def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
#判断url格式，因为从百度查询所得不一定是question，目前只爬question
if url.find('question') < 0:
self.logger.warn('Question supported only:%s', url)
return None
article_id = re.findall(r'question/(\d+)', url)[0]
self.session.randomSleep()
response = self.session.get(url, headers=CRAWL_ARTICLE_HEADERS)
soup = BeautifulSoup(response)
main = soup.find('div', attrs={'id': "data"}).attrs['data-state']
articleJson = json.loads(main)
questionJson = articleJson['entities']['questions'][article_id]
title = questionJson['title']
contentSoup = BeautifulSoup(questionJson['editableDetail'])
content = contentSoup.text
author_id = questionJson['author']['id']
author_name = questionJson['author']['name']
createTimeInFloat = questionJson['created']
publish_datetime = time.strftime('%Y-%m-%d %H:%M:%S',
time.localtime(createTimeInFloat))
reply_count = questionJson['commentCount']
read_count = questionJson['visitCount']
collect_count = questionJson['followerCount']
article = Article(article_id, self.channel.channel_id, title, content,
publish_datetime, url, author_id, author_name)
article.statistics.reply_count = reply_count
article.statistics.read_count = read_count
article.statistics.collect_count = collect_count
return article
def __filterRemovedArticle(self, articleList, entityId, eventId=None):
'''
与remove表格对比，进行文章过滤
返回不存在remove表中的文章list
'''
if len(articleList) == 0:
return []
if eventId is not None:
tableName = Constants.TABLE_SA_EVENT_ARTICLE_REMOVE + Constants.TABLE_NAME_DELIMITER + entityId
eventCondition = ' event_id=%d and ' % eventId
else:
tableName = Constants.TABLE_SA_ARTICLE_REMOVE + Constants.TABLE_NAME_DELIMITER + entityId
eventCondition = ''
# 在remove表里查找文章
selectSql = '''
SELECT TID, CHANNEL_ID FROM %s where %s (%s)
'''
whereClauseList = map(
lambda article: '(TID="%s" and CHANNEL_ID=%d)' %
(article.tid, article.channel_id), articleList)
self.dbProxy.execute(
selectSql %
(tableName, eventCondition, ' or '.join(whereClauseList)))
resultList = self.dbProxy.fetchall() # 查询返回结果集
removedArticleList = map(lambda x: Article(x[0], x[1]), resultList)
filteredArticle = filter(lambda x: x not in removedArticleList,
articleList)
#self.logger.debug('originalList:%s', map(lambda article: article.tid, articleList))
#self.logger.debug('removedArticleList:%s', map(lambda article: article.tid, removedArticleList))
#self.logger.debug('filteredArticle:%s', map(lambda article: article.tid, filteredArticle))
return filteredArticle
def __searchByPage(self,keywordList,startTime,endTime,page):
# 在豆瓣内部搜索框搜索
page = str(page) #url接收的是str格式
search_url = "https://www.douban.com/group/search?start="+page+"&cat=1013&sort=time&q="
#cat:按话题搜索 sort：按最新发布时间分类 q:搜索关键词
query = urllib.quote(' '.join(keywordList).encode('utf-8'))
search_url = search_url+str(query)
url_page = self.session.get(search_url,headers=self.headers)
soup = BeautifulSoup(url_page, "lxml")
# self.logger.debug(soup)
main_wrap = soup.find('div', attrs={'class': "article"})
main_article_list = main_wrap.find('div',attrs={'class':"topics"})
articleList = list()
hasnext = True
if main_article_list is not None:
title_list = main_article_list.findAll('tr', {'class': 'pl'})
for title in title_list:
article_publishtime = title.find('td', attrs={'class': "td-time"}).attrs['title']
urlTime = time.strptime(article_publishtime, "%Y-%m-%d %H:%M:%S")
Y, M, D, H = urlTime[0:4]
urlTime2 = datetime.datetime(Y, M, D, H)
urlTime2 = time.mktime(urlTime2.timetuple())
#转换成时间戳来比较 float类型
startTime = endTime - datetime.timedelta(days=2)
startTimeIntSecond = time.mktime(startTime.timetuple())
endTimeIntSecond = time.mktime(endTime.timetuple())
#如果符合时间范围就爬取
if urlTime2 >= startTimeIntSecond and urlTime2 <= endTimeIntSecond:
article_url = title.find('a').attrs['href']
self.logger.debug(article_url)
(content, author_name, tid,like_count) = self.crawlArticle(article_url)
if content is None: # 话题已被删除或则其他格式
continue
article_title = title.find('td', attrs={'class': "td-subject"}).text
article = Article(tid, self.channel.channel_id, article_title, content, article_publishtime, article_url, None,
author_name)
self.crawlComment(article)
reply_count = title.find('td', attrs={'class': "td-reply"}).text.strip()
reply_count = re.sub(u'回应', '', reply_count) # 回复数量去除中文保留数字
article.statistics.reply_count = reply_count
article.statistics.like_count = like_count
if article not in articleList:
articleList.append(article)
else:
print len(articleList)
hasnext=False
break
return (articleList, hasnext)
def crawlArticle(self, url, **kwargs):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
# TBD, 转发情况目前先不考虑
searchTitle = ''
referer = None
if (kwargs['flag'] == 0):
if 'referer' not in kwargs:
return None
if 'title' in kwargs:
searchTitle = kwargs['title']
else:
searchTitle = ''
try:
self.session_public_article.randomSleep()
if kwargs['flag'] == 0:
referer = kwargs['referer']
headers = HEADERS_ARTICLE.copy()
headers['Referer'] = referer
response = self.session_public_article.get(
url, textRspOnly=False, headers=HEADER_PUBLIC_ARTICLE)
else:
response = self.session_public_article.get(
url, textRspOnly=False, headers=HEADER_PUBLIC_ARTICLE)
mid = re.findall(r'var mid = .*"(\d+)";',
response.text)[0] + '-' + re.findall(
r'var idx = .*"(\d+)";', response.text)[0]
soup = BeautifulSoup(response.text, 'lxml')
main = soup.find('div', {'id': "img-content"})
title = main.find('h2').text.strip()
content = main.find('div', {'id': "js_content"}).text.strip()
profile = main.find('div', {'class': "profile_inner"})
author_id = profile.find('span').text.strip()
author_name = profile.find('strong').text.strip()
publish_datetime = kwargs['publish_datetime']
article = Article(mid,
self.channel.channel_id,
title,
content,
publish_datetime=publish_datetime,
url=url,
author_id=author_id,
author_name=author_name,
meta_info='{refer="%s"}' % referer)
self.logger.debug('Successfully parse article:%s', title)
return article
except:
self.logger.error('Fail to get article for %s: %s due to %s', url,
searchTitle, traceback.format_exc())
return None
def __fetchOldArticleList(self, articleList, articleCount=100):
'''
从全局文章表，获取尚未消亡的文章id，而且这些文章并不在本次爬虫爬回来的记录里
'''
#用来查询总页数
selectSql_count = 'SELECT COUNT(*) FROM %s where extinct="N" and channel_id=%d '
sql2 = selectSql_count % (Constants.TABLE_SA_ARTICLE,
self.channel.channel_id)
#获取旧文章的sql
selectSql = 'SELECT TID,title, publish_datetime,url, meta_info FROM %s where extinct="N" and channel_id=%d '
sql = selectSql % (Constants.TABLE_SA_ARTICLE, self.channel.channel_id)
if len(articleList) > 0:
whereClauseList = map(
lambda article: ' tid<>"%s" ' % (article.tid), articleList)
sql += ' and (%s)' % (' and '.join(whereClauseList))
sql2 += ' and (%s)' % (' and '.join(whereClauseList))
sql2 += ' order by add_datetime desc'
self.dbProxy.execute(sql2)
resultList2 = self.dbProxy.fetchall()
# print '12456789sssssssssssssssssss'
# print resultList2 #((53,),)
resultList2 = re.findall(r'\d+', str(resultList2)) #返回一个list
# print resultList2[0]
if int(resultList2[0]) > int(articleCount):
randpage = random.randint(
0, int(math.ceil(float(resultList2[0]) / articleCount)))
else:
randpage = 0 #用来随机取数据库页数
sql += ' order by add_datetime desc limit %d,%d' % (randpage,
articleCount)
self.dbProxy.execute(sql)
resultList = self.dbProxy.fetchall()
return map(
lambda item: Article(item[0],
self.channel.channel_id,
title=item[1],
publish_datetime=item[2],
url=item[3],
meta_info=item[4]), resultList)
def __filterRemovedArticle(self, articleList, entityId, eventId=None):
'''
与remove表格对比，进行文章过滤
返回不存在remove表中的文章list
'''
if len(articleList) == 0:
return []
if eventId is not None:
tableName = Constants.TABLE_SA_EVENT_ARTICLE_REMOVE + Constants.TABLE_NAME_DELIMITER + entityId
eventCondition = ' event_id=%d and ' % eventId
start_datetime, end_datetime = self.__fetchEventTime(
entityId, eventId)
# 过滤掉不在该事件开始时间和结束之间内的文章
article_new_list = list()
for article in articleList:
if (str(article.publish_datetime) > str(start_datetime)) and (
str(article.publish_datetime) < str(end_datetime)):
article_new_list.append(article)
articleList = article_new_list
else:
tableName = Constants.TABLE_SA_ARTICLE_REMOVE + Constants.TABLE_NAME_DELIMITER + entityId
eventCondition = ''
# 在remove表里查找文章
selectSql = '''
SELECT TID, CHANNEL_ID FROM %s where %s (%s)
'''
whereClauseList = map(
lambda article: '(TID="%s" and CHANNEL_ID=%d)' %
(article.tid, article.channel_id), articleList)
self.dbProxy.execute(
selectSql %
(tableName, eventCondition, ' or '.join(whereClauseList)))
resultList = self.dbProxy.fetchall() # 查询返回结果集
removedArticleList = map(lambda x: Article(x[0], x[1]), resultList)
filteredArticle = filter(lambda x: x not in removedArticleList,
articleList)
return filteredArticle
def __seperateNewOldArticles(self, articleList, entityId=None):
'''
查询全局文章表，区分新文章和旧文章
'''
if len(articleList) == 0:
return ([], [])
