def extract_answerers_followers(user_id, user_id_list):
mongo = Mongo_1()
follower_id = []
if user_id in user_id_list:
temp_list = mongo.db.haoyouguanxi.answerer_follower.find(
{"user_id": user_id})
for list in temp_list:
if list.get('follower') is not None:
for l in list.get('follower'):
follower_id.append(l.get('follower_id'))
break
if len(follower_id) == 0:
temp_list1 = mongo.db.haoyouguanxi.answerer_followers.find(
{"user_id": user_id})
for list in temp_list1:
if list.get('follower') is not None:
for l in list.get('follower'):
follower_id.append(l.get('follower_id'))
break
follower_id.append('user_exists')
mongo.client.close()
return follower_id
else:
mongo.client.close()
return follower_id
def extract_answerID():
answerID_list = []
mongo = Mongo_1()
temp_list = mongo.db.question_answers.find({}, {
'answer_num': 1,
'answers': 1
})
# 旧的回答
temp_list1 = mongo.db.answer.find({}, {'answer_num': 1, 'answers': 1})
for item in temp_list:
if item.get('answer_num') == 0:
continue
else:
answer_list = item.get('answers')
for answer in answer_list:
answerID_list.append(answer.get('answer_id'))
for item in temp_list1:
if item.get('answer_num') == 0:
continue
else:
answer_list = item.get('answers')
for answer in answer_list:
answerID_list.append(answer.get('answer_id'))
mongo.client.close()
# 去重
return list(set(answerID_list))
def __init__(self, fileNum):
self.user_id = None
self.url_damin = 'https://www.zhihu.com/people/'
self.is_del = False
# self.proxy = None
self.userID_list = []
self.fileNum = fileNum
self.file = None
self.start = None
self.end = None
self.type = None
self.user_info = None
self.mongo = Mongo_1()
self.mongo1 = Mongo()
self.current_proxy = None
self.content = None
self.headers = {
'Accept':
'textml,application/json,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
'Accept-Encoding': 'gzip, deflate, sdch',
'Accept-Language':
'en-US,en;q=0.8,zh-CN;q=0.6,zh;q=0.4,zh-TW;q=0.2',
'Host': 'www.zhihu.com',
'User-Agent':
'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/43.0.2357.81 Safari/537.36',
'Referer': 'http://www.zhihu.com/',
'Cookie': None,
'x-udid': None,
}
self.get_userinfo()
def extract_last_followers():
mongo = Mongo_1()
temp = mongo.db.followers_last.find({})
user_id_list = []
for l in temp:
user_id_list.append(l.get('user_id'))
mongo.client.close()
return user_id_list
def extract_questionUrl():
questionUrl_list = []
mongo = Mongo_1()
items = mongo.db.question_url.find(
{}, {"question_url": 1}) #find()----检测字符串中是否包含子字符串str
for item in items:
questionUrl_list.append(item['question_url'])
mongo.client.close()
return questionUrl_list
def extract_question_followers(question_id):
mongo = Mongo_1()
lists = mongo.db.english_followers_new.find({"question_id": question_id})
follower_id = []
for l in lists:
if l.get('followers') is not None:
for l1 in l.get('followers'):
follower_id.append(l1)
mongo.client.close()
return follower_id
def extract_answer_voters(answer_id):
mongo = Mongo_1()
lists = mongo.db.voters_old.find({"answer_id": answer_id})
voter_id = []
for l in lists:
if l.get('voters') is not None:
for l1 in l.get('voters'):
voter_id.append(l1)
mongo.client.close()
return voter_id
def extract_question_answers(question_id):
mongo = Mongo_1()
temp_list = mongo.db.answer.find({"question_id": question_id})
answer_id = []
for list in temp_list:
if list.get('answer_num') is not None:
answer_id.append(list.get('answer_num'))
break
mongo.client.close()
return answer_id
