import re,requests
from lxml import etree
from concurrent.futures import ThreadPoolExecutor,ProcessPoolExecutor
def urls(url):
res=requests.get(url)
html=re.compile('lazysrc="(?P<img_urls>.*?)"',re.S)
image_urls=html.finditer(res.text)
for url in image_urls:
res2=requests.get(url.group('img_urls'))
s=url.group('img_urls')
i = s.split('.')[2].split('/')[-1]
with open(f'tupian2/{i}.jpg', mode='wb') as f:
f.write(res2.content)
if __name__ == '__main__':
with ThreadPoolExecutor(20) as t:
for i in range(1,10):
t.submit(urls,f'https://www.3gbizhi.com/meinv/index_{i}.html')
print('下载完成')