def excute_search_task():
keywords = get_search_keywords()
for each in keywords:
celery.send_task('celery_tasks.weibo.search.search_keyword',
args=(each[0], each[1]),
queue='search_crawler',
routing_key='for_search_info')
def excute_user_task():
seeds = get_seed_ids()
if seeds:
for seed in seeds:
celery.send_task('celery_tasks.weibo.user.crawl_person_infos',
args=(seed.uid, ),
queue='user_crawler',
routing_key='for_user_info')
def excute_comment_task():
# 只解析了根评论，而未对根评论下的评论进行抓取，如果有需要的同学，可以适当做修改
weibo_datas = wb_data.get_weibo_comment_not_crawled()
for weibo_data in weibo_datas:
celery.send_task('celery_tasks.weibo.comment.crawl_comment_page',
args=(weibo_data.weibo_id, ),
queue='comment_crawler',
routing_key='comment_info')
def excute_home_task():
# 这里的策略由自己指定，可以基于已有用户做主页抓取，也可以指定一些用户,我这里直接选的种子数据库中的uid
id_objs = get_home_ids()
for id_obj in id_objs:
celery.send_task('celery_tasks.weibo.home.crawl_weibo_datas',
args=(id_obj.uid, ),
queue='home_crawler',
routing_key='home_info')
def excute_user_personal_adver_task(adv_message):
seeds = get_seed()
if seeds:
for seed in seeds:
sleep(random.randint(1, 6))
print(seed.uid)
celery.send_task('celery_tasks.weibo.user.excute_personal_adver',
args=(seed.uid, adv_message),
queue='personal_adver',
routing_key='for_adver')
def excute_login_task():
infos = login_info.get_login_info()
# Clear all stacked login celery_tasks before each time for login
Cookies.check_login_task()
log.crawler.info('The excute_login_task is starting...')
for info in infos:
celery.send_task('celery_tasks.weibo.login.login_task',
args=(info.name, info.password, info.source),
queue='login_task',
routing_key='login_task')
time.sleep(10)
def excute_repost_task():
# regard current weibo url as the original url, you can also analyse from the root url
weibo_datas = wb_data.get_weibo_repost_not_crawled()
crawler.info('There are {} repost urls have to be crawled'.format(
len(weibo_datas)))
for weibo_data in weibo_datas:
celery.send_task('celery_tasks.weibo.repost.crawl_repost_page',
args=(weibo_data.weibo_id, weibo_data.uid),
queue='repost_crawler',
routing_key='repost_info')
def crawl_comment_page(mid):
limit = conf.get_max_comment_page() + 1
# 这里为了马上拿到返回结果，采用本地调用的方式
first_page = crawl_comment_by_page(mid, 1)
total_page = comment.get_total_page(first_page)
if total_page < limit:
limit = total_page + 1
for page_num in range(2, limit):
celery.send_task('celery_tasks.weibo.comment.crawl_comment_by_page',
args=(mid, page_num),
queue='comment_page_crawler',
routing_key='comment_page_info')
def crawl_person_infos(uid):
"""
Crawl user info and their fans and followers
For the limit of weibo's backend, we can only crawl 5 pages of the fans and followers.
We also have no permissions to view enterprise's followers and fans info
:param uid: current user id
:return: None
"""
if not uid:
return
user = user_get.get_profile(uid)
# If it's enterprise user, just skip it
if user.verify_type == 2:
set_seed_other_crawled(uid)
return
# Crawl fans and followers
celery.send_task('celery_tasks.weibo.user.crawl_follower_fans',
args=(uid, ),
queue='fans_followers',
routing_key='for_fans_followers')
def search_keyword(keyword, keyword_id):
cur_page = 1
encode_keyword = url_parse.quote(keyword)
while cur_page < limit:
cur_url = url.format(encode_keyword, cur_page)
search_page = get_page(cur_url)
if not search_page:
crawler.warning(
'No result for keyword {}, the source page is {}'.format(
keyword, search_page))
return
search_list = parse_search.get_search_info(search_page)
# Because the search results are sorted by time, if any result has been stored in mysql,
# we need not crawl the same keyword in this turn
for wb_data in search_list:
rs = get_wb_by_mid(wb_data.weibo_id)
if rs:
crawler.info(
'keyword {} has been crawled in this turn'.format(keyword))
return
else:
insert_weibo_data(wb_data)
insert_keyword_wbid(keyword_id, wb_data.weibo_id)
# send task for crawling user info
celery.send_task('celery_tasks.weibo.user.crawl_person_infos',
args=(wb_data.uid, ),
queue='user_crawler',
routing_key='for_user_info')
if 'page next S_txt1 S_line1' in search_page:
cur_page += 1
else:
crawler.info(
'keyword {} has been crawled in this turn'.format(keyword))
return
def jd_seckill_timer_relogin(self):
task_id = self.request.id
logger = cl.getLogger('jd_seckill_timer_relogin')
logger.info('重新登录任务开始，task_id：' + task_id)
ppool = ProxyStore.get_proxyPoolstores()
mongdb_conn = DBStore.get_datastores()
mydb = mongdb_conn['JD']
jd_users = mydb.Users.find({"status": 1})
for jd_user in jd_users:
fetch_result = celery.send_task("celery_tasks.jd_seckill.jd_seckill.jd_seckill_relogin_task",
queue='jd_seckill_presell',
args=(jd_user["username"], jd_user["password"], 4099139, ppool.getProxy()))
def crawl_weibo_datas(uid):
limit = get_max_home_page()
cur_page = 1
while cur_page <= limit:
url = home_url.format(uid, cur_page)
html = get_page(url)
weibo_datas = get_wbdata_fromweb(html)
if not weibo_datas:
crawler.warning('用户id为{}的用户主页微博数据未采集成功，请检查原因'.format(uid))
return
insert_weibo_datas(weibo_datas)
domain = public.get_userdomain(html)
cur_time = int(time.time() * 1000)
ajax_url_0 = ajax_url.format(domain, 0, domain, uid, cur_page,
cur_page, cur_time)
ajax_url_1 = ajax_url.format(domain, 1, domain, uid, cur_page,
cur_page, cur_time + 100)
if cur_page == 1:
total_page = get_total_page(crawl_ajax_page(ajax_url_1))
if total_page < limit:
limit = total_page
cur_page += 1
celery.send_task('celery_tasks.weibo.home.crawl_ajax_page',
args=(ajax_url_0, ),
queue='ajax_home_crawler',
routing_key='ajax_home_info')
celery.send_task('celery_tasks.weibo.home.crawl_ajax_page',
args=(ajax_url_1, ),
queue='ajax_home_crawler',
routing_key='ajax_home_info')
# coding:utf-8
from apps.celery_init import celery
from celery_tasks.weibo import login
if __name__ == '__main__':
# you should execute this file, because celery timer will execute login delayed
# login.excute_login_task()
celery.send_task('apps.celery_init.start_add_task',
args=(1, 2),
queue='start_add_task',
routing_key='start_add_task')