# _*_ coding: utf-8 _*_

'''
模擬知乎登陸并使用tesseract-ocr識別驗證碼

Author: Insomnia
Version: 0.0.1
Date: 2017-09-07
Language: Python3.6.2
Editor: Sublime Text3
'''

import requests, os, time, re
from bs4 import BeautifulSoup
from PIL import Image

# 當前項目路徑
cur_path = os.getcwd() + '/'

class ZhiHuSpider(object):
    '''
    本類主要用于實現模擬知乎登陸并使用tesseract-ocr識別驗證碼

    Attribute:
        session: 建立會話
        url_signin: 登陸頁面鏈接
        url_login: 登陸接口
        url_captcha: 驗證碼鏈接
        headers： 請求頭部信息
        num: 識別驗證碼次數
    '''

    def __init__(self):
        self.session = requests.Session();
        self.url_signin = 'https://www.zhihu.com/#signin'
        self.url_login = 'https://www.zhihu.com/login/email'
        self.url_captcha = 'https://www.zhihu.com/captcha.gif?r=%d&type=login' % (time.time() * 1000)
        self.headers = {'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_12_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/60.0.3112.101 Safari/537.36'}
        self.num = 1

    def get_captcha(self):
        '''
        獲取并使用tesseract-ocr識別驗證碼

        Returns: 
            返回驗證碼
        '''
        # 創建文件夾存放驗證碼
        if not os.path.exists(cur_path + 'captcha'):
            os.mkdir(cur_path + 'captcha')

        captcha_text = ''
        while True:
            # 下載驗證碼圖片
            captcha = self.session.get(self.url_captcha, headers=self.headers).content
            captcha_path = cur_path + 'captcha/captcha.gif' # 驗證碼圖片路徑
            captcha_path_new = cur_path + 'captcha/captcha_new.gif' # 處理后的驗證碼圖片路徑
            with open(captcha_path, 'wb') as f:
                f.write(captcha)

            # 圖片處理便于識別文字：彩色轉灰度，灰度轉二值，二值圖像識別
            im = Image.open(captcha_path)
            w, h = im.size
            im = im.convert('L') # convert()用于不同模式圖像之間的轉換
            threshold = 100 # 圖片降噪處理
            table = []
            for i in range(256):
                if i < threshold:
                    table.append(0)
                else:
                    table.append(1)
            im = im.point(table, '1')
            w_new, h_new = w * 2, h * 2 
            im = im.resize((w_new, h_new), Image.ANTIALIAS) # 圖片放大
            im.save(captcha_path_new)

            # 執行tesseract-ocr識別驗證碼
            captcha_text_path = cur_path + 'captcha/captcha_text' # 識別驗證碼后驗證碼文字的存放路徑
            cmd = '/usr/bin/tesseract %s %s' % (captcha_path_new, captcha_text_path)
            os.system(cmd)
            time.sleep(2)
            with open('%s.txt' % captcha_text_path, 'r') as f:
                try:
                    captcha_text = f.read().strip()
                    print('第 %d 次識別的驗證碼為：%s' % (self.num, captcha_text))
                    regex = re.compile('^[0-9a-zA-Z]{4}$') # 正則表達式
                    if captcha_text and re.search(regex, captcha_text):
                        break;
                    else:
                        print('驗證碼無效！重新識別中...')
                        self.num += 1
                        # time.sleep(2) # 避免過于頻繁的訪問
                except Exception as e:
                    print('Exception:', e)
                    break;
        return captcha_text

    def login(self, username, password):
        '''
        登陸接口

        Args:
            username: 登陸賬戶
            password: 登陸密碼
        Returns:
            返回登錄結果list
        '''
        soup = BeautifulSoup(self.session.get(self.url_signin, headers=self.headers).content, 'html.parser')
        # 獲取xsrf_token
        xsrf = soup.find('input', attrs={'name': '_xsrf'}).get('value')
        post_data = {
            '_xsrf': xsrf,
            'email': username,
            'password': password,
            'captcha': self.get_captcha()
        }
        login_ret = self.session.post(self.url_login, post_data, headers=self.headers).json()
        return login_ret

    def get_index_topic(self):
        '''
        獲取首頁第一條話題記錄
        '''
        print('-'*50, '獲取知乎首頁第一條話題記錄', '-'*50, sep="\n")
        soup = BeautifulSoup(self.session.get(self.url_signin, headers=self.headers).content, 'html.parser')
        item = soup.find('div', attrs={'class': 'Card TopstoryItem TopstoryItem--experimentFont18'})
        popover = item.find('div', attrs={'class': 'Feed-title'}).find('div', attrs={'class': 'Popover'}).find('div').get_text() # 話題
        print('來自話題：%s' % popover)
        title = item.find('div', attrs={'class': 'ContentItem AnswerItem'}).find('meta', attrs={'itemprop': 'name'}).get('content') # 標題
        print('title: %s' % title)
        content = item.find('div', attrs={'class': 'ContentItem AnswerItem'}).find('span', attrs={'class': 'RichText CopyrightRichText-richText'}).get_text() # 內容
        print('content: %s' % content)

if  __name__ == '__main__':
    zhihu = ZhiHuSpider()
    # 循環嘗試登陸
    while True:
        ret = zhihu.login('知乎登陸郵箱', '知乎密碼')
        if ret['r'] == 0:
            print('登陸成功! ')
            zhihu.get_index_topic()
            break
        else:
            print('登陸失敗: %s' % ret['msg'])
  

運行效果圖

image.png