if entityId is None:
selectSql = 'select tid, channel_id from %s where ' % Constants.TABLE_SA_ARTICLE
else:
selectSql = 'select tid, channel_id from %s where ' % (
Constants.TABLE_SA_ARTICLE + Constants.TABLE_NAME_DELIMITER +
entityId)
whereClauseList = map(
lambda article: '(tid="%s" and channel_id=%d)' %
(article.tid, article.channel_id), articleList)
self.dbProxy.execute(selectSql + ' or '.join(whereClauseList))
resultList = map(lambda x: Article(x[0], x[1]),
self.dbProxy.fetchall())
existingArticleList = filter(lambda x: x in resultList, articleList)
newArticleList = filter(lambda x: x not in resultList, articleList)
return (existingArticleList, newArticleList)
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
:param url:
:return:返回一个Article实例
'''
cookies = None
html = self.session.download(url=url,
encoding='utf-8',
data=None,
timeout=10,
retry=3,
addr=True,
cookies=cookies)
if html:
article_url = html['url']
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
article_id = re.findall(r'c_\d+', article_url)[0]
article_id = article_id[2:]
soup = BeautifulSoup(html['html'], "html.parser")
main1 = soup.find('div', attrs={'class': "widthMain main"})
main2 = soup.find('div', attrs={'class': "main pagewidth"})
main3 = soup.find('body', attrs={'class': "streamline-page"})
main4 = soup.find('div', attrs={'class': "h-title"})
main5 = soup.find('div', attrs={'id': "article"})
main6 = soup.find('div', attrs={'id': "Title"})
main7 = soup.find('div', attrs={'class': "article"})
if main1 is not None:
self.logger.debug("main1")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain1(main1)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main2 is not None:
self.logger.debug("main2")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain2(main2)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main3 is not None:
self.logger.debug("main3")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain3(main3)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main3 is None and main4 is not None:
self.logger.debug("main4")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain4(soup)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main5 is not None:
self.logger.debug("main5")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain5(main5)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main6 is not None:
self.logger.debug("main6")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain6(soup)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if main7 is not None:
self.logger.debug("main7")
Ttitle, Ttime, Tauthor, Tcontent = self.crawlMain7(soup)
if (Ttitle != None and Ttime != None and Tcontent != None):
article = Article(tid=article_id,
channel_id=self.channel.channel_id,
title=Ttitle,
author_name=Tauthor,
publish_datetime=Ttime,
content=Tcontent,
url=article_url)
return article
if (main1 is None and main2 is None and main3 is None
and main4 is None and main5 is None and main6 is None
and main7 is None):
self.logger.error(u"存在另外一种html格式：：%s", url)
return
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
article = None
html = self.session.download(url,
encoding='utf-8',
data=None,
isJson=False,
timeout=10,
retry=3,
addr=True)
article_url = html['url']
# if article_url.find(self.channel.url)<0:
# self.logger.warn('Unrelated url found:%s',url)
# continue
# self.logger.debug(article_url)
soup = BeautifulSoup(html['html'], 'html.parser')
noResultDivList = soup.findAll('div', {'class': 'pl_noresult'})
if len(noResultDivList) > 0:
hasnext = False
self.logger.info('No result')
return article
main = soup.find('div', attrs={'class': "left_section"})
if main:
Ttitle = main.find('div', attrs={
'id': "j_core_title_wrap"
}).find('h1')
Ttitle1 = main.find('div', attrs={
'id': "j_core_title_wrap"
}).find('h3')
if Ttitle:
Ttitle = Ttitle.text.strip()
elif Ttitle1:
Ttitle = Ttitle1.text.strip()
else:
Ttitle = ''
# self.logger.debug(u'标题%s',Ttitle)
data_field = main.find('div', attrs={
'id': "j_p_postlist"
}).find('div').attrs['data-field'].strip()
data_field = json.loads(data_field)
publish_datetime = data_field['content']
if 'date' in publish_datetime.keys():
publish_datetime = publish_datetime['date']
else:
publish_datetime = main.find('div',
attrs={
'id': "j_p_postlist"
}).find('div').find_all(
'span',
attrs={'class': "tail-info"
})[-1].text.strip()
publish_datetime = re.findall(
r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2}(:\d{2})?)',
publish_datetime)[0]
if len(publish_datetime[4]) > 1:
publish_datetime = publish_datetime[
0] + '-' + publish_datetime[1] + '-' + publish_datetime[
2] + ' ' + publish_datetime[3] + publish_datetime[4]
else:
publish_datetime = publish_datetime[
0] + '-' + publish_datetime[1] + '-' + publish_datetime[
2] + ' ' + publish_datetime[3] + ':00'
Tid = data_field['author']['user_id']
Tauthor = data_field['author']['user_name']
Treply = soup.find('li', attrs={
'class': "l_reply_num"
}).find('span').text.strip() #总回复数
Tcontent = main.find('div', attrs={
'id': "j_p_postlist"
}).find('div').find('cc').text.strip()
article = Article(Tid,
self.channel.channel_id,
Ttitle,
Tcontent,
publish_datetime,
url=article_url,
author_id=None,
author_name=Tauthor,
meta_info=None)
article.statistics.reply_count = Treply
else:
self.logger.warn(u'很抱歉，该贴已被删除。%s', article_url)
return article
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
html = self.session.download(url,
encoding='utf-8',
data=None,
isJson=False,
timeout=10,
retry=3,
addr=True)
article_url = html['url']
self.logger.debug(article_url)
soup = BeautifulSoup(html['html'], 'html.parser')
main = soup.find('div', attrs={'id': "bd"})
main1 = soup.find('div', attrs={
'class': "wd-question"
}) #论坛提问帖子http://bbs.tianya.cn/post-730-5795-1-1.shtml
article = None
if main:
Ttitle = main.find('h1').find('span').text
Ttime = main.find('div', attrs={
'class': "atl-info"
}).find_all('span')[1].text.strip()
Ttime = re.sub(u'[\u4e00-\u9fa5]+：', '', Ttime)
Tid = main.find('div', attrs={
'class': "atl-info"
}).find_all('span')[0].find('a').attrs['uid'].strip()
Tauthor = main.find('div', attrs={
'class': "atl-info"
}).find_all('span')[0].find('a').attrs['uname'].strip()
Tclick = main.find('div', attrs={
'class': "atl-info"
}).find_all('span')[2].text.strip()
Tclick = re.sub(u'[\u4e00-\u9fa5]+：', '', Tclick)
Tclick = int(Tclick)
Treply = main.find('div', attrs={
'class': "atl-info"
}).find_all('span')[3].text.strip()
Treply = re.sub(u'[\u4e00-\u9fa5]+：', '', Treply)
Treply = int(Treply)
Tlike = main.find('a', attrs={'class': "zantop"}).attrs['_count']
Tcontent = main.find('div',
attrs={
'class': "bbs-content clearfix"
}).text.strip()
article = Article(Tid,
self.channel.channel_id,
Ttitle,
Tcontent,
Ttime,
url=article_url,
author_id=None,
author_name=Tauthor)
article.statistics.reply_count = Treply
article.statistics.click_count = Tclick
article.statistics.like_count = Tlike
elif main1:
Ttitle = main1.find('h1').find('span').text
Ttime = main1.find('div').attrs['js_replytime']
Ttime = re.findall(r'(\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2})',
Ttime)[0]
Tid = main1.find('div').attrs['_host']
Tauthor = main1.find('div', attrs={
'class': "q-info"
}).find('a').text
Tclick = main1.find('div').attrs['js_clickcount']
Treply = main1.find('div').attrs['js_powerreply']
Tcontent = main1.find('div', attrs={'class': "q-content atl-item"})
if Tcontent:
Tcontent = Tcontent.find('div', attrs={
'class': "text"
}).text.strip()
else:
Tcontent = ''
article = Article(Tid,
self.channel.channel_id,
Ttitle,
Tcontent,
Ttime,
url=article_url,
author_id=None,
author_name=Tauthor)
article.statistics.reply_count = Treply
article.statistics.click_count = Tclick
return article
def crawlArticle(self, req_data):
hasnext = False
req_url = 'http://120.55.151.61/Treehole/V4/Cave/getList.action'
data_r = self.s.post(url=req_url, data=req_data, headers=self.headers)
data_r = data_r.text
true = True
false = False
data_j = eval(data_r)
data_js = json.dumps(data_j)
data_dict = simplejson.loads(data_js)
data = data_dict['data']
timestamp = data['timestampLong']
timestampLong = data['timestampLong'] / 1000
timestampLong_struct = datetime.datetime.fromtimestamp(timestampLong)
timestampLong_str = timestampLong_struct.strftime('%Y-%m-%d %H:%M:%S')
messageBO = data['messageBOs']
for each in messageBO: # 20条记录保存在一组里
if 'studentBO' in each:
content = each.get('content', False)
if content:
publish_timeStamp = each['issueTime'] / 1000
datetime_struct = datetime.datetime.fromtimestamp(
publish_timeStamp)
publish_time = datetime_struct.strftime(
'%Y-%m-%d %H:%M:%S')
if datetime.datetime.strptime(
publish_time,
'%Y-%m-%d %H:%M:%S') < self.startTime:
hasnext = False
break
schoolName = each['schoolName']
readCount = each['readCount']
relayedCount = each['relayedCount']
reply_count = each['comments']
content = each['content']
if len(content) > 8:
title = content[:8] + '...'
