def fail_action(self, values):
'''
消息动作处理失败之后，更改队列中间件中该消息的失败次数并记录执行机器的IP
如果达到该机器的最大尝试失败次数，则更改队列中间件中该消息的状态为未处理，目的让其它机器再次尝试去处理该消息
:param values: 消息动作处理之后的结果
'''
update_sql = """
update hainiu_queue set fail_times=fail_times+1,fail_ip='%s' where id=%s;
"""
update_sql_1 = """
update hainiu_queue set type=1 where id=%s
"""
try:
d = DBUtil(config._HAINIU_DB)
id = values[0]
u = Util()
ip = u.get_local_ip()
sql = update_sql % (ip, id)
d.execute_no_commit(sql)
if (self.try_num == Consumer._WORK_TRY_NUM):
sql = update_sql_1 % id
d.execute_no_commit(sql)
d.commit()
except:
self.rl.exception()
self.rl.error(sql)
d.rollback()
finally:
d.close()
def put_inner_to_queue():
redis_util = RedisUtill()
'''
'''
page_show_num = 10
# 统计hainiu_queue 未处理的记录数
select_queue_count_sql = """
select count(*) from hainiu_queue where type=%s and is_work=0 and fail_times=0;
"""
# 插入hainiu_queue表
insert_queue_sql = """
insert into hainiu_queue (type,action,params) values (%s, %s, %s);
"""
logger = LogUtil().get_logger("download_news_queue", "download_news_queue")
db_util = DBUtil(_HAINIU_DB)
db_util.execute_no_commit("set NAMES utf8mb4;")
try:
# 统计hainiu_queue 未处理的记录数
sql_params = [2]
res1 = db_util.read_one(select_queue_count_sql, sql_params)
queue_count = res1[0]
start_time = time.time()
if queue_count >= 5:
logger.info("hainiu_queue 有 %d 条未处理的记录，不需要导入！" % queue_count)
return None
inner_count = 0
for ip in ips:
key_list = []
scan_limit_to_queue_table(ip, port, 0, 'down:*', 20, key_list)
inner_count = inner_count + len(key_list)
# 根据key列表上Redis里获取value列表
values = redis_util.get_values_batch_keys(key_list)
# 导入hainiu_queue表
insert_queue_record = []
for value in values:
queue_param = json.loads(value)
a_url = queue_param['a_url']
insert_queue_record.append((2, a_url, value))
db_util.executemany_no_commit(insert_queue_sql,
insert_queue_record)
db_util.commit()
# 把导入表后的key列表从redis里删掉
redis_util.delete_batch(key_list)
end_time = time.time()
run_time = end_time - start_time
logger.info("本地导入 %d 条数据， 用时 %.2f 秒" % (inner_count, run_time))
except Exception, e:
traceback.print_exc(e)
db_util.rollback()
def queue_items(self):
'''
通过悲观锁+事务+更新状态来实现多个机器串行拿取数据，
并把其封装成HainiuConsumerAction对象实例列表返回
'''
select_sql = """
select id,action,params
from hainiu_queue where type=%s and is_work=%s and fail_ip!=%s and fail_times<%s limit %s for update;
"""
# 更新SQL-拼字符串
update_sql = """
update hainiu_queue set is_work=1 where id in (%s);
"""
c_actions = []
# 用于装id，来更新
ids = []
db_util = DBUtil(_HAINIU_DB)
try:
# sql_params = [1, 0, _QUEUE_NEWS_FIND['MAX_FAIL_TIMES'], _QUEUE_NEWS_FIND['LIMIT_NUM']]
# 屏蔽ip查询的参数
ip = Util().get_local_ip()
sql_params = [
1, 0, ip, _QUEUE_NEWS_FIND['MAX_FAIL_TIMES'],
_QUEUE_NEWS_FIND['LIMIT_NUM']
]
# ({},{})
res1 = db_util.read_dict(select_sql, sql_params)
for row in res1:
id = row['id']
ids.append(str(id))
act = row['action']
params = row['params']
c_action = NewsFindConsumerAction(id, act, params)
c_actions.append(c_action)
if len(ids) > 0:
db_util.execute_no_commit(update_sql % ",".join(ids))
db_util.commit()
except Exception, e:
db_util.rollback()
traceback.print_exc(e)
def success_action(self, values):
#1）记录种子url最后爬取成功数， （用来校验最后的爬取是否成功）；
#2）在hainiu_queue 表中删除已经爬取成功的url；
seed_update_sql = """
