CSDN 已下载资源自动批量评论脚本

2014/10/12 Python Total: 6944 palavras, Tempo: 20 minutos de leitura

用 Python 实现自动批量打分评论指定 CSDN 账号内所有下载过待评论的资源。

GitHub 仓库地址:https://github.com/mzlogin/csdncommenter

可通过 pip 安装运行:

pip install csdncommenter
csdncommenter

背景

CSDN 账号过一段时间就会累积几十个下载过但是未评论打分的资源,虽然现在上传了一些资源供别人下载后基本不愁积分,但是为了可持续发展,还是把评论一下就能顺手拿了的这种积分不客气地收入囊中吧!不过手动一个一个去评论真的很蛋疼……特别是 CSDN 还设了两个评论间隔不能小于 60 秒、刚刚下载的资源十分钟内不能评论的限制,评论几十个就得至少花个几十分钟折腾,所以想想这种耗时、无脑的活还是交给程序来完成吧。

对于这类模拟 HTTP 请求然后可能频繁用到页面解析和正则表达式之类的活,用 C++ 写还是有点蛋疼的,用我那半生不熟的 Python 练练手正合适。

遂在 GitHub 上建了个仓库开工,地址在这里:https://github.com/mzlogin/csdncommenter

Update 2016/08/10:当前 CSDN 貌似已经取消了评论返积分的规则,我看了下我的得分记录,最近一次评论得分是在 2015/11/15。

分析

使用 Fiddler 把登录 - 到待评论页面 - 评论的完整流程抓了一下,整理程序逻辑大致如下:

注:如下 HTTP 请求均使用同一个 SESSION。

  1. 手动输入 CSDN 的用户名和密码。

  2. GET 方法从 https://passport.csdn.net/account/login 页面获取 ltexecution_eventId 等参数。

  3. 将第 1 步中的用户名和密码,还有第 2 步中得到的参数 POST 给 https://passport.csdn.net/account/login ,从 Response 中判断是否登录成功——我采用的依据是 status_code 为 200 且 Reponse 内容中有 lastLoginIP

  4. GET 方法从 http://download.csdn.net/my/downloads 页面获取已下载资源总页数。从最后一个 pageliststyhref 中得到。

  5. 根据第 4 步中得到的总页数,根据每个页面 num 拼得 url 为 http://download.csdn.net/my/downloads/num ,使用 GET 方法访问之拿到该页面中所有待评论资源 ID。从所有 class="btn-comment"a 标签的 href 中得到。

  6. 在进行第 5 步的过程中,如果 num 为 1 的页面里有 <span class="btn-comment"> 存在,那说明存在 10 分钟以内下载,暂时不能评论的资源,这时循环检查最多 11 次,每次检查完如果发现还需要等待就过一分钟再检查,进到不再需要等待或者超过 11 次为止。

  7. 对第 5 步中得到的所有待评论资源 ID 依次进行间隔至少 60S 的打分评论,根据资源的现有评星打分,不对资源评分造成不良影响。根据打出的 1 到 5 星,对应一句英文短句评论。出乎我意料的是评论这一步竟然也是用 GET 就可以做,http://download.csdn.net/index.php/comment/post_comment 后面带上 sourceidcontent(评论内容)、rating(打分)和 t(时间戳)参数就可以。评论成功会返回 ({"succ":1}),失败会返回「两次评论需要间隔 60 秒」、「您已经发表过评论」等之类的 msg

    获取资源的现有评星的方法是从 http://download.csdn.net/detail/<资源所有者 ID>/<资源 ID> 页面获取 <span style="width:75px" class="star-yellow"></span> 这一段内容,其中的 75px 表示为五星,如果是 0px 表示为零星,即每加一星增加 15px。

最终运行截图如下:

CSDN 自动批量打分评论

确认这种方式能有效拿到 CSDN 的分数:

CSDN 自动评论得分

总结

  1. 用 Python 干这种类型的活还是很有优势的,requests 和 BeautifulSoup 简直神器啊!
  2. 我那点蹩脚的 Python 底子之所以能还比较顺利地把这个流程写下来,实际上也得亏 CSDN 对请求的验证相对较松,比如像我代码里那样写, User-Agent 是带有 Python 字样的,而且很显然不是浏览器在访问,但 CSDN 并未对此作限制。

源码

没有找到从 Github Pages 引用 Github 仓库里的源码的方法,所以把 py 文件放到一个 gist 里了,引用如下

(Gist 前几天被伟大的墙封了,还是直接贴上代码吧。2014/11/5 update)

(GitHub 仓库:mzlogin/csdncommenter,现在可以通过 pip 安装使用了 pip install csdncommenter 然后 csdncommenter。2015/10/27 update)

# File   : csdncommenter.py
# Author : Zhuang Ma
# E-mail : chumpma(at)gmail.com
# Website: https://mazhuang.org
# Date   : 2016-07-26
import requests
from BeautifulSoup import BeautifulSoup
import getpass
import time
import random
import re
import urllib
import traceback

class CsdnCommenter():
    """Csdn operator"""
    def __init__(self):
        self.sess = requests.Session()

    def login(self):
        """login and keep session"""
        username = raw_input('username: ')
        password = getpass.getpass('password: ')
        url = 'https://passport.csdn.net/account/login'
        html = self.getUrlContent(self.sess, url)
        if html is None:
            return False
        soup = BeautifulSoup(html)

        lt = self.getElementValue(soup, 'name', 'lt')
        execution = self.getElementValue(soup, 'name', 'execution')
        _eventId = self.getElementValue(soup, 'name', '_eventId')

        data = {
                'username' : username,
                'password' : password,
                'lt' : lt,
                'execution' : execution,
                '_eventId' : _eventId
                }

        response = None
        try:
            response = self.sess.post(url, data)
        except:
#             traceback.print_exc()
            pass

        return self.isLoginSuccess(response)

    def autoComment(self):
        """main handler"""
        if self.getSourceItems() is False:
            print('No source can comment!')
            return

        print('Total %d source(s) wait for comment.' % len(self.sourceitems))

        nhandled = 0
        for sourceid in self.sourceitems.keys():
            left = len(self.sourceitems) - nhandled

            sec = random.randrange(61,71)
            print('Wait %d seconds for start. %s source(s) left.' % (sec, left))
            time.sleep(sec)

            self.comment(sourceid)
            nhandled += 1

        print('Finished!')