elif len(content) <= 0:
title = u'内容为图片'
else:
title = content
studentBO = each['studentBO']
studentBO_id = studentBO['studentId']
studentBO_name = studentBO['nickName']
likeCount = each['likeCount']
article_id = each['messageId']
publish_method = each['source']
publish_method = publish_method[publish_method.find('_') +
1:]
data2 = 'platform=1&phoneBrand=HONOR&phoneVersion=24&versionNumber=9.1.1&phoneModel=BLN-AL10&sharePlatform=4&channel=huaweiMarket&plateId=1&messageId=' + str(
article_id) + '&'
article_urlPOST = 'http://120.55.151.61/Treehole/V4/Message/preShare.action'
data_urlPOST = self.s.post(url=article_urlPOST,
data=data2,
headers=self.headers)
data3 = data_urlPOST.text
data_j3 = eval(data3)
data_js3 = json.dumps(data_j3)
data_dict3 = simplejson.loads(data_js3)
data4 = data_dict3['data']
article_url = data4['shareUrl']
article = Article(article_id,
self.channel.channel_id,
title,
content,
publish_time,
article_url,
author_id=studentBO_id,
author_name=studentBO_name,
publish_method=publish_method,
entity=schoolName)
article.statistics.reply_count = reply_count
article.statistics.like_count = likeCount
article.statistics.read_count = readCount
article.statistics.forward_count = relayedCount
# print article.__str__()
if article is not None and article not in self.articleList:
self.articleList.append(article)
# print datetime.datetime.strptime(timestampLong_str,'%Y-%m-%d %H:%M:%S')
# print '1111111'
if datetime.datetime.strptime(timestampLong_str,
'%Y-%m-%d %H:%M:%S') < self.startTime:
hasnext = False
return ('', hasnext)
else:
hasnext = True
res_data = 'timestamp=' + str(
timestamp
) + '&preMoodTimestap=' + str(
timestamp
) + '&platform=1&phoneBrand=HONOR&phoneVersion=24&versionNumber=9.1.1&phoneModel=BLN-AL10&type=1&channel=huaweiMarket&'
return (res_data, hasnext)
def __searchByPage(self, keywordList, startTime, endTime, page):
# 在豆瓣内部搜索框搜索
page = str(page) #url接收的是str格式
query = urllib.quote(' '.join(keywordList).encode('utf-8'))
params = {
'c': 'blog',
'range': 'article', # 包含全部
'by': 'all',
'sort': 'time',
'col': '',
'source': '',
'from': '',
'country': '',
'size': '',
'time': '',
'a': '',
'isown': 'false',
'page': page,
'dpc': '',
'q': query
}
search_url = "http://search.sina.com.cn/"
url_page = self.session.get(search_url,
params=params,
headers=self.headers)
soup = BeautifulSoup(url_page, "lxml")
# self.logger.debug(soup)
main_wrap = soup.find('div', attrs={'class': "result-boxes"})
if main_wrap is None: #为了防止因网络问题而导致的查找不到
self.logger.debug(u'第一次查找没有结果再找一遍中')
url_page = self.session.get(search_url,
params=params,
headers=self.headers)
soup = BeautifulSoup(url_page, "lxml")
main_wrap = soup.find('div', attrs={'class': "result-boxes"})
if main_wrap is None: #再找一遍如果还是没有才可以算没有结果
self.logger.debug(u'第二次查找没有结果，该关键词没有结果')
return (list(), False)
main_article_list = main_wrap.findAll(
'div', attrs={'class': "box-result clearfix"})
hasnext = True
if main_article_list is not None:
print '1'
for title in main_article_list:
print '2'
article_publishtime_wrap = title.findAll('p')[1]
article_publishtime = article_publishtime_wrap.find(
'span', attrs={
'class': "fgray_time"
}).text.strip()
self.logger.error(article_publishtime)
urlTime = time.strptime(article_publishtime,
"%Y-%m-%d %H:%M:%S")
Y, M, D, H = urlTime[0:4]
urlTime2 = datetime.datetime(Y, M, D, H)
urlTime2 = time.mktime(urlTime2.timetuple())
#转换成时间戳来比较 float类型
startTime = endTime - datetime.timedelta(days=5)
startTimeIntSecond = time.mktime(startTime.timetuple())
endTimeIntSecond = time.mktime(endTime.timetuple())
#如果符合时间范围就爬取
if urlTime2 >= startTimeIntSecond and urlTime2 <= endTimeIntSecond:
print '3'
title_wrap = title.find('h2')
article_url_parent = title_wrap.find('a').attrs['href']
article_url_index = article_url_parent.rfind('/')
article_url_child = article_url_parent[article_url_index +
1:]
article_url = 'http://blog.sina.com.cn/s/blog_' + article_url_child + '.html'
Ttitle = title_wrap.find('a').text.strip()
self.logger.debug(article_url)
meta_info = article_url_child
(content, reply_count, like_count, read_count,
collect_count, forward_count) = self.crawlArticle(
article_url, meta_info)
self.logger.debug(like_count)
if content is None: # 话题已被删除或则其他格式
print '756236'
continue
tid = 't_' + article_url_child
author_name = article_publishtime_wrap.find(
'a', attrs={
'class': "rib-author"
}).text
article = Article(tid,
self.channel.channel_id,
Ttitle,
content,
article_publishtime,
article_url,
None,
author_name,
meta_info=meta_info)
# self.crawlComment(article)
article.statistics.reply_count = reply_count
article.statistics.like_count = like_count
article.statistics.read_count = read_count
article.statistics.collect_count = collect_count
article.statistics.forward_count = forward_count
# self.logger.debug(article)
if article not in self.articleList:
self.articleList.append(article)
else:
print '78956'
print len(self.articleList)
hasnext = False
break
return hasnext
class JianShuCrawler(object):
def __init__(self, channel, logger=None):
'''
:param channel:
:param logger:
'''
if logger is None:
self.logger = Logging.getLogger(Logging.LOGGER_NAME_DEFAULT)
else:
self.logger = logger
self.channel = channel #获取媒体实体
# 设置请求头和代理
self.headers = {
'Accept':
'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
'Accept-Encoding':
'gzip, deflate, br',
'Accept-Language':
'zh-CN,zh;q=0.9',
'Host':
'www.jianshu.com',
'User-Agent':
'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36',
}
self.headers_1 = {
'Host': 'www.jianshu.com',
'Connection': 'keep-alive',
'Connection-Length': '0',
'Accept': 'application/json',
'User-Agent':
'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/67.0.3396.99 Safari/537.36',
'Accept-Encoding': 'gzip,deflate,br',
'Accept-Language': 'zh-CN,zh;q=0.9'
}
#用户信息
self.usename = '13432881156'
self.password = '******'
#用cookie进行登录
self.session = SessionCrawler()
self.session.cookies = cookielib.LWPCookieJar(
filename='cookie') #读取cookie
try:
self.session.cookies.load(ignore_discard=True)
except:
print('未能加载cookie')
if self.islogin():
print('已经登录简书')
else:
self.login(self.usename, self.password)
def searchArticle(self, keywordList, endTime):
'''
根据关键字数组，开始时间和结束时间范围搜索文章
:param keywordList: 关键字数组
:param endTime: 搜索时间范围结束
:return:
'''
if self.islogin():
page = 0 # 页数
articleList = list()
hasnext = True # 开始爬虫
while hasnext:
page += 1
if page == 5:
break
(articleListInPage,
hasnext) = self._searchByPage(keywordList, endTime, page)
is_article = list()
for article in articleListInPage:
if article not in articleList:
is_article.append(article)
articleList.extend(is_article)
print(len(articleList))
self.logger.debug('总共抓取文章有:%s' % len(articleList))
return articleList
else:
self.logger.error(u'未登录简书，请确认登录')
return None
def _searchByPage(self, keywordList, endTime, page):
hasnext = True
articleList = list()
page = str(page) # 在简书内部搜索框搜索
search_url_1 = "https://www.jianshu.com/search/do?q="
search_url_2 = "&type=note&page=" + page + "&order_by=published_at&time_range=a_day"
# q：是关键词搜索，page：页数,order_by是排序，time_range是按天排序
query = urllib.quote(' '.join(keywordList).encode('utf-8'))
search_url = search_url_1 + str(query) + search_url_2 # 抓取页面的网址
csrf = self.getCsrf(keywordList, page)
post_data = {'X-CSRF-Token': csrf}
# 获取页面信息
url_page = self.session.post(search_url,
headers=self.headers_1,
data=post_data,
textRspOnly=False)
if url_page.status_code == 200:
self.logger.debug(u'已经获取中大新闻页面')
else:
self.logger.debug(u'中大新闻获取完毕')
return (articleList, False) #已经爬取到新闻界面的最后一页
# 以下对中大新闻页面进行解析
articleJson = None
try:
articleJson = json.loads(url_page.text)
except Exception, e:
self.logger.error(e)
if articleJson is None:
self.logger.error('articleJson is None')
return
allInfo = articleJson['entries'] # allinfo代表字典0:{}\1:{}.....
