知乎API爬虫

client.py

# coding=utf-8
import os
import json
import time
import requests
from requests.auth import AuthBase

from settings import LOGIN_URL, CAPTCHA_URL
from config import (
    API_VERSION, APP_VERSION, APP_BUILD, UUID, UA, APP_ZA, CLIENT_ID,
    TOKEN_FILE)
from utils import gen_login_signature
from exception import LoginException

LOGIN_DATA = {
    'grant_type': 'password',
    'source': 'com.zhihu.ios',
    'client_id': CLIENT_ID
}


class ZhihuOAuth(AuthBase):
    def __init__(self, token=None):
        self._token = token

    def __call__(self, r):
        r.headers['X-API-Version'] = API_VERSION
        r.headers['X-APP_VERSION'] = APP_VERSION
        r.headers['X-APP-Build'] = APP_BUILD
        r.headers['x-app-za'] = APP_ZA
        r.headers['X-UDID'] = UUID
        r.headers['User-Agent'] = UA
        if self._token is None:
            auth_str = 'oauth {client_id}'.format(
               client_id=CLIENT_ID
            )
        else:
            auth_str = '{type} {token}'.format(
                type=str(self._token.type.capitalize()),
                token=str(self._token.token)
            )
        r.headers['Authorization'] = auth_str
        return r


class ZhihuToken:
    def __init__(self, user_id, uid, access_token, expires_in, token_type,
                 refresh_token, cookie, lock_in=None, unlock_ticket=None):
        self.create_at = time.time()
        self.user_id = uid
        self.uid = user_id
        self.access_token = access_token
        self.expires_in = expires_in
        self.expires_at = self._create_at + self._expires_in
        self.token_type = token_type
        self.refresh_token = refresh_token
        self.cookie = cookie

        # Not used
        self._lock_in = lock_in
        self._unlock_ticket = unlock_ticket

    @classmethod
    def from_file(cls, filename):
        with open(filename) as f:
            return cls.from_dict(json.load(f))

    @staticmethod
    def save_file(filename, data):
        with open(filename, 'w') as f:
            json.dump(data, f)

    @classmethod
    def from_dict(cls, json_dict):
        try:
            return cls(**json_dict)
        except TypeError:
            raise ValueError(
                '"{json_dict}" is NOT a valid zhihu token json.'.format(
                    json_dict=json_dict
                ))

class ZhihuClient:
    def __init__(self, username=None, passwd=None, token_file=TOKEN_FILE):
        self._session = requests.session()
        self._session.verify = False
        self.username = username
        self.passwd = passwd

        if os.path.exists(token_file):
            self._token = ZhihuToken.from_file(token_file)
        else:
            self._login_auth = ZhihuOAuth()
            json_dict = self.login()
            ZhihuToken.save_file(token_file, json_dict)
        self._session.auth = ZhihuOAuth(self._token)

    def login(self):
        data = LOGIN_DATA.copy()
        data['username'] = self.username
        data['password'] = self.passwd
        gen_login_signature(data)

        if self.need_captcha():
            captcha_image = self.get_captcha()
            with open(CAPTCHA_FILE, 'wb') as f:
                f.write(captcha_image)
            print('Please open {0} for captcha'.format(
                os.path.abspath(CAPTCHA_FILE)))

            captcha = input('captcha: ')
            os.remove(os.path.abspath(CAPTCHA_FILE))
            res = self._session.post(
                CAPTCHA_URL,
                auth=self._login_auth,
                data={'input_text': captcha}
            )
            try:
                json_dict = res.json()
                if 'error' in json_dict:
                    raise LoginException(json_dict['error']['message'])
            except (ValueError, KeyError) as e:
                raise LoginException('Maybe input wrong captcha value')

        res = self._session.post(LOGIN_URL, auth=self._login_auth, data=data)
        try:
            json_dict = res.json()
            if 'error' in json_dict:
                raise LoginException(json_dict['error']['message'])
            self._token = ZhihuToken.from_dict(json_dict)
            return json_dict
        except (ValueError, KeyError) as e:
            raise LoginException(str(e))

    def need_captcha(self):
        res = self._session.get(CAPTCHA_URL, auth=self._login_auth)
        try:
            j = res.json()
            return j['show_captcha']
        except KeyError:
            raise LoginException('Show captcha fail!')


if __name__ == '__main__':
    client = ZhihuClient('YOUR_USERNAME', 'YOUR_PASSWORD')

config.py

API_VERSION = '3.0.42'
APP_VERSION = '3.28.0'
APP_BUILD = 'release'
UUID = 'AJDA7XkI9glLBWc85sk-nJ_6F0jqALu4AlY='
UA = 'osee2unifiedRelease/3.28.0 (iPhone; iOS 10.2; Scale/2.00)'
APP_ZA = 'OS=iOS&Release=10.2&Model=iPhone8,1&VersionName=3.28.0&VersionCode=558&Width=750&Height='
CLIENT_ID = '8d5227e0aaaa4797a763ac64e0c3b8'
APP_SECRET = b'ecbefbf6b17e47ecb9035107866380'

TOKEN_FILE = 'token.json'

exception.py

class LoginException(Exception):
    def __init__(self, error):
        self.error = error

    def __repr__(self):
        return 'Login Fail: {}'.format(self.error)

    __str__ = __repr__

settings.py

ZHIHU_API_ROOT = 'https://api.zhihu.com'
LOGIN_URL = ZHIHU_API_ROOT + '/sign_in'
CAPTCHA_URL = ZHIHU_API_ROOT + '/captcha'

utils.py

import hashlib
import hmac
import time

from config import APP_SECRET


def gen_login_signature(data):
    data['timestamp'] = str(int(time.time()))

    params = ''.join([
        data['grant_type'],
        data['client_id'],
        data['source'],
        data['timestamp'],
    ])

    data['signature'] = hmac.new(
        APP_SECRET, params.encode('utf-8'), hashlib.sha1).hexdigest()

Hi，我是 zhihu-oauth 的作者~

今天在群里通过朋友分享看到了你公众号上的那篇文章，过感谢你帮着宣传了一波我的项目~

说起来你可能不信，我以前在专栏里写过一点项目介绍，然后被删了 =。= 所以说知乎是不是大度的公司还不一定呢……（参见这里）

另外开源软件重要的就是合作，虽然这个项目目前几乎都是我一个人开发的，但还是希望更多的人能够参与进来~

如果你对代码风格，项目组织，实现设计或者其他方面又什么建议的话不妨开个 Issue 一起讨论下 OvO，当然能协助下开发就更好啦。

毕竟作为一个还没毕业的学生，对代码和项目的理解都不太深，如果有人能帮忙指点一二也是很感激的。

以上。

再次感谢。

@7sDream, 被你发现了 ( *・ω・)✄╰ひ╯

看来Pythonista还是爬豆瓣比较安全...

平心而论，你写的非常好，无论当时甚至现在的我也未必能写出来这种带感的项目。Python开发者应该是各语言中最烦人的一类，成天在纠结代码规范、Pythonic或者Idiomatic的感觉，尤其是多了看多了之后渐渐的就会有自己的品味，。提个Issue倒不至于，至多算是对一些地方实现和处理上和我不同，这是非常主观的感觉，无所谓对错 - 编程的乐趣之一就是有无数种方法去完成一个目标。

说具体了, 就是感觉由于炫技增加了很多不环保的代码，也大幅的提高了项目的阅读复杂度。我比较喜欢简单粗暴，因为平时工作中别人的用了复杂的表达方法会增加我维护的成本，已所不欲勿施于人的原则。btw. 我以前也喜欢炫技，以别人看不懂，只能我来改为傲。