    def getSourceItems(self):
        """get (sourceid,username) couples wait for comment"""
        self.sourceitems = dict()
        pagecount = self.getPageCount()
        if pagecount == 0:
            return False

        print('Pagecount is %d.' % pagecount)

        pattern = re.compile(r'/detail/([^/]+)/(\d+)#comment')

        for n in range(1, pagecount + 1):
            if n == 1:
                self.waitUncommentableSourceIfNecessary()
            url = 'http://download.csdn.net/my/downloads/%d' % n
            html = self.getUrlContent(self.sess, url)
            if html is None:
                continue
            soup = BeautifulSoup(html)
            sourcelist = soup.findAll('a', attrs={'class' : 'btn-comment'})
            if sourcelist is None or len(sourcelist) == 0:
                continue
            for source in sourcelist:
                href = source.get('href', None)
                if href is not None:
                    rematch = pattern.match(href)
                    if rematch is not None:
                        self.sourceitems[rematch.group(2)] = rematch.group(1)

        return len(self.sourceitems) > 0

    def waitUncommentableSourceIfNecessary(self):
        """souce cannot comment within 10 minutes after download"""
        url = 'http://download.csdn.net/my/downloads/1'
        maxMinutes = 11
        for i in range(0, maxMinutes):
            html = self.getUrlContent(self.sess, url)
            if html is None:
                break
            soup = BeautifulSoup(html)
            sourcelist = soup.findAll('span', attrs={'class' : 'btn-comment'})
            if sourcelist is None or len(sourcelist) == 0:
                print('None uncommentable source now!')
                break
            print('Waiting for uncommentable source count down %d minutes.' % (maxMinutes-i))
            time.sleep(60)

    def getPageCount(self):
        """get downloaded resources page count"""
        url = 'http://download.csdn.net/my/downloads'
        html = self.getUrlContent(self.sess, url)
        if html is None:
            print('Get pagecount failed')
            return 0
        soup = BeautifulSoup(html)

        pagelist = soup.findAll('a', attrs={'class' : 'pageliststy'})
        if pagelist is None or len(pagelist) == 0:
            return 0

        lasthref = pagelist[len(pagelist) - 1].get('href', None)
        if lasthref is None:
            return 0
        return int(filter(str.isdigit, str(lasthref)))

    def comment(self, sourceid):
        """comment per source"""
        print('sourceid %s commenting...' % sourceid)
        contents = [
                'It just soso, but thank you all the same.',
                'Neither good nor bad.',
                'It is a nice resource, thanks for share.',
                'It is useful for me, thanks.',
                'I have looking this for long, thanks.'
                ]
        rating = self.getSourceRating(sourceid)
        print('current rating is %d.' % rating)
        if rating == 0: # nobody comments
            rating = 3
        content = contents[rating - 1]
        t = '%d' % (time.time() * 1000)

        paramsmap = {
                'sourceid' : sourceid,
                'content' : content,
                'rating' : rating,
                't' : t
                }
        params = urllib.urlencode(paramsmap)
        url = 'http://download.csdn.net/index.php/comment/post_comment?%s' % params
        html = self.getUrlContent(self.sess, url)
        if html is None or html.find('({"succ":1})') == -1:
            print('sourceid %s comment failed! response is %s.' % (sourceid, html))
        else:
            print('sourceid %s comment succeed!' % sourceid)

    def getSourceRating(self, sourceid):
        """get current source rating"""
        rating = 3
        url = 'http://download.csdn.net/detail/%s/%s' % (self.sourceitems[sourceid], sourceid)
        html = self.getUrlContent(self.sess, url)
        if html is None:
            return rating

        soup = BeautifulSoup(html)
        ratingspan = soup.findAll('span', attrs={'class': 'star-yellow'})
        if ratingspan is None or len(ratingspan) == 0:
            return rating

        ratingstyle = ratingspan[0].get('style', None)

        if ratingstyle is None:
            return rating

        rating = int(filter(str.isdigit, str(ratingstyle))) / 15
        return rating


    @staticmethod
    def getElementValue(soup, element_name, element_value):
        element = soup.find(attrs={element_name : element_value})
        if element is None:
            return None
        return element.get('value', None)

    @staticmethod
    def isLoginSuccess(response):
        if response is None or response.status_code != 200:
            return False
        return -1 != response.content.find('lastLoginIP')

    @staticmethod
    def getUrlContent(session, url):
        html = None
        try:
            response = session.get(url)
            if response is not None:
                html = response.text
        except requests.exceptions.ConnectionError as e:
#             traceback.print_exc()
            pass

        return html

def main():
    csdn = CsdnCommenter()
    while csdn.login() is False:
        print('Login failed! Please try again.')
    print('Login succeed!')

    csdn.autoComment()

if __name__ == '__main__':
    main()

Informações:

Search

    Tabela de conteudos