for info in allInfo:
pre_publishedTime = info["first_shared_at"]
publishedTime_1 = ''.join(
re.findall('[^A-Za-z]', pre_publishedTime[0:-5])) # 文章发表时间
publishedTime = publishedTime_1[0:10] + ' ' + publishedTime_1[10:18]
#print(publishedTime)
urlTime = time.strptime(publishedTime, '%Y-%m-%d %H:%M:%S')
Y, M, D, H = urlTime[0:4]
urlTime2 = datetime.datetime(Y, M, D, H)
# 转化成时间戳来比较float
urlTime2 = time.mktime(urlTime2.timetuple()) # 文章发表时间的时间戳
startTime = endTime - datetime.timedelta(
days=self.channel.search_ranges)
startTimeIntSecond = time.mktime(startTime.timetuple()) # 开始爬取的时间戳
endTimeIntSecond = time.mktime(endTime.timetuple()) # 结束爬取的时间错
if urlTime2 >= startTimeIntSecond and urlTime2 <= endTimeIntSecond:
# 获取文章链接，爬取文章内容
pre_article_url = info['slug']
articleUrl = "https://www.jianshu.com/p/" + pre_article_url # 文章的链接
(content, authorName,
title) = self.crawlArticle(articleUrl) #爬取文章内容
#print(title)
if content is None: # 话题已被删除或则其他格式
print("没有爬到文章")
continue
authorId = info["user"]["id"] #作者id
likeCount = info["likes_count"] # 点赞数
readCount = info["views_count"] # 文章阅读数
replyCount = info["public_comments_count"] # 文章评论数
tid = info["id"] # 文章的id
article = Article(tid, self.channel.channel_id, title, content,
publishedTime, articleUrl, authorId,
authorName)
#self.crawlComment(article)
article.statistics.reply_count = replyCount
article.statistics.like_count = likeCount
article.statistics.read_count = readCount
if (article is not None) and (article not in articleList):
articleList.append(article)
else:
print('结束爬虫')
hasnext = False
break
return (articleList, hasnext)
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
cookies = 'U_TRS1=000000fa.9fe376b4.58573ebc.bde2f2c3; UOR=,vip.stock.finance.sina.com.cn,; vjuids=3923fcfb8.15914cd122a.0.e347599b65a6; SINAGLOBAL=183.63.92.250_1482112700.861930; SUB=_2AkMvC7H0f8NhqwJRmP4WzWzrb4xwzgnEieLBAH7sJRMyHRl-yD83qlNetRBAqqE4nv4pjjxQaUfLZo_Os-Bxsw..; SUBP=0033WrSXqPxfM72-Ws9jqgMF55529P9D9WFZzJ6nbHTRfVEqOXp-S.5z; SGUID=1482721389362_efec0e8d; vjlast=1488765553.1489054965.10; bdshare_firstime=1492414283526; _ct_uid=58f46f61.537a7929; lxlrtst=1492423120_o; rotatecount=2; Apache=59.42.29.149_1492670298.869113; ULV=1492670299361:18:6:6:59.42.29.149_1492670298.869113:1492670298484; afpCT=1; CNZZDATA1252916811=1442218969-1492654141-http%253A%252F%252Fnews.sina.com.cn%252F%7C1492664941; UM_distinctid=15b8a154522e79-0a3f79bddc9d05-4e45042e-100200-15b8a154523a49; CNZZDATA5399792=cnzz_eid%3D349789736-1492650802-http%253A%252F%252Fnews.sina.com.cn%252F%26ntime%3D1492667002; U_TRS2=00000095.1c285e96.58f85761.e07aa962; lxlrttp=1492423120'
html = self.session.download(url, encoding='utf-8', data=None, timeout=10, retry=3, addr=True, cookies=cookies)
if html:
article_url = html['url']
article_url = re.findall(r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
self.logger.info('[SinaNews]'+article_url)
#获取发布时间
date = re.findall(r'/(\d{4}-\d{2}-\d{2})/', article_url)
if len(date) == 0:
return None
# if date[0] < '2015-07-01':
# html = self.session.download(url, encoding='gbk', data=None, timeout=10, retry=3, addr=True)
soup = BeautifulSoup(html['html'], 'lxml')
main = soup.find('div', attrs={'class': "wrap-inner"})
main1 = soup.find('div', attrs={'class': "Main clearfix"})
main2 = soup.find('div', attrs ={'class': "main-content w1240"})
#第一种网页格式
if main is not None:
self.logger.debug('走第一种格式')
#获取标题
Ttitle = main.find('h1', attrs={'id': "artibodyTitle"})
if Ttitle is None:
self.logger.error('[SinaNews]' + '缺少标题，无法构成文章，可能已被修改格式，本文停止爬取')
return
else:
Ttitle = Ttitle.text.strip()
#获取发布时间
Ttime = main.find('span', attrs={'class': 'time-source'})
if Ttime is not None:
Ttime = Ttime.text.strip()
Ttime = re.findall(r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2}).*', Ttime)[0]
Ttime = Ttime[0] + '-' + Ttime[1] + '-' + Ttime[2] + ' ' + Ttime[3]
else:
self.logger.error('[SinaNews]' + '缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取')
return
if len(Ttime) == 16:
Ttime = Ttime + ':00'
#获取作者信息
Tauthor = soup.find('span', attrs={'class': "time-source"})
if Tauthor is not None:
Tauthor = Tauthor.find('a')
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = None
else:
Tauthor = None
#获取内容
Tcontent = main.find('div', attrs={'id': "artibody"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error('[SinaNews]' + '缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取')
return
#获取评论数
try:
channel = re.findall(r"channel: '(.*)',", html['html'])[0]
newsid = re.findall(r"newsid: '(.*)',", html['html'])[0]
data = {
'format': 'js',
'channel': channel,
'newsid': newsid,
'group': '',
'compress': '1',
'ie': 'gbk',
'oe': 'gbk',
'page': '1',
'page_size': '20'
}
re_url = 'http://comment5.news.sina.com.cn/page/info'
html1 = self.session.download(re_url, encoding='utf-8', data=data, isJson=False, timeout=10, retry=3, addr=False)
html1 = re.sub(r'(.*=)\{', '{', html1)
html1 = json.loads(html1)
totalcount = html1['result']['count']['show']
Treply = totalcount
except:
Treply = None
# 获取文章的id
articleid = re.findall(r'([a-z]{8}\d{7})', article_url)[0]
article = Article(tid=articleid, channel_id=self.channel.channel_id, title=Ttitle, publish_datetime=Ttime, content=Tcontent, url=article_url, author_name=Tauthor)
article.statistics.reply_count = Treply
return article
#第二种网页格式
elif main1 is not None:
self.logger.debug('走第二种格式')
#获取标题
Ttitle = main1.find('h1', attrs={'id': "artibodyTitle"})
if Ttitle is None:
self.logger.error('[SinaNews]' + '缺少标题，无法构成文章，可能已被修改格式，本文停止爬取')
return
else:
Ttitle = Ttitle.text.strip()
#获取时间
Ttime = main1.find('span', attrs={'id': "pub_date"})
if Ttime is not None:
Ttime = Ttime.text.strip()
Ttime = re.findall(r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2})', Ttime)[0]
Ttime = Ttime[0] + '-' + Ttime[1] + '-' + Ttime[2] + ' ' + Ttime[3]
else:
self.logger.error('[SinaNews]' + '缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取')
return
if len(Ttime) == 16:
Ttime = Ttime + ':00'
#获取发布者
Tauthor = main1.find('span', attrs={'id': "media_name"})
if Tauthor is not None:
Tauthor = Tauthor.find('a').text.strip()
else:
Tauthor = None
#获取内容
Tcontent = main1.find('div', attrs={'id': "artibody"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error('[SinaNews]' + '缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取')
return
try:
channel = re.findall(r"channel: '(.*)',", html['html'])[0]
newsid = re.findall(r"newsid: '(.*)',", html['html'])[0]
data = {
'format': 'js',
'channel': channel,
'newsid': newsid,
'group': '',
'compress': '1',
'ie': 'gbk',
'oe': 'gbk',
'page': '1',
'page_size': '20'
}
re_url = 'http://comment5.news.sina.com.cn/page/info'
html1 = self.session.download(re_url, encoding='utf-8', data=data, isJson=False, timeout=10, retry=3, addr=False)
html1 = re.sub(r'(.*=)\{', '{', html1)
html1 = json.loads(html1)
totalcount = html1['result']['count']['show']
Treply = totalcount
except:
Treply = None
# 获取文章的id
articleid = re.findall(r'([a-z]{8}\d{7})', article_url)[0]
article = Article(tid=articleid, channel_id=self.channel.channel_id,title=Ttitle, content=Tcontent, publish_datetime=Ttime, url=article_url, author_name=Tauthor)
article.statistics.reply_count = Treply
return article
#第三种网页格式
elif main2 is not None:
self.logger.debug(u'第三种格式')
#获取标题
Ttitle = main2.find('div', attrs={'class': "second-title"})
if Ttitle is None:
self.logger.error('[SinaNews]' + '缺少标题，无法构成文章，可能已被修改格式，本文停止爬取')
return
else:
Ttitle = Ttitle.text.strip()
# 获取时间
Ttime = main2.find('span', attrs={'class': "date"})
if Ttime is not None:
Ttime = Ttime.text.strip()
Ttime = re.findall(r'(\d{4})\D(\d{2})\D(\d{2})\D ' '(\d{2}:\d{2}).*', Ttime)[0]
Ttime = Ttime[0] + '-' + Ttime[1] + '-' + Ttime[2] + ' ' + Ttime[3]
else:
self.logger.error('[SinaNews]' + '缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取')
return
if len(Ttime) == 16:
Ttime = Ttime + ':00'
# 获取发布者
Tauthor = main2.find('a', attrs={'class': "source"})
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = None
# 获取内容
Tcontent = main2.find('div', attrs={'id': "article"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error('[SinaNews]' + '缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取')
return
# 获取评论数
try:
channel = re.findall(r"channel: '(.*)',", html['html'])[0]
newsid = re.findall(r"newsid: '(.*)',", html['html'])[0]
data = {
'format': 'js',
'channel': channel,
'newsid': newsid,
'group': '',
