爬虫验证码识别-01--古诗词登录

Stubborn

看了代码，你也许会的好傻，哈哈，我也觉得是，请跳过，嘿，看下验证码啊！  这个辣眼的提示

# -*- coding: utf-8 -*-
# @Time    : 2019-04-10 14:47
# @Author  : Ein
# @File    : tesserocr验证码识别-古诗文.py
# @Software: PyCharm
'''
g1RRZ+h02qqY4rNct0PPw/ELkgPaoF82SoPcA8lIEM2q0bQsHXhB6ANpGwCX25N5r3jqUq9XpNkweWhcQD7j1iaAp4kG4xG48jx3iow1bBtRhU+9fqw1Y/Tp3g0=
C93BE1AE
'''

import requests,os
from lxml import etree
def yzm_file_path():
    currrent_path = os.path.realpath(__file__)  # 文件绝对路径
    current_dir = os.path.split(currrent_path)[0]
    return current_dir + r"/yzm.jpg"
def down_html():
    currrent_path = os.path.realpath(__file__)  # 文件绝对路径
    current_dir = os.path.split(currrent_path)[0]
    return current_dir + r"/htj.html"

headers = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64)"
}
session = requests.Session()

def download_yzm():
    url = 'https://so.gushiwen.org/user/login.aspx?from='
    resp = session.get(url=url,headers=headers).text
    tree = etree.HTML(resp)
    #图片地址：
    img_url = 'https://so.gushiwen.org' + tree.xpath("//img[@id='imgCode']/@src")[0]
    #获取验证码令牌
    VIEWSTATE = tree.xpath("//input[@type='hidden'][1]/@value")[0]
    VIEWSTATEGENERATOR = tree.xpath("//input[@type='hidden'][1]/@value")[1]
    #下载到本地
    response = session.get(url=img_url,headers=headers)
    yzm_path = yzm_file_path()
    with open(yzm_path, "wb") as f:
        f.write(response.content)
    return VIEWSTATE,VIEWSTATEGENERATOR

def poss_pwd(vi,vie):
    pwd_url = 'https://so.gushiwen.org/user/login.aspx?from=http%3a%2f%2fso.gushiwen.org%2fuser%2fcollect.aspx'
    yzm = input("嘿，看下验证码啊！=")
    data = {
        '__VIEWSTATE': vi,
        '__VIEWSTATEGENERATOR': vie,
        'from':'http://so.gushiwen.org/user/collect.aspx',
        'email':'812745685@qq.com',
        'pwd': 'asd123..',
        'code': yzm,
        'denglu': '登录',
    }
    path = down_html()
    response = session.post(url=pwd_url,data=data)
    with open(path,'w',encoding='utf8') as fp:
        fp.write(response.text)


def main():
    vi,vie = download_yzm()
    poss_pwd(vi,vie)

if __name__ == '__main__':
    main()

goolming

你好，可以加个好友交流一下吗