update hainiu_web_seed set last_crawl_internally=%s,last_crawl_externally=%s,last_crawl_time=now() where md5=%s;
"""
queue_delete_sql = """
delete from hainiu_queue where id=%s
"""
db_util = DBUtil(_HAINIU_DB)
try:
sql_param = [values[1], values[2], values[3]]
db_util.execute_no_commit(seed_update_sql, sql_param)
sql_param = [values[0]]
db_util.execute_no_commit(queue_delete_sql, sql_param)
db_util.commit()
except Exception, e:
traceback.print_exc(e)
db_util.rollback()
def queue_items(self):
# 屏蔽ip的查询方式
select_sql='''
select id, action, params from web_queue where type=%s
and is_work=%s and fail_ip != %s and fail_times < %s limit 0, %s for update;
'''
update_sql='''
update web_queue set is_work=1 where id in(%s);
'''
db_util = DBUtil(_ZZ_DB)
try:
ip = Util().get_local_ip()
sql_params = [1, 0, ip, _QUEUE_ZZ["MAX_FAIL_TIMES"], _QUEUE_ZZ["LIMIT_NUM"]]
res = db_util.read_dict(select_sql, sql_params)
actions = []
ids = []
for row in res:
id = row["id"]
ids.append(str(id))
action = row["action"]
params = row["params"]
# 封装对象
c_action = WebConsumerAction(id, action, params)
actions.append(c_action)
if len(actions) != 0:
# 更新 is_work=1
db_util.execute_no_commit(update_sql % ",".join(ids))
db_util.commit()
except Exception, err:
actions = []
db_util.rollback()
traceback.print_exc(err)
def fail_action(self, values):
# 每次失败都需要更新ip 和 失败次数
update_sql1='''
update web_queue set fail_ip = %s , fail_times = fail_times + 1 where id = %s;
'''
# 当失败次数到达每台机器的最大重试次数，就将该记录的is_work=0 ,让其重试
update_sql2='''
update web_queue set is_work = 0 where id = %s;
'''
# 更新seed表状态
update_seed_sql = '''
update web_seed set fail_times=fail_times + 1,fail_ip=%s where md5 =%s;
'''
# 更新externally表状态
update_exter_sql = '''
update web_seed_externally set fail_times=fail_times + 1,fail_ip=%s where a_md5 =%s;
'''
db_util = DBUtil(_ZZ_DB)
try:
id = values[0]
ip = Util().get_local_ip()
# 每次更新失败ip 和失败次数
# queue表
sql_params = [ip, id]
db_util.execute_no_commit(update_sql1, sql_params)
# seed 表
sql_params = [ip, values[1]]
db_util.execute(update_seed_sql, sql_params)
# externally表
db_util.execute(update_exter_sql, sql_params)
if self.current_retry_num == _QUEUE_ZZ["C_RETRY_TIMES"] - 1:
db_util.execute_no_commit(update_sql2 % id)
db_util.commit()
except Exception,err:
db_util.rollback()
traceback.print_exc(err)
def success_action(self, values):
delete_sql = """
delete from hainiu_queue where id=%s;
"""
update_hainiu_news_seed_sql = """
update hainiu_web_seed set last_crawl_internally=%s,last_crawl_externally=%s,last_crawl_time=now() where md5="%s";"""
try:
d = DBUtil(config._HAINIU_DB)
id = values[5]
sql = delete_sql % id
d.execute_no_commit(sql)
sql = update_hainiu_news_seed_sql % (values[3],values[4],values[0])
d.execute_no_commit(sql)
d.commit()
except:
self.rl.exception()
self.rl.error(sql)
d.rollback()
d.commit()
finally:
d.close()
def fail_action(self, values):
update_sql = """
update hainiu_queue set fail_times=fail_times+1,fail_ip='%s' where id=%s;
"""
update_sql_1 = """
update hainiu_queue set type=1 where id=%s;
"""
update_hainiu_news_seed_sql = """
update hainiu_web_seed set fail_times=fail_times+1,fail_ip="%s" where md5="%s";
"""
try:
d = DBUtil(config._HAINIU_DB)
id = values[5]
u = Util()
ip = u.get_local_ip()
sql = update_sql % (ip, id)
d.execute_no_commit(sql)
main_md5 = values[0]