'compress': '1',
'ie': 'gbk',
'oe': 'gbk',
'page': '1',
'page_size': '20'
}
re_url = 'http://comment5.news.sina.com.cn/page/info'
html1 = self.session.download(re_url, encoding='utf-8', data=data, isJson=False, timeout=10, retry=3, addr=False)
html1 = re.sub(r'(.*=)\{', '{', html1)
html1 = json.loads(html1)
totalcount = html1['result']['count']['show']
Treply = totalcount
except:
Treply = None
# 获取文章的id
articleid = re.findall(r'([a-z]{8}\d{7})', article_url)[0]
article = Article(tid=articleid, channel_id=self.channel.channel_id, title=Ttitle, publish_datetime=Ttime, content=Tcontent, url=article_url, author_name=Tauthor)
article.statistics.reply_count = Treply
return article
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
cookies = 'Province=020; City=020; usertrack=c+5+hljsm+B+cg5MA7YDAg==; vjuids=-7517fab.15b5c40e631.0.a042d54907b81; _ntes_nnid=5e90ea8f4ef321150e3b5d43f68870c8,1491901408828; _ntes_nuid=5e90ea8f4ef321150e3b5d43f68870c8; UM_distinctid=15b5c41b7836eb-0fd2f7e510ef22-4e45042e-100200-15b5c41b78461b; __gads=ID=18c804c9f3ead780:T=1491901995:S=ALNI_MYWNxLkcHVgXyExP9eeFcD-mj7SiQ; afpCT=1; CNZZDATA1256734798=337963631-1491900970-http%253A%252F%252Fnews.163.com%252F%7C1492767097; CNZZDATA1256336326=1559830613-1491900088-http%253A%252F%252Fnews.163.com%252F%7C1492765460; vjlast=1491901409.1492754596.11; ne_analysis_trace_id=1492768109053; vinfo_n_f_l_n3=09c375e3d4394d15.1.13.1491901408836.1492766182939.1492768266676; s_n_f_l_n3=09c375e3d4394d151492768109056'
html = self.session.download(url,
encoding='gbk',
data=None,
timeout=10,
retry=3,
addr=True,
cookies=cookies)
if html:
article_url = html['url']
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
self.logger.info(article_url)
soup = BeautifulSoup(html['html'], 'html.parser')
main = soup.find('div', attrs={'class': "post_content_main"})
main1 = soup.find('div', attrs={'class': "ep-content-main"})
#第一种网页格式
if main is not None:
#获取标题
Ttitle = main.find('h1')
if Ttitle is None:
self.logger.error('[NeteaseNews]' +
'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取')
return
else:
Ttitle = Ttitle.text.strip()
#获取发布时间
Ttime = main.find('div', attrs={'class': "post_time_source"})
if Ttime is not None:
Ttime = Ttime.text.strip()
Ttime = re.findall(r'\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}',
Ttime)[0]
else:
self.logger.error('[NeteaseNews]' +
'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取')
return
if len(Ttime) == 16:
Ttime = Ttime + ':00'
#获取发布作者
Tauthor = main.find('div', attrs={'class': "post_time_source"})
if Tauthor is not None:
Tauthor = Tauthor.find('a')
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = None
#获取发布内容
Tcontent = main.find('div', attrs={'class': "post_text"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
dr = re.compile(r'<[^>]+>', re.S)
Tcontent = dr.sub('', Tcontent)
else:
self.logger.error('[NeteaseNews]' +
'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取')
return
# 获取评论数
articleid = ""
try:
articleid = re.findall(r'"docId" : "(.*)",',
html['html'])[0]
re_url = 'http://sdk.comment.163.com/api/v1/products/a2869674571f77b5a0867c3d71db5856/threads/' + articleid
html1 = self.session.download(url=re_url,
encoding='utf-8',
data=None,
isJson=True,
timeout=10,
retry=3)
Treply = html1["tcount"]
except:
Treply = None
self.logger.error('[NeteaseComment]url:' + article_url +
', tid:' + articleid + ', %s' +
traceback.format_exc())
finally:
article = Article(tid=articleid,
channel_id=self.channel.channel_id,
title=Ttitle,
content=Tcontent,
publish_datetime=Ttime,
url=article_url,
author_name=Tauthor)
article.statistics.reply_count = Treply
return article
#第二种网页格式
elif main1 is not None:
#标题
Ttitle = main1.find('h1')
if Ttitle is None:
self.logger.error('[NeteaseNews]' +
'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取')
return
else:
Ttitle = Ttitle.text.strip()
#发布的时间
Ttime = main1.find('div',
attrs={'class': "ep-time-source cDGray"})
Ttime1 = main1.find('div', attrs={'class': "ep-info cDGray"})
if Ttime is not None:
Ttime = Ttime.text.strip()
Ttime = re.findall(
r'\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}')[0]
elif Ttime1 is not None:
Ttime = Ttime1.text.strip()
Ttime = re.findall(
r'\d{4}-\d{2}-\d{2} \d{2}:\d{2}:\d{2}')[0]
else:
self.logger.error('[NeteaseNews]' +
'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取')
return
if len(Ttime) == 16:
Ttime = Ttime + ':00'
#获取作者信息
Tauthor = main1.find('div',
attrs={'class': "ep-time-soure cDGray"})
Tauthor1 = main1.find('div',
attrs={'class': "ep-source cDGray"})
if Tauthor is not None:
Tauthor = Tauthor.find('a')
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = None
elif Tauthor1 is not None:
Tauthor = Tauthor1.find('span')
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
print Tauthor
Tauthor = re.findall(r'来源：(.*)"', Tauthor)[0]
else:
Tauthor = None
else:
Tauthor = None
#获取内容
Tcontent = main1.find('div', attrs={'id': "endText"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
dr = re.compile(r'<[^>]+>', re.S)
Tcontent = dr.sub('', Tcontent)
else:
self.logger.error('[SinaNews]' +
'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取')
return
#获取评论数
try:
articleid = re.findall(r'"docId" : "(.*)",',
html['html'])[0]
re_url = 'http://sdk.comment.163.com/api/v1/products/a2869674571f77b5a0867c3d71db5856/threads/' + articleid
html1 = json.load(
self.session.download(re_url,
encoding='utf-8',
data=None,
isJson=True,
timeout=10,
retry=3))
Treply = html1['tcount']
except:
Treply = None
self.logger.error('[NeteaseComment]url:' + article_url +
', tid:' + articleid + ', %s' +
traceback.format_exc())
finally:
article = Article(tid=articleid,
channel_id=self.channel.channel_id,
title=Ttitle,
content=Tcontent,
publish_datetime=Ttime,
url=article_url,
author_name=Tauthor)
self.logger.debug("[crawlArticle]" + article.tid)
article.statistics.reply_count = Treply
return article
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
:return:返回一个article实例
'''
article_id = re.findall(r'\d+', url)[0]
html = self.session.get(url=url, headers=CRAWL_ARTICLE_HEADERS)
soup = BeautifulSoup(html.text, 'lxml')
if (url.find('zhuanlan') != -1):
main = soup.find('textarea', attrs={'id': "preloadedState"})
try:
if main is None:
return main
sub = re.findall(r'new Date\(\".*\"\)', main.text)[0]
sub = re.findall(r'".*"', sub)[0]
maintext = re.subn(r'new Date\(\".*\"\)', sub, main.text)[0]
try:
articleJson = json.loads(
maintext)["database"]["Post"][article_id]
except:
return None
#获取标题
title = articleJson["title"]
self.logger.info(title)
authorName = articleJson["author"]
contentSoup = BeautifulSoup(articleJson["content"])
content = contentSoup.text
commentCount = articleJson["commentCount"]
collapsedCount = articleJson["collapsedCount"]
likeCount = articleJson["likeCount"]
publishedTime = articleJson["publishedTime"][0:18]
timeArray = time.strptime(publishedTime, "%Y-%m-%dT%H:%M:%S")
publishedTime = time.strftime("%Y-%m-%d %H:%M:%S", timeArray)
article = Article(article_id, self.channel.channel_id, title,
content, publishedTime, url, None,
authorName)
article.statistics.reply_count = commentCount
article.statistics.like_count = likeCount
article.statistics.collect_count = collapsedCount
self.session.close()
return article
except:
return None
if (url.find('question') != -1):
# 获取标题
try: # 针对404页面
main = soup.find('div', attrs={
'id': "data"
}).attrs['data-state']
articleJson = json.loads(main)
questionJson = articleJson['entities']['questions'][article_id]
title = questionJson['title']
self.logger.info(title)
contentSoup = BeautifulSoup(questionJson['editableDetail'],
'lxml')
content = contentSoup.text
author_id = questionJson['author']['id']
author_name = questionJson['author']['name']
createTimeInFloat = questionJson['created']
publish_datetime = time.strftime(
'%Y-%m-%d %H:%M:%S', time.localtime(createTimeInFloat))
reply_count = questionJson['commentCount']
read_count = questionJson['visitCount']
collect_count = questionJson['followerCount']
article = Article(article_id, self.channel.channel_id, title,
content, publish_datetime, url, author_id,
author_name)
article.statistics.reply_count = reply_count
article.statistics.read_count = read_count
article.statistics.collect_count = collect_count
return article
except:
return None
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
html = self.session.download(url,
encoding='utf-8',