def __init__(self, fileNum):
self.fileNum = fileNum
self.mongo = Mongo_1()
self.userID_list = []
self.id = None
self.file = None
self.start = None
self.end = None
self.driver = None
self.get_people_detail()
def extract_answer_comments(answer_id):
mongo = Mongo_1()
temp_list = mongo.db.comment.find({"answer_id": answer_id})
comment_id = []
for list in temp_list:
if list.get('comment_num') is not None:
comment_id.append(list.get('comment_num'))
break
mongo.client.close()
return comment_id
def extract_last_voters():
mongo = Mongo_1()
temp = mongo.db.voters_last.find({})
user_id_list = []
for l in temp:
user_id_list.append(l.get('user_id'))
mongo.client.close()
return user_id_list
# if __name__ == "__main__":
# print extract_answer_voters(255313072)
def extract_voters_following(user_id, user_id_list):
mongo = Mongo_1()
following_id = []
if user_id in user_id_list:
temp_list = mongo.db.haoyouguanxi.voter_following.find(
{"user_id": user_id})
for list in temp_list:
if list.get('following') is not None:
for l in list.get('following'):
following_id.append(l.get('following_id'))
break
mongo.client.close()
following_id.append('user_exists')
return following_id
else:
mongo.client.close()
return following_id
def __init__(self, fileNum):
# self.question_url = question_url
# self.id = question_url.replace('/question/', '')
self.question_url = None
self.id = None
self.log_url = None
self.url = None
# self.log_url = 'https://www.zhihu.com' + question_url + '/log'
# self.url = 'https://www.zhihu.com' + question_url #'/answers/created'
self.url_domain = 'https://www.zhihu.com'
self.content = None
self.is_del = False
self.fileNum = fileNum
self.file = None
self.start = None
self.end = None
# self.proxy = None
self.type = None
self.questionUrl_list = None
self.topic_list = []
self.log = []
self.mongo = Mongo_1()
self.current_proxy = None
self.headers = {
'Accept':
'textml,application/json,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8',
'Accept-Encoding': 'gzip, deflate, sdch',
'Accept-Language':
'en-US,en;q=0.8,zh-CN;q=0.6,zh;q=0.4,zh-TW;q=0.2',
'Host': 'www.zhihu.com',
'User-Agent':
'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/59.0.3071.86 Safari/537.36',
'Referer': 'http://www.zhihu.com/',
'Cookie': None,
'x-udid': None,
}
self.get_question_content()
def __init__(self):
self.user_list = []
self.mongo = Mongo_1()
def __init__(self):
self.mongo = Mongo_1()
def extract_voters_info():
mongo = Mongo_1()
temp_list = mongo.db.voters_info_1111.find({}, {"user_id": 1})
dt = {list.get('user_id'): "" for list in temp_list}
mongo.client.close()
return dt
# self.mongo.db.question_followers_userDetail_2222.insert(detail)
# self.mongo.db.voters_userDetail_2222.insert(detail)
# self.mongo.db.commenters_userDetail.insert(detail)
self.delLogger(logger)
self.mongo.client.close()
# 删除日志手柄
def delLogger(self, myLogger):
for myHandler in myLogger.handlers:
myHandler.close()
myLogger.removeHandler(myHandler)
if __name__ == '__main__':
userID_list = []
mongo = Mongo_1()
items = mongo.db.answerers.find()
for item in items:
userID_list.append(item.get('user_id'))
mongo.client.close()
ud = UserDetail()
#driver = webdriver.Chrome()
driver = webdriver.PhantomJS(executable_path=r'D:\PhantomJS\phantomjs-2.1.1-windows\bin\phantomjs.exe')
dt = re.sub(r'[^0-9]', '', str(datetime.datetime.now().strftime('%Y-%m-%d-%H:%M:%S')))
for i in range(0, 2367):
ud.get_people_detail(userID_list[i], i , dt)
#time.sleep(1)