模拟登陆豆瓣

login_db.py

from requests import session, get
from re import search, compile
from PIL import Image


class Douban:
    def __init__(self):
        self.url = 'https://accounts.douban.com/login'
        self.head = {
            'User-Agent':
            ('Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 '
             '(KHTML, like Gecko) Chrome/59.0.3071.115 Safari/537.36')
        }
        self.responce = get(self.url).text
        self.ssion = session()
        self.data = {
            'source': 'None',
            'redir': 'https://www.douban.com',
            'form_email': '',
            'form_password': '',
            'login': '登录'
        }

    def needCatpcha(self):
        return search(r'l>验证码', self.responce)

    def getCatpcha(self):
        img = compile(r'img id="captcha_image" src="(.*?)"')
        img_url = img.findall(self.responce)[0]
        jpg = get(img_url).content
        with open('captcha.jpg', 'wb') as f:
            f.write(jpg)
        Image.open('captcha.jpg').show()
        captcha_solu = input('input captcha: ')
        capt = compile(r'captcha\?id=(.*?)\&amp')
        captcha_id = capt.findall(img_url)[0]
        self.data = {
            'source': 'None',
            'redir': 'https://www.douban.com',
            'form_email': '',
            'form_password': '',
            'captcha-solution': captcha_solu,
            'captcha-id': captcha_id,
            'login': '登录'
        }

    def main(self):
        if self.needCatpcha():
            self.getCatpcha()
            self.ssion.post(self.url, headers=self.head, data=self.data)
        else:
            self.ssion.post(self.url, headers=self.head, data=self.data)
        return self.ssion


db = Douban()
ssion = db.main()
print(ssion.get('https://www.douban.com/').text)