data=None,
timeout=10,
retry=3,
addr=True)
if html:
article_url = html['url']
if article_url.find(self.channel.url) < 0:
self.logger.warn('Unrelated url found:%s', url)
return None
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
self.logger.debug('[iFengnews]' + article_url)
soup = BeautifulSoup(html['html'], 'lxml') # 'html.parser' 解析器
main = soup.find('div', attrs={'class': "main"})
main1 = soup.find('div', attrs={'class': "yc_main"})
if main is not None:
self.logger.debug(u'走第一种格式')
Ttitle = main.find('h1').text.strip() # 标题
if Ttitle is None:
self.logger.error(u'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
url)
return
else:
Tpublishtime = main.find('span', attrs={
'class': "ss01"
}).text.strip()
if Tpublishtime is None:
self.logger.error(
u'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ', url)
return
else:
Tpublishtime = Tpublishtime.replace(u'年', '-').replace(
u'月', '-').replace(u'日', '')
Tauthor = main.find('a', attrs={'target': "_blank"})
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = 'None'
Tcontent = main.find('div',
attrs={'id': "main_content"})
# print Tcontent
# Tcontent = Tcontent.find('p')
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error(
u'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
url)
return
doc_url = re.findall(r'"commentUrl":"(.*)",',
html['html'])
if doc_url:
doc_url = doc_url[0]
else:
doc_url = url
data1 = {
'callback': 'newCommentListCallBack',
# 'doc_url': 'http://gd.ifeng.com/a/20171010/6053241_0.shtml',
'doc_url': doc_url,
'job': '1',
'callback': 'newCommentListCallBack'
}
re_url = 'http://comment.ifeng.com/get.php'
html1 = self.session.download(re_url,
encoding='gbk',
data=data1,
timeout=10,
retry=3,
addr=False,
isJson=True)
# html1 = json.loads( html1[html1.find('=') + 1:html1.rfind(';c')] )
Treply = html1['count']
if len(html1['comments']) is not 0:
articleid = html1['comments'][0]['article_id']
else:
articleid = article_url
articleid = articleid[articleid.find('a/') + 2:
-6] # 由于数据库字段长度有限，所以截取部分作为ID
self.logger.warn(u'地址 %s 没有评论因此以url地址部分字段作为tid',
article_url)
meta_info = doc_url
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Tpublishtime,
article_url,
None,
Tauthor,
meta_info=meta_info)
article.statistics.reply_count = Treply
self.logger.info(article)
return article
##对第二种格式的爬取
if main1 is not None:
self.logger.debug(u'走第二种格式')
Ttitle = main1.find('h1').text.strip()
if Ttitle is None:
self.logger.error(u'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
url)
return
else:
Tpublishtime = main1.find('span').text.strip()
if Tpublishtime is None:
self.logger.error(
u'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ', url)
return
# return Tpublishtime
Tauthor = main1.find('a', attrs={'target': "_blank"})
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = 'None'
Tcontent = main1.find('div', attrs={'class': "yc_con_txt"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.warn(
u'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ', url)
return
doc_url = re.findall(r'"commentUrl":"(.*)",', html['html'])
if doc_url:
doc_url = doc_url[0]
else:
doc_url = url
data1 = {
'callback': 'newCommentListCallBack',
# 'doc_url': 'http://gd.ifeng.com/a/20171010/6053241_0.shtml',
'doc_url': doc_url,
'job': '1',
'callback': 'newCommentListCallBack'
}
re_url = 'http://comment.ifeng.com/get.php'
html1 = self.session.download(re_url,
encoding='gbk',
data=data1,
timeout=10,
retry=3,
addr=False,
isJson=True)
# html1 = json.loads( html1[html1.find('=') + 1:html1.rfind(';c')] )
try:
Treply = html1['count']
except:
Treply = None
if len(html1['comments']) is not 0:
articleid = html1['comments'][0]['article_id']
else:
articleid = url.strip()
articleid = articleid[articleid.find('a/') +
2:-6] # 由于数据库字段长度有限，所以截取部分作为ID
self.logger.warn(u'地址 %s 没有评论因此以url地址部分字段作为tid',
article_url)
meta_info = doc_url
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Tpublishtime,
article_url,
None,
Tauthor,
meta_info=meta_info)
article.statistics.reply_count = Treply
self.logger.info(article)
return article
if (main is None) and (main1 is None):
self.logger.warn(u"存在另外一种html格式：：%s", url)
return
def __parseSearchPage(self, data):
'''
@return: (articleList,hasnext)
'''
articleList = list()
hasnext = False
soup = BeautifulSoup(data, "lxml")
self.logger.info(soup)
# check if no result
noResultDivList = soup.findAll('div', {'class': 'pl_noresult'})
if len(noResultDivList) > 0:
hasnext = False
self.logger.info('No result')
return (articleList, hasnext)
# find page bar to check if more
pageDivList = soup.findAll('div', {'class': 'W_pages'})
if len(pageDivList) > 0:
pageDiv = pageDivList[0]
if len(pageDiv.findAll('a',
{'class': 'page next S_txt1 S_line1'})) > 0:
hasnext = True
if hasnext is False:
self.logger.info('The last page')
root_1 = soup.findAll('div', {"action-type": "feed_list_item"})
# self.logger.debug(root_1)
for r in root_1:
root_2 = r.find('div', {'class': "content clearfix"})
mid = r.attrs['mid']
article_url = root_2.find('div', {
'class': "feed_from W_textb"
}).findNext('a').attrs['href']
self.logger.debug('1 %s', article_url)
if not article_url.startswith('http:'):
article_url = 'http:' + article_url
# self.logger.debug(article_url)
root_content = root_2.find('p', {'class': "comment_txt"})
long_content = root_content.find('a', {'action-type': "fl_unfold"})
try:
link_content = root_content.find('a').attrs['href']
link_content = ' 原文链接： ' + link_content
except:
link_content = ''
if long_content:
content_url = 'http://s.weibo.com/ajax/direct/morethan140?' + long_content.attrs[
'action-data']
self.session.randomSleep()
response = self.session.get(content_url, textRspOnly=False)
try:
content_html = response.json()['data']['html']
content = BeautifulSoup(
content_html, 'html.parser').text.strip().replace(
"'", "''").replace("%", "\%").replace(":", "\:")
except Exception, e:
self.logger.debug('Exception: %s' % e)
continue
else:
content = root_content.text.strip().replace("'", "''").replace(
"%", "\%").replace(":", "\:")
content = content + link_content
# self.logger.error(content)
title = content[:30].replace("'", "''").replace("%", "\%").replace(
":", "\:") + ' '
author_id = r.attrs['tbinfo']
author_id = re.findall(r'ouid=(\d+)', author_id)[0]
author_name = root_2.find('a').attrs['nick-name']
publish_datetime = root_2.find('a', {
'class': "W_textb"
}).attrs['date']
try:
publish_datetime = time.strftime(
'%Y-%m-%d %H:%M:%S',
time.localtime(float(publish_datetime) / 1000))
except:
continue
article = Article(mid,
self.channel.channel_id,
title,
content,
publish_datetime,
url=article_url,
author_id=author_id,
author_name=author_name)
# fetch statistics
root_3 = r.find('div', {'class': "feed_action clearfix"})
soup_li_list = root_3.findAll('li')
self.__parseStatistics(article, soup_li_list)
# print mid, article_url, add_datetime, channeltype, channel, title, content, author_id, author_name, \
# publish_datetime, reply_count, read_count, like_count, collect_count, forward_count
if article not in articleList:
articleList.append(article)
def crawlNewArticle(self, url):
'''
爬取url前缀为new.qq.com的文章
:param url:
:return:
'''
html = self.session.download(url,
encoding='gbk',
data=None,
timeout=10,
retry=3,
addr=True)
if html:
article_url = html['url']
if article_url.find('new.qq.com/omn') < 0:
self.logger.warn('Unrelated url found:%s', url)
return
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
self.logger.debug('[TencentNew]' + article_url)
soup = BeautifulSoup(html['html'], 'html.parser')
script_tags = soup.head.find_all('script')
data = dict()
for tag in script_tags:
text = re.search(r'window.DATA = (.*)', tag.text, re.S)
if text:
data = json.loads(text.group(1))
tid = data['article_id']
title = data['title']
author_name = data['media']
author_id = data['media_id']
publish_datetime = data['pubtime']
comment_id = data['comment_id']
main = soup.find('div', attrs={'class': 'qq_conent clearfix'})
t_content = ''
if main is not None:
contents = main.find_all('p', {'class': 'one-p'})
for content in contents:
if content.string is None:
continue
t_content += str(content.get_text().strip())
get_comment_count_url = 'https://coral.qq.com/article/%s/commentnum?callback=_article%scommentnum' % (
comment_id, comment_id)
comment_data = self.session.download(get_comment_count_url)
comment_data = re.search(
r'_article%scommentnum\((.*)\)' % comment_id, comment_data)
