列维坦油画（搜狐）

最新推荐文章于 2024-09-17 07:32:06 发布

Elisiiiii

最新推荐文章于 2024-09-17 07:32:06 发布

阅读量155

点赞数

文章标签：爬虫

本文链接：https://blog.csdn.net/weixin_48176210/article/details/112427954

版权

import os,shutil
from urllib.request import urlopen, Request,urlretrieve
from bs4 import BeautifulSoup
url = 'https://www.sohu.com/a/286956359_301394'
headers = { "user-agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_6) ""AppleWebKit/537.36 (KHTML, like Gecko) Chrome/81.0.4044.129 Safari/537.36"}
ret = Request(url, headers=headers)
html = urlopen(ret)
bs = BeautifulSoup(html, 'html.parser')
img_path= bs.find('article')
index=0
if os.path.exists("img_download"):
    shutil.rmtree("img_download")
    os.makedirs("img_download")
else:
    os.makedirs("img_download")
    for img_paths in img_path.find_all('img'):
        img_paths=img_paths.attrs['src']
        index=index+1
        print(img_paths)
        url = '{}'.format(img_paths)
        dir = os.path.abspath('img_download')
        work_path = os.path.join(dir, '{}.png'.format(index))
        urlretrieve(url, work_path)

http://5b0988e595225.cdn.sohucs.com/images/20190106/a90ff17487e84c05ad1ac4b827900874.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/d303f3cc4dff46718b65a9e753e84234.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/8b38d5a1039d445085516c85209c7173.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/600f531a1e784923be8ef69b6dae1e74.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/aaf53b9fbf4e4712a3ce6fb81d559252.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/f5fad84d65964cf7b9b631d905388b53.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/528a726489ba4a9bb8174e866993f97a.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/bc5edbe568c2400284967f5b7f23f8f7.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/27eaea4dc5dd4ff892d597f9b173f809.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/a35fa32948ed4c4ab0c7c19fb9819bfc.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/fff13f94c4444d52888aee37d8330c0f.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/297cd24aa0a54a8291fe3d8813e7165e.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/591c1a57744e4de9a432dfec9721ffd5.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/e5b5978f9b5243cab8801e6ba3f5b88d.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/14f2bf90a00c4be79a4972fe8835e952.jpeg
http://5b0988e595225.cdn.sohucs.com/images/20190106/cb488b3d8e47455a841c40a2ff48a518.jpeg