WebsiteBase.py

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

######################
# Author : 高明飞
# Data   : 2016-07-22
# Brief  : 用于获取网站新通知的爬虫基类
######################

import requests, re, time, sqlite3, json, logging
from bs4 import BeautifulSoup


class WebsiteBase:
    # Name : 网站名称
    # DBName ： 数据库名称，不要包含后缀
    # AgentID ： 微信发布时需要用到的AgentID
    # CheckContent ： 是否需要打开URL检查内容，True or False
    # KeyWords : 过滤用关键词List，如果不需要设置为[]
    # KeyWordsThreshold : 关键词阈值，内容页包含的关键词个数超过这个值才认为符合要求
    # SpecialKeyWords : 特殊关键词，不受阈值的影响，只要包含一个即符合要求
    # encoding ： 网站的编码格式，不设置的话默认为utf-8
    def __init__(self, Name, DBName, AgentID, CheckContent, KeyWords, KeyWordsThreshold, SpecialKeyWords = [], encoding = 'utf-8'):
        self.Name = Name
        self.DBName = DBName + '.db'
        self.DBCheckedName = DBName + '_checked.db'
        self.AgentID = AgentID
        self.CheckContent = CheckContent
        self.KeyWords = KeyWords
        self.encoding = encoding
        self.KeyWordsThreshold = KeyWordsThreshold
        self.SpecialKeyWords = SpecialKeyWords
        # Error Status
        self.err = 0

        # Wchat ID & Password
        f = open('wchat')
        self.corpid = f.readline().strip()
        self.corpsecret = f.readline().strip()
        f.close()

        # Init DB
        conn = sqlite3.connect(self.DBName)
        cursor = conn.cursor()
        try:
            cursor.execute(
                'create table Articles (Title TEXT, Brief TEXT, URL TEXT, DATE TEXT, Published INTEGER)')
        except:
            pass
        cursor.close()
        conn.commit()
        conn.close()
        # Use this DB to record checked messages.
        conn = sqlite3.connect(self.DBCheckedName)
        cursor = conn.cursor()
        try:
            cursor.execute('create table URL (URL TEXT)')
        except:
            pass
        cursor.close()
        conn.commit()
        conn.close()

    def GET(self):
        returnErr = None
        # Open DB
        conn = sqlite3.connect(self.DBName)
        cursor = conn.cursor()
        conn2 = sqlite3.connect(self.DBCheckedName)
        cursor2 = conn2.cursor()

        logging.warning('Getting : ' + self.Name + '......')

        PageRange = self.GetPageRange()

        for p in PageRange:
            logging.warning('  Getting : ' + str(p))

            try:
                time.sleep(3)
                response = self.GetMainPage(p)
                response.encoding = self.encoding
            except Exception as err:
                returnErr = err
                logging.error('    ' + str(p))
                logging.error('    ' + repr(err))
                continue

            soup = BeautifulSoup(response.text, 'html5lib')

            soup = self.GetEnclose(soup)
            tags = self.GetTags(soup)
            logging.warning('    Number of Pages : ' + str(len(tags)))

            for tag in tags:
                try:
                    # Get Title
                    Title = self.GetTitle(tag)
                    if not Title:
                        continue
                    logging.warning('      Checking : ' + Title)

                    # Get URL
                    ContentURL = self.GetURL(tag)

                    # Get time
                    PublishTime = self.GetPublishTime(tag)
                except Exception as err:
                    logging.error('      Format Error!')
                    logging.error('    ' + repr(err))
                    continue

                # Check DB
                cursor2.execute(
                    "select * from URL where URL = ?",
                    [(ContentURL)]
                )
                # Already exists
                if cursor2.fetchone():
                    continue

                # Addition Check
                if not self.AdditionCheck(tag):
                    continue

                if self.KeyWords:
                    # Check Title
                    # 标题仅检查特殊关键词，以此减少各种误报情况
                    specialflagcount = 0
                    keywordstring = ' 关键词：'
                    for keyword in self.SpecialKeyWords:
                        if (Title.count(keyword) > 0):
                            specialflagcount += Title.count(keyword)
                            keywordstring = keywordstring + keyword + '；'

                    if specialflagcount == 0 and not self.CheckContent:
                        # Next Title
                        continue

                    # Check Content
                    flagcount = 0
                    if self.CheckContent and ContentURL != '':
                        try:
                            time.sleep(3)
                            header = {
                                'User-Agent': 'Mozilla/5.0 (Windows NT 6.3; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/50.0.2661.102 Safari/537.36'}
                            response = requests.get(ContentURL, timeout=21, headers = header)
                            response.encoding = self.encoding
                        except Exception as err:
                            # returnErr = err
                            logging.error('    ' + ContentURL)
                            logging.error('    ' + repr(err))
                            continue

                        for keyword in self.KeyWords:
                            if (response.text.count(keyword)) > 0:
                                flagcount += response.text.count(keyword)
                                keywordstring = keywordstring + keyword + '；'
                                