comment_dict = eval(comment_data.group(1))
reply_count = comment_dict['data']['commentnum']
meta_info = '{"commentid":"%s"}' % comment_id
article = Article(tid=tid,
channel_id=self.channel.channel_id,
title=title,
content=t_content,
publish_datetime=publish_datetime,
url=article_url,
author_id=author_id,
author_name=author_name,
meta_info=meta_info)
article.statistics.reply_count = reply_count
return article
return None
def crawlNewsArticle(self, url):
'''
爬取url前缀为news.qq.com和gd.qq.com的文章
:param url:
:return:
'''
html = self.session.download(url,
encoding='gbk',
data=None,
timeout=10,
retry=3,
addr=True)
if html:
article_url = html['url']
if article_url.find('news.qq.com') < 0 and article_url.find(
'gd.qq.com') < 0:
self.logger.warn('Unrelated url found:%s', url)
return None
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
self.logger.debug('[TencentNews]' + article_url)
soup = BeautifulSoup(html['html'], 'html.parser')
main = soup.find('div', attrs={'id': "Main-Article-QQ"})
main1 = soup.find('div', attrs={'id': "Main-P-QQ"})
if main is not None:
Ttitle = main.find('h1').text.strip() #标题
Ttime = main.find('span', attrs={'class':
"article-time"}) #发布时间
Ttime1 = main.find('span', attrs={'class': "a_time"})
Ttime2 = main.find('span', attrs={'class': "pubTime"})
if Ttime is not None:
Ttime = Ttime.text.strip()
elif Ttime1 is not None:
Ttime1 = Ttime1.text.strip()
Ttime = Ttime1
elif Ttime2 is not None:
Ttime2 = Ttime2.text.strip()
Ttime = Ttime2
else:
Ttime = datetime.datetime.now().strftime(
'%Y-%m-%d %H:%M:%S')
if len(Ttime) == 16:
Ttime = Ttime + ':00'
Tauthor = main.find('span', attrs={'class': "a_source"})
Tauthor1 = main.find('span', attrs={'class': "color-a-1"})
if Tauthor is not None:
#Tauthor = Tauthor.find('a').text.strip()
Tauthor = Tauthor.text.strip()
elif Tauthor1 is not None:
#Tauthor1 = Tauthor1.find('a').text.strip()
Tauthor1 = Tauthor1.text.strip()
Tauthor = Tauthor1
else:
Tauthor = None
Tcontent = main.find('div',
attrs={'id': "Cnt-Main-Article-QQ"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
Tcontent = None
articleid = re.findall(r'id:\'(\d+)\',', html['html'])[0]
try:
commentid = re.findall(r'cmt_id = (\d+);', html['html'])[0]
meta_info = '{"commentid":"%s"}' % commentid
except:
commentid = None
meta_info = None
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Ttime,
article_url,
None,
Tauthor,
meta_info=meta_info)
if commentid is not None:
try:
re_url = 'http://coral.qq.com/article/' + commentid + '/commentnum'
html1 = json.loads(
self.session.download(re_url,
encoding='utf-8',
data=None,
timeout=10,
retry=3))
Treply = int(html1['data']['commentnum'])
except Exception:
traceInfo = traceback.format_exc()
self.logger.error(
'Faile to parse comment for %s (cid=%s):%s',
articleid, commentid, traceInfo)
Treply = None
article.statistics.reply_count = Treply
return article
elif main1 is not None:
Ttitle = soup.find('meta', attrs={
'name': "Description"
}).attrs['content'] # 标题
Ttime = re.findall(
r"pubtime\D+(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2})\',",
html['html'])
if Ttime is not None:
Ttime = Ttime[0]
Ttime = Ttime[0] + '-' + Ttime[1] + '-' + Ttime[
2] + ' ' + Ttime[3]
else:
Ttime = datetime.datetime.now().strftime(
'%Y-%m-%d %H:%M:%S')
if len(Ttime) == 16:
Ttime = Ttime + ':00'
Tauthor = re.findall(r'para = {\s+name: \"(.*)\",',
html['html'])
if Tauthor is not None:
Tauthor = Tauthor[0]
else:
Tauthor = None
con_url = re.sub(r'\.htm\?.*', '.hdBigPic.js', article_url)
con_html = self.session.download(con_url,
encoding='gbk',
data=None,
timeout=10,
retry=3)
con_list = re.findall(r'<p>(.*?)</p>', con_html)
if con_list is not None:
TT = []
for i in con_list:
if i.strip() not in TT:
TT.append(i)
Tcontent = ''.join(TT)
else:
Tcontent = None
articleid = re.findall(r'id:\'(\d+)\',', html['html'])[0]
try:
commentid = re.findall(r'aid\D+(\d+)\",', html['html'])[0]
meta_info = '{"commentid":"%s"}' % commentid
except:
commentid = None
meta_info = None
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Ttime,
article_url,
None,
Tauthor,
meta_info=meta_info)
try:
if commentid is not None:
re_url = 'http://coral.qq.com/article/batchcommentnum'
data1 = {'targetid': articleid}
html1 = json.loads(
self.session.download(re_url,
encoding='utf-8',
data=data1,
timeout=10,
retry=3))
Treply = int(html1['data'][0]['commentnum'])
else:
Treply = None
except:
Treply = None
article.statistics.reply_count = Treply
return article
return None
def crawlArticle(self, url):
'''
根据url爬取文章内容和统计信息
@return: 返回一个Article实例
'''
cookies = '_HY_lvt_c5928ed57343443aa175434a09ea2804=1492582419784; _HY_CTK_c5928ed57343443aa175434a09ea2804=0acc9a79b115c4ca1931c41e303bec28; BAIDU_SSP_lcr=https://www.baidu.com/link?url=NszYD2w_HgkPWqrzDQ3WKApYldw_9MpUVun9r-R09M7r0dh09MUwTHzG087WaJrhBwMCY-7pDfds4xjtWArRf2xh01DHOWWWd9DpBnHwZ03&wd=&eqid=84d5edfe0003dbdd0000000658f6c280; ALLYESID4=0DB901C6E627D980; sso_c=0; wdcid=5838509dcecc0a53; _people_ip_new_code=510000; UM_distinctid=15b802cdbd364d-02e7f218c26ae6-4e45042e-100200-15b802cdbd49ce; wdses=62d3f0f698d07532; sfr=1; CNZZDATA1260954200=1457761991-1492499618-null%7C1492580619; CNZZDATA1260954203=33096124-1492503288-null%7C1492578888; CNZZDATA1256327855=1768205365-1492503342-null%7C1492578947; wdlast=1492582420'
html = self.session.download(url,
data=None,
isJson=False,
timeout=10,
retry=3,
addr=True,
cookies=cookies)
soup = BeautifulSoup(html['html'], 'html.parser') # 'html.parser' 解析器
try:
meta = soup.find('meta').attrs['content']
except:
self.logger.warn(u'找不到meta里的content')
return
# self.logger.error('%s',meta)
if "GB2312" in meta:
encoding1 = 'GB2312'
elif "UTF-8" in meta:
encoding1 = 'UTF-8'
elif "utf-8" in meta:
encoding1 = 'utf-8'
else:
encoding1 = 'gbk'
html = self.session.download(url,
encoding=encoding1,
data=None,
isJson=False,
timeout=10,
retry=3,
addr=True,
cookies=cookies)
# 不同网页编码格式让其重新下载一遍
soup = BeautifulSoup(html['html'], 'html.parser') # 'html.parser' 解析器
main = soup.find('body')
if html:
article_url = html['url']
# self.logger.debug(article_url)
if article_url.find(self.channel.url) < 0:
self.logger.warn('Unrelated url found:%s', url)
return None
# if '.html' not in article_url:
# self.logger.error(u'非文章类型网址：%s ',article_url)
# return
try:
article_url = re.findall(
r'.*?\.html|.*?\.htm|.*?\.shtml|.*?\.shtm', article_url)[0]
except:
self.logger.error(u'网址后缀不符合：%s ', article_url)
return
self.logger.debug('[peoplenews]' + article_url)
articleid = article_url
articleid = articleid[articleid.find('cn/') +
3:-5] # 由于数据库字段长度有限，所以截取部分作为ID
self.logger.warn(u'地址 %s 以url地址部分字段作为tid', article_url)
if 'bbs1' not in article_url:
main1 = soup.find(
'div', attrs={'class': "i_cont"}
) #http://health.people.com.cn/n1/2017/1011/c14739-29579836.html
main2 = soup.find(
'div', attrs={'class': "text_c"}
) #http://rencai.people.com.cn/n/2014/0721/c244800-25311391.html
if (main1 is None) and (main2 is None):
self.logger.debug(u'走main')
try:
Ttitle = main.find('h1').text.strip() # 标题
except:
self.logger.error(u'Ttitle存在走了main部分却不满足其他格式的的url：：%s',
article_url)
return
if Ttitle is None:
self.logger.error(
u'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
Ttitle_crawl = soup.find('div',
attrs={'class':
"box01"}) #对应一种格式
if Ttitle_crawl is None:
self.logger.error(
u'Ttitle_crawl存在走了main部分却不满足其他格式的的url：：%s',
article_url)
return
try:
Tpublishtime = Ttitle_crawl.find('div',
attrs={
'class': "fl"
}).text.strip()
except:
self.logger.error(u'main中发布时间不匹配')
return
if Tpublishtime is None:
self.logger.error(
u'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
# self.logger.error(Tpublishtime)
Tpublishtime = re.findall(
r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2}(:\d{2})?)',
Tpublishtime)[0]
# Tpublishtime = Tpublishtime[:18]
if len(Tpublishtime[4]) > 1:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + Tpublishtime[4]
else:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + ':00'
# Tpublishtime = Tpublishtime.replace(u'年', '-').replace(u'月', '-').replace(u'日', '')
Tauthor = Ttitle_crawl.find(
'a', attrs={'target': "_blank"})
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = 'None'
Tcontent = soup.find('div',
attrs={'class': "box_con"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error(
u'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
Treply = None #这种格式下没有这些统计可以获取
meta_info = None
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Tpublishtime,
