 2016-8-10 11:01:52
|
显示全部楼层
- import urllib.request
- import os
- import re
- import time
- from selenium import webdriver
- def url_open(url1):
- req=urllib.request.Request(url1,headers={'User-Agent':'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/46.0.2490.86 Safari/537.36'})
- response=urllib.request.urlopen(req)
- a=response.read()
- return a
-
- #url='http://www.10x10.co.kr/shopping/category_prd.asp?itemid=1517740&disp=103103103&pRtr=fulldesign'
- url='http://www.1300k.com/shop/goodsDetail.html?f_sp=&f_goodsno=215023298934'
- #url='http://www.fishc.com'
- a=webdriver.Chrome()
- a.get(url)
- #time.sleep(30)
- c=a.page_source
- #re1=r'src="([https?://.*\.jpg|https?://.*\.png]{0,85})"'
- re1=r'src="(https?://.*\.(?:jpg|png)){0,85}"'
- imglist=re.findall(re1,c)
- print(imglist)
- os.mkdir('E:/test\\pic')
- os.chdir('E:/test\\pic')
- for n in imglist:
- filename=n.split('/')[-1]
- try:
- with open(filename,'wb') as f:
- img=url_open(n)
- f.write(img)
- except Exception:
- pass
复制代码 |
|