要破解这个验证码当然不是啥难事，诚然可以自己去训练模型，然而大量的训练集测试数据还得自己去打码。
#baidu-aip 智能识别验证码
from aip import AipOcr
import requests
from lxml import etree
import time
""" 你的 APPID AK SK """
APP_ID = 'APP_ID'
API_KEY = 'API_KEY'
SECRET_KEY = 'SECRET_KEY'
#识别验证码图片
def tranformImgCode(imgPath):
client = AipOcr(APP_ID, API_KEY, SECRET_KEY)
im = open(imgPath, 'rb').read()
return client.basicGeneral(im)["words_result"][0]["words"] #取验证码
t = int(round(time.time() * 1000)) #时间戳
session = requests.Session() #维持会话,可以让我们在跨请求时保存某些参数
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'}
#获取验证码图片
url = 'https://so.gushiwen.cn/user/login.aspx?from=http://so.gushiwen.cn/user/collect.aspx' #登陆地址
page_text = session.get(url=url,headers=headers).text #获取登陆地址内容
#解析验证码图片地址
tree = etree.HTML(page_text)
img_src = 'https://so.gushiwen.cn'+tree.xpath('//img[@id="imgCode"]/@src')[0]+'?t='+str(t)
VIEWSTATE = tree.xpath('//*[@id="__VIEWSTATE"]/@value')
VIEWSTATEGENERATOR = tree.xpath('//*[@id="__VIEWSTATEGENERATOR"]/@value') #动态数据
#将验证码保存到本地
img_data= session.get(img_src,headers=headers).content
with open(r'D:\代码保存\a.jpg','wb') as fp:
fp.write(img_data)
#识别验证码
code_text = tranformImgCode(r'D:\代码保存\a.jpg')
print(code_text)
login_url = 'https://so.gushiwen.cn/user/login.aspx?from=http%3a%2f%2fso.gushiwen.cn%2fuser%2fcollect.aspx'
data = {
'__VIEWSTATE': VIEWSTATE,
'__VIEWSTATEGENERATOR': VIEWSTATEGENERATOR,
'from': 'http://so.gushiwen.cn/user/collect.aspx',
'email': '网页账号',
'pwd': '网页密码',
'code': code_text,
'denglu': '登录',
}
#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
page_text_login = session.post(url=login_url,headers=headers,data=data).text
#保存网页
with open(r'D:\代码保存\gushiwen.html','w',encoding='utf-8') as fp:
fp.write(page_text_login)basicGeneral换成basicAccurate精度要高些。