article_url,
None,
Tauthor,
meta_info=meta_info)
article.statistics.reply_count = Treply
#self.logger.info(article)
return article
elif (main1 is not None):
self.logger.debug(u'走main1')
Ttitle = main1.find('h2') # 标题
if Ttitle is None:
self.logger.error(
u'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
Ttitle = Ttitle.text.strip()
try:
Tpublishtime = main1.find('div',
attrs={
'class': "artOri"
}).text.strip()
except:
self.logger.error(u'main1中发布时间不匹配')
return
if Tpublishtime is None:
self.logger.error(
u'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
Tpublishtime = re.findall(
r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2}(:\d{2})?)',
Tpublishtime)[0]
# self.logger.error(Tpublishtime)
# Tpublishtime = Tpublishtime[:18]
if len(Tpublishtime[4]) > 1:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + Tpublishtime[4]
else:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + ':00'
Tauthor = main1.find('div', attrs={
'class': "artOri"
}).find('a', attrs={'target': "_blank"})
# self.logger.debug(u"作者:%s",Tauthor)
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = 'None'
Tcontent = main1.find('div', attrs={'class': "artDet"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error(
u'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
Treply = None # 这种格式下没有这些统计可以获取
meta_info = None
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Tpublishtime,
article_url,
None,
Tauthor,
meta_info=meta_info)
article.statistics.reply_count = Treply
self.logger.info(article)
return article
elif (main2 is not None):
self.logger.debug(u'走main2')
Ttitle = main2.find('h2', attrs={'class': "one"}) # 标题
if Ttitle is None:
self.logger.error(
u'缺少标题，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
Ttitle = Ttitle.text.strip()
try:
Tpublishtime = main2.find('span',
attrs={
'id': "p_publishtime"
}).text.strip()
except:
self.logger.error(u'main2中发布时间不匹配')
return
if Tpublishtime is None:
self.logger.error(
u'缺少文章发布时间，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
else:
Tpublishtime = re.findall(
r'(\d{4})\D(\d{2})\D(\d{2})\D(\d{2}:\d{2}(:\d{2})?)',
Tpublishtime)[0]
# self.logger.error(Tpublishtime)
# Tpublishtime = Tpublishtime[:18]
if len(Tpublishtime[4]) > 1:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + Tpublishtime[4]
else:
Tpublishtime = Tpublishtime[0] + '-' + Tpublishtime[1] + '-' + Tpublishtime[2] + ' ' + \
Tpublishtime[3] + ':00'
Tauthor = main2.find('span', attrs={
'id': "p_origin"
}).find('a', attrs={'target': "_blank"})
# self.logger.debug(u"作者:%s",Tauthor)
if Tauthor is not None:
Tauthor = Tauthor.text.strip()
else:
Tauthor = 'None'
Tcontent = main2.find('div',
attrs={'class': "show_text"})
if Tcontent is not None:
Tcontent = Tcontent.text.strip()
Tcontent = re.sub(r'\n|\t', '', Tcontent)
else:
self.logger.error(
u'缺少文章内容，无法构成文章，可能已被修改格式，本文停止爬取::该网站为 %s ',
article_url)
return
Treply = None # 这种格式下没有这些统计可以获取
meta_info = None
article = Article(articleid,
self.channel.channel_id,
Ttitle,
Tcontent,
Tpublishtime,
article_url,
None,
Tauthor,
meta_info=meta_info)
article.statistics.reply_count = Treply
self.logger.info(article)
return article
else:
self.logger.warn(u'存在另外一种html格式 %s', article_url)
def __updateOldArticleToArticleHistoryTable(self,
articleList,
currentTableName,
historyTableName,
isEventTable=False):
'''
更新到文章历史表
@param currentTableName: 当前文章表：全局文章表、实体文章表或者实体事件文章表
@param historyTableName: 历史文章表：全局文章表、实体文章表或者实体事件文章表
@param eventId: 如果更新到实体事件文章表，则需要提供事件id，否则为None
'''
if len(articleList) > 0:
if isEventTable is False:
eventIdFieldName = ''
else:
eventIdFieldName = ',EVENT_ID'
#找寻老文章
selectSql = '''
SELECT TID, CHANNEL_ID %s FROM %s where %s
'''
whereClauseList = map(
lambda article: '(TID="%s" and CHANNEL_ID=%d)' %
(article.tid, article.channel_id), articleList)
self.dbProxy.execute(selectSql %
(eventIdFieldName, currentTableName,
' or '.join(whereClauseList)))
resultList = self.dbProxy.fetchall()
if isEventTable:
existingArticleList = map(
lambda item: Article(item[0], item[1], eventId=item[2]),
resultList)
else:
existingArticleList = map(
lambda item: Article(item[0], item[1]), resultList)
toBeUpdateArticleList = list()
for item in existingArticleList:
index = articleList.index(item)
obj = copy.copy(articleList[index])
obj.eventId = item.eventId
toBeUpdateArticleList.append(obj)
if len(toBeUpdateArticleList) > 0:
n = datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S')
if isEventTable is False:
eventIdFieldName = ''
else:
eventIdFieldName = 'EVENT_ID,'
insertSql = '''
INSERT INTO %s (TID, %s CHANNEL_ID,
READ_COUNT,LIKE_COUNT, REPLY_COUNT,
FORWARD_COUNT, COLLECT_COUNT, HEAT, ADD_DATETIME)
VALUES %s
'''
valueList = list()
for article in toBeUpdateArticleList:
statistics = article.statistics
if isEventTable is False:
eventIdFieldValue = ''
else:
eventIdFieldValue = str(article.eventId) + ','
valueList.append(
'("%s", %s %d, %s, %s, %s, %s, %s, %s, "%s")' %
(article.tid, eventIdFieldValue, article.channel_id,
statistics.read_count if statistics.read_count
is not None else Constants.DEFAULT_NUM,
statistics.like_count if statistics.like_count
is not None else Constants.DEFAULT_NUM,
statistics.reply_count if statistics.reply_count
is not None else Constants.DEFAULT_NUM,
statistics.forward_count if statistics.forward_count
is not None else Constants.DEFAULT_NUM,
statistics.collect_count if statistics.collect_count
is not None else Constants.DEFAULT_NUM,
statistics.heat if statistics.heat is not None else
Constants.DEFAULT_NUM, n))
if len(valueList) > 0:
self.dbProxy.execute(insertSql %
(historyTableName, eventIdFieldName,
','.join(valueList)))
self.dbProxy.commit()
def __parse_page_url(self, page_url, count_page=0):
'''
获取每一页的微博信息
@return: article_list
'''
user_agent = random.choice(AGENTS)
headers = {
'Accept':
'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8',
'Accept-Encoding': 'gzip, deflate',
'Accept-Language': 'zh-CN,zh;q=0.8',
'Host': 's.weibo.com',
'Referer': 'http://s.weibo.com/',
'User-Agent': user_agent
}
self.change_cookie()
cookies = self.__load_cookies_from_lwp(self.user_name)
response = self.session.get(url=page_url,
headers=headers,
cookies=cookies)
# 若此cookie无效，则重新登录获取新的cookie
if re.match(r'File not .*', response) is not None:
self.__login(self.user_name)
cookies = self.__load_cookies_from_lwp(self.user_name)
response = self.session.get(url=page_url,
headers=headers,
cookies=cookies)
html = etree.HTML(response)
# 内容list
card_wrap_list = html.xpath(
'//div[@class="m-wrap"]/div[@id="pl_feedlist_index"]//div[@class="card-wrap"]'
)
article_list = list()
for card_wrap in card_wrap_list:
try:
mid = card_wrap.xpath('./@mid')[0]
except Exception as e:
continue
info = card_wrap.xpath('.//div[@class="content"]')[0]
author_name = info.xpath('.//a[@class="name"]/text()')[0]
author_id = re.search(
r'/(\d+).?refer_flag',
info.xpath('.//a[@class="name"]/@href')[0]).group(1)
txt = info.xpath('.//p[@class="txt"]')[0]
content = txt.xpath('string(.)').strip()
title = content[:30].replace("'", "''").replace("%", "\%").replace(
":", "\:") + ' '
article_info = info.xpath('.//p[@class="from"]/a')[0]
article_url = self.HTTP + article_info.xpath('./@href')[0]
statistics = card_wrap.xpath(
'.//div[@class="card-act"]//li/a/text()')
collect_count = re.search(r'(\d+)', statistics[0])
forward_count = re.search(r'(\d+)', statistics[1])
reply_count = re.search(r'(\d+)', statistics[2])
like_count = card_wrap.xpath(
'.//div[@class="card-act"]//li/a/em/text()')
if collect_count:
collect_count = collect_count.group(1).strip()
else:
collect_count = 0
if forward_count:
forward_count = forward_count.group(1).strip()
else:
forward_count = 0
if reply_count:
reply_count = reply_count.group(1).strip()
else:
reply_count = 0
if len(like_count):
like_count = like_count[0]
else:
like_count = 0
publish_datetime = self.__parse_time(article_url, cookies)
article = Article(mid,
self.channel.channel_id,
title,
content,
url=article_url,
publish_datetime=publish_datetime,
author_id=author_id,
author_name=author_name)
article.statistics.collect_count = collect_count
article.statistics.forward_count = forward_count
article.statistics.reply_count = reply_count
article.statistics.like_count = like_count
if article not in article_list:
article_list.append(article)
self.logger.debug('%d page crawl article is %s' %
(count_page, len(article_list)))
return article_list