                        for keyword in self.SpecialKeyWords:
                            if (response.text.count(keyword)) > 0:
                                specialflagcount += response.text.count(keyword)
                                keywordstring = keywordstring + keyword + '；'

                        # Update Checked DB
                        cursor2.execute(
                            "insert into URL (URL) values (?)",
                            [(ContentURL)]
                        )
                        if flagcount < self.KeyWordsThreshold and specialflagcount == 0:
                            # Next Title
                            continue

                # Get Brief
                if self.KeyWords:
                    BriefString = self.GetBrief(tag, str(flagcount) + keywordstring)
                else:
                    BriefString = self.GetBrief(tag, '')

                # Update DB
                cursor.execute(
                    "select * from Articles where Title = ? and Brief = ? and URL = ? and DATE = ?",
                    (Title, BriefString, ContentURL, PublishTime)
                )
                # Already exist
                if cursor.fetchone():
                    continue

                cursor.execute(
                    "insert into Articles (Title, Brief, URL, DATE, Published) values (?, ?, ?, ?, ?)",
                    (Title, BriefString, ContentURL, PublishTime, 0)
                )
                logging.warning('    Updating : ' + Title + '......')

        # Close DB
        cursor.close()
        conn.commit()
        conn.close()
        cursor2.close()
        conn2.commit()
        conn2.close()

        if returnErr:
            raise returnErr

    def Update(self):
        logging.warning('Updating : ' + self.Name + '......')

        # Init Wchat
        access_token = self.InitWchat()
        if not access_token:
            return

        # Open DB
        conn = sqlite3.connect(self.DBName)
        cursor = conn.cursor()

        cursor.execute("select * from Articles where Published = 0")
        unpublished = cursor.fetchall()

        for record in unpublished:
            # Publish
            logging.warning('  Publishing : ' + record[0] + '......')

            newsdata = {'touser': '@all',
                        'msgtype': 'news',
                        'agentid': self.AgentID,
                        'news': {'articles': [{
                            'title': record[0],
                            'description': record[1],
                            'url': record[2]
                        }]}}

            try:
                # time.sleep(1)
                r = requests.post('https://qyapi.weixin.qq.com/cgi-bin/message/send', params=access_token,
                                  data=json.dumps(newsdata, ensure_ascii=False).encode('utf-8'), timeout=21)
            except Exception as err:
                logging.error('    Publish Error!')
                logging.error('    ' + repr(err))
            else:
                if 'errcode' in r.json() and r.json()['errcode'] == 0:
                    logging.warning('    Publish Success!')
                    cursor.execute("update Articles set Published = 1 where Title = ? and URL = ? and DATE = ?",
                                   (record[0], record[2], record[3]))
                else:
                    logging.error('    Publish Error!')
                    logging.error('    ' + json.dumps(newsdata, ensure_ascii=False))
                    logging.error('    ' + r.json())

        # Close DB
        cursor.close()
        conn.commit()
        conn.close()

    def ReportErrStatus(self, errstr, errstate):
        access_token = self.InitWchat()
        if not access_token:
            return

        if errstate:
            newsdata = {'touser': 'g199209',
                        'msgtype': 'news',
                        'agentid': 0,
                        'news': {'articles': [{
                            'title': '云端程序错误',
                            'description': self.Name + ' :\r\n' + errstr,
                        }]}}
        else:
            newsdata = {'touser': 'g199209',
                        'msgtype': 'news',
                        'agentid': 0,
                        'news': {'articles': [{
                            'title': '云端程序正常运行',
                            'description': self.Name + ' :\r\n云端程序已从上次错误中恢复，现已正常运行~\r\n' +
                            '错误共发生' + str(self.err) + '次',
                        }]}}

        try:
            r = requests.post('https://qyapi.weixin.qq.com/cgi-bin/message/send', params=access_token,
                              data=json.dumps(newsdata, ensure_ascii=False).encode('utf-8'), timeout=21)
        except:
            logging.error('Send Wchat Report Error!')

    def InitWchat(self):
        # Init Wchat
        Auth = {'corpid': self.corpid,
                'corpsecret': self.corpsecret}
        try:
            r = requests.get('https://qyapi.weixin.qq.com/cgi-bin/gettoken', params=Auth, timeout=21)
        except Exception as err:
            logging.error('Wchat Init Timeout!!')
            logging.error(repr(err))
            return None
        else:
            if 'access_token' in r.json():
                return {'access_token': r.json()['access_token']}
            else:
                logging.error('Wchat Init Error!!')
                return None

    def GetPageRange(self):
        pass

    def GetMainPage(self, page):
        pass

    def GetEnclose(self, soup):
        pass

    def GetTags(self, soup):
        pass

    def GetTitle(self, tag):
        pass

    def GetURL(self, tag):
        pass

    def GetPublishTime(self, tag):
        pass

    def AdditionCheck(self, tag):
        pass

    def GetBrief(self, tag, keywordstring):
        pass