[已解决]关于爬虫

xiaosi4081

1. import re
   
2. import requests
   
3. from bs4 import BeautifulSoup
   
4. 
   
5. def get_html(url):
   
6.     header={
   
7.         'User-Agent': 'Mozilla / 5.0(Windows NT 10.0;WOW64) AppleWebKit / 537.36(KHTML, likeGecko) Chrome / 78.0.3904.108Safari / 537.36QIHU360EE',
   
8.         'Referer': 'https: // www.so.com / s?src = 360chrome_newtab_search & q = % E5 % B8 % 8C % E5 % B3 % B6 % E3 % 81 % 82 % E3 % 81 % 84 % E3 % 82 % 8ASec - Fetch - Mode: navigate'
   
9.     }
   
10.     html=requests.get(url=url,headers=header)
    
11.     #print(html.text)
    
12.     return html
    
13. 
    
14. # def get_list(html):
    
15. #     html=html.text
    
16. #     soup=BeautifulSoup(html,'html.parser')
    
17. #     img_re=re.compile('https://p\d[.]ssl[.]qhimgs1[.]com/sdr/400__/.*？[.]jpg')
    
18. #     img_list=soup.find_all(img_re,soup)
    
19.     # print(img_list)
    
20.     #a=re.compile(r'<img style=".*?src="(.*?[.]jpg)"',re.S)
    
21.     #link_list=re.findall(a,html.text)
    
22. 
    
23.     #return link_list
    
24. def get_list(html):
    
25.     soup = BeautifulSoup(html.text,'html.parser')
    
26.     link_list1 = soup.find_all("img")
    
27.     link_list = []
    
28.     for i in link_list1:
    
29.         link_list.append(img_list1["src"])
    
30.     #print(link_list)
    
31.     return link_list
    
32. 
    
33. 
    
34. def main():
    
35.     url='https://image.so.com/i?src=360pic_normal&z=1&i=0&cmg=15484592.3836743514792807400.1594087443636.3574&q=%E5%B8%8C%E5%B2%9B%E3%81%82%E3%81%84%E3%82%8A'
    
36.     html=get_html(url)
    
37.     img_list=get_list(html)
    
38.     #print(img_list)
    
39. 
    
40.     for img_url in img_list:
    
41.         img_name='美眉/'+img_url.split('/')[-1]
    
42.         print(img_name)
    
43.         with open(img_name,'wb') as f:
    
44.             html=get_html(img_list)
    
45.             print(html.content)
    
46.             f.write(html.content)
    
47. 
    
48. 
    
49. 
    
50. if __name__ == "__main__":
    
51.     main()

复制代码