class TaskManager():
def __init__(self):
self._oracledb = OracleDB()
self._redisdb = RedisDB()
self._news_url_table = 'news:news_urls'
self._news_urls_dupefilter = 'news:news_urls_dupefilter'
def get_task_count(self):
'''
@summary: redis 中是否有待做的url
---------
---------
@result:
'''
return self._redisdb.zget_count(self._news_url_table)
def get_ever_depth_count(self, total_depth=5):
'''
@summary:
---------
@param total_depth: 不包含。 以客户角度的层数
---------
@result:
'''
depth_count_info = {}
total_count = 0
for depth in range(total_depth):
key = '第%s层url数' % (depth + 1)
depth_count_info[key] = self._redisdb.sget_count(
self._news_urls_dupefilter + str(depth))
total_count += depth_count_info[key]
depth_count_info['总url数'] = total_count
return depth_count_info
def get_task_from_oracle(self):
tasks = []
offset = 0
while True:
# 取任务
task_sql = '''
select *
from (select t.id, t.name, t.position, t.url, t.depth, rownum r
from TAB_IOPM_SITE t
where classify = 1
and t.mointor_status = 701
and (t.position != 35 or t.position is null)
and rownum < {page_size})
where r >= {offset}
'''.format(page_size=offset + ONE_PAGE_SIZE, offset=offset)
results = self._oracledb.find(task_sql)
offset += ONE_PAGE_SIZE
if not results: break
# 拼装成json格式的url
for task in results:
website_id = task[0]
website_name = task[1]
website_position = task[2]
website_url = task[3]
website_domain = tools.get_domain(website_url)
spider_depth = task[4]
remark = {
'website_name': website_name,
'website_position': website_position,
'website_url': website_url,
'website_domain': website_domain,
'spider_depth': spider_depth
}
url_dict = {
'site_id': 1,
'url': website_url,
'depth': 0,
'remark': remark,
'retry_times': 0
}
tasks.append(url_dict)
return tasks
def add_task_to_redis(self, tasks):
for task in tasks:
url = task.get('url')
if url:
url_id = tools.get_sha1(url)
if self._redisdb.sadd(self._news_urls_dupefilter, url_id):
self._redisdb.zadd(self._news_url_table, task, prioritys=0)
# 下面是统计每层url数量用的表
self._redisdb.sadd('news:news_urls_dupefilter0', url_id)
def clear_task(self):
# 清空url指纹表
self._redisdb.sdelete('news:news_urls_dupefilter')
# 下面是统计每层url数量用的表
self._redisdb.sdelete('news:news_urls_dupefilter0')
self._redisdb.sdelete('news:news_urls_dupefilter1')
self._redisdb.sdelete('news:news_urls_dupefilter2')
self._redisdb.sdelete('news:news_urls_dupefilter3')
self._redisdb.sdelete('news:news_urls_dupefilter4')
class CheckNewArticle():
def __init__(self):
self._oracledb = OracleDB()
self._redisdb = RedisDB()
self._wechat_sogo = WechatSogou()
def get_wait_check_account(self):
'''
@summary:
---------
@param :
---------
@result:
'''
# 取抓取完的公众号，且最近发布时间已过去两小时，则再次监测是否又发布新文章
before_tow_hours = tools.timestamp_to_date(
tools.get_current_timestamp() - 60 * 60 * 2)
sql = '''
select t.id,
t.domain,
t.name,
to_char(t.last_article_release_time, 'yyyy-mm-dd hh24:mi:ss'),
t.biz
from TAB_IOPM_SITE t
where t.biz is not null
and mointor_status = 701
and t.spider_status = 603
and (t.last_article_release_time is null or
t.last_article_release_time <=
to_date('{}', 'yyyy-mm-dd hh24:mi:ss'))
'''.format(before_tow_hours)
accounts = self._oracledb.find(sql)
# 若无抓取完的公众号，且redis中无抓取任务，则数据库中非603任务可能为丢失任务，需要重新下发
if not accounts and not self._redisdb.sget_count('wechat:account'):
sql = '''
select t.id,
t.domain,
t.name,
to_char(t.last_article_release_time, 'yyyy-mm-dd hh24:mi:ss'),
t.biz
from TAB_IOPM_SITE t
where t.biz is not null
and mointor_status = 701
and t.spider_status ！= 603
'''
accounts = self._oracledb.find(sql)
return accounts
def check_new_article(self, account):
oralce_id, account_id, account_name, last_article_release_time, biz = account
article_release_time = self._wechat_sogo.get_article_release_time(
account_id=account_id, account=account_name)
print(article_release_time)
if article_release_time:
last_article_release_time = last_article_release_time or ''
if article_release_time >= tools.get_current_date(
'%Y-%m-%d'
) and article_release_time > last_article_release_time:
print('{} 有新文章发布，等待抓取。 发布时间：{}'.format(account_name,
article_release_time))
sql = '''
update TAB_IOPM_SITE t set t.spider_status = 601,
t.last_article_release_time =
to_date('{}', 'yyyy-mm-dd hh24:mi:ss')
where id = {}
'''.format(article_release_time, oralce_id)
# 多线程， 数据库需每个线程持有一个
oracledb = OracleDB()
oracledb.update(sql)
oracledb.close()
# 入redis， 作为微信爬虫的任务池
data = (oralce_id, account_id, account_name,
last_article_release_time, biz)
self._redisdb.sadd('wechat:account', data)