Python自学记录--验证码识别模拟登陆

菠萝儿

1. #!/usr/bin/env python
   
2. # coding:utf-8
   
3. 
   
4. import requests
   
5. from hashlib import md5
   
6. from lxml import etree
   
7. import time
   
8. 
   
9. t = int(round(time.time() * 1000)) #时间戳
   
10. 
    
11. class Chaojiying_Client(object):
    
12. 
    
13.     def __init__(self, username, password, soft_id):
    
14.         self.username = username
    
15.         password =  password.encode('utf8')
    
16.         self.password = md5(password).hexdigest()
    
17.         self.soft_id = soft_id
    
18.         self.base_params = {
    
19.             'user': self.username,
    
20.             'pass2': self.password,
    
21.             'softid': self.soft_id,
    
22.         }
    
23.         self.headers = {
    
24.             'Connection': 'Keep-Alive',
    
25.             'User-Agent': 'Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 5.1; Trident/4.0)',
    
26.         }
    
27. 
    
28.     def PostPic(self, im, codetype):
    
29.         """
    
30.         im: 图片字节
    
31.         codetype: 题目类型 参考 http://www.chaojiying.com/price.html
    
32.         """
    
33.         params = {
    
34.             'codetype': codetype,
    
35.         }
    
36.         params.update(self.base_params)
    
37.         files = {'userfile': ('ccc.jpg', im)}
    
38.         r = requests.post('http://upload.chaojiying.net/Upload/Processing.php', data=params, files=files, headers=self.headers)
    
39.         return r.json()
    
40. 
    
41.     def ReportError(self, im_id):
    
42.         """
    
43.         im_id:报错题目的图片ID
    
44.         """
    
45.         params = {
    
46.             'id': im_id,
    
47.         }
    
48.         params.update(self.base_params)
    
49.         r = requests.post('http://upload.chaojiying.net/Upload/ReportError.php', data=params, headers=self.headers)
    
50.         return r.json()
    
51. 
    
52. 
    
53. 
    
54. def tranformImgCode(imgPath,imgType):
    
55.     chaojiying = Chaojiying_Client('超级鹰账号', '超级鹰密码', '软件ID')   
    
56.     im = open(imgPath, 'rb').read()                                                
    
57.     return chaojiying.PostPic(im,imgType)['pic_str']
    
58. 
    
59. # print(tranformImgCode(r'D:\代码保存\超级鹰_Python\a.jpg',1902))
    
60. 
    
61. session = requests.Session() #维持会话,可以让我们在跨请求时保存某些参数
    
62. 
    
63. headers = {'User-Agent':'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.198 Safari/537.36'}
    
64. #识别验证码
    
65. url = 'https://so.gushiwen.cn/user/login.aspx?from=http://so.gushiwen.cn/user/collect.aspx'
    
66. 
    
67. page_text = session.get(url=url,headers=headers).text
    
68. 
    
69. #解析验证码图片地址
    
70. tree = etree.HTML(page_text)
    
71. 
    
72. img_src = 'https://so.gushiwen.cn'+tree.xpath('//img[@id="imgCode"]/@src')[0]+'?t='+str(t)
    
73. print(img_src)
    
74. 
    
75. 
    
76. #将验证码保存到本地
    
77. img_data= session.get(img_src,headers=headers).content
    
78. 
    
79. with open(r'D:\代码保存\a.jpg','wb') as fp:
    
80.     fp.write(img_data)
    
81. 
    
82. #识别验证码
    
83. code_text = tranformImgCode(r'D:\代码保存\a.jpg',1902)
    
84. print(code_text)
    
85. 
    
86. login_url = 'https://so.gushiwen.cn/user/login.aspx?from=http%3a%2f%2fso.gushiwen.cn%2fuser%2fcollect.aspx'
    
87. 
    
88. data = {
    
89.     '__VIEWSTATE': 'uoQIcMK9IobQkeY9MVAmtKoq33r3OLASfUdWmTdVKWkPtsNNpQMuqH92Trfdy05CDl1WFTmp5wys9JzcE+3wziAytRHxPGrKiZP9/29sMhAoT3AIsd0WDlhIqoU=',
    
90.     '__VIEWSTATEGENERATOR': 'C93BE1AE',
    
91.     'from': 'http://so.gushiwen.cn/user/collect.aspx',
    
92.     'email': '网站账号',
    
93.     'pwd': '网站密码',
    
94.     'code': code_text,
    
95.     'denglu': '登录',
    
96. }
    
97. 
    
98. #点击登陆按钮发起请求，获取了登陆成功后对应的页码源码数据
    
99. page_text_login = session.post(url=login_url,headers=headers,data=data).text
    
100. 
     
101. 
     
102. with open(r'D:\代码保存\gushiwen.html','w',encoding='utf-8') as fp:
     
103.     fp.write(page_text_login)
     

复制代码

weigar

加油！易起学习努力