sql = update_hainiu_news_seed_sql % (ip, main_md5)
d.execute_no_commit(sql)
if (self.try_num == Consumer._WORK_TRY_NUM):
sql = update_sql_1 % (id)
d.execute_no_commit(sql)
d.commit()
except:
self.rl.exception()
self.rl.error(sql)
d.rollback()
d.commit()
finally:
d.close()
def fail_action(self, values):
ip = Util().get_local_ip()
db_util = DBUtil(_HAINIU_DB)
#1）记录hainiu_queue表错误次数和ip；
# is_success,self.id,len(inner_list),len(exter_list),md5
queue_update_sql1 = """
update hainiu_queue set fail_times=fail_times+1,fail_ip=%s where id=%s;
"""
#2）当某个机器的错误次数达到了当前机器设定的最大重试次数，把hainiu_queue
# 表对应的记录的 is_work = 0，让其他机器重试；
queue_update_sql2 = """
update hainiu_queue set is_work=0 where id=%s;
"""
#3）更新种子表的失败次数、失败ip；队列表的数据不删除，有可能是因为目标网站把ip给封了，
# 在某个时间，写个脚本，把失败的队列数据改状态和失败次数和失败ip，重新爬取试试。
seed_update_sql = """
update hainiu_web_seed set fail_times=fail_times+1,fail_ip=%s where md5=%s
"""
try:
sql_params = [ip, values[0]]
db_util.execute_no_commit(queue_update_sql1, sql_params)
# 比较失败次数
if self.current_retry_num == _QUEUE_NEWS_FIND['C_RETRY_TIMES'] - 1:
sql_params = [self.id]
db_util.execute_no_commit(queue_update_sql2, sql_params)
sql_params = [ip, values[3]]
db_util.execute_no_commit(seed_update_sql, sql_params)
db_util.commit()
except Exception, e:
traceback.print_exc(e)
db_util.rollback()
def fail_action(self, values):
ip = Util().get_local_ip()
db_util = DBUtil(_HAINIU_DB)
#1）记录队列表错误次数和ip；
queue_update_sql1 = """
update hainiu_queue set fail_times=fail_times+1,fail_ip=%s where id=%s;
"""
#2）当某个机器的错误次数达到了当前机器设定的最大重试次数，把hainiu_queue 表对应的记录的
#is_work = 0，让其他机器重试；
queue_update_sql2 = """
update hainiu_queue set is_work=0 where id=%s;
"""
#3）更新内链表的失败次数和失败ip，队列表的数据不删除；
inner_update_sql = """
update hainiu_web_seed_internally set fail_times=fail_times+1,fail_ip=%s where md5=%s and a_md5=%s
"""
try:
# 1)
sql_params = [ip, values[0]]
db_util.execute_no_commit(queue_update_sql1, sql_params)
# 2)
# 比较失败次数
if self.current_retry_num == _QUEUE_NEWS_FIND['C_RETRY_TIMES'] - 1:
sql_params = [self.id]
db_util.execute_no_commit(queue_update_sql2, sql_params)
sql_params = [ip, values[1], values[2]]
db_util.execute_no_commit(inner_update_sql, sql_params)
db_util.commit()
except Exception, e:
db_util.rollback()
traceback.print_exc(e)
def success_action(self, values):
db_util = DBUtil(_HAINIU_DB)
time_util = TimeUtil()
#1）在hainiu_queue 表中删除已经下载成功的url；
queue_delete_sql = """
delete from hainiu_queue where id=%s
"""
#2）更新内链表的最后更新时间；
inner_update_sql = """
update hainiu_web_seed_internally set update_time= %s where a_md5=%s AND
md5=%s
"""
update_time = time_util.get_timestamp()
try:
sql_param = [values[0]]
db_util.execute_no_commit(queue_delete_sql, sql_param)
sql_param = [update_time, values[1], values[2]]
db_util.execute_no_commit(inner_update_sql, sql_param)
db_util.commit()
except Exception, e:
traceback.print_exc(e)
db_util.rollback()
def action(self):
is_success = True
t = TimeUtil()
u = Util()
hu = HtmlUtil()
r = RequestUtil()
in_values = []
ex_values = []
a_href = ''
main_md5 = u.get_md5(self.url)
now_time = datetime.now()
update_time = int(time.mktime(now_time.timetuple()))
create_time = update_time
create_day = int(t.now_day().replace('-', ''))
create_hour = int(t.now_hour())
try:
html = r.http_get_phandomjs(self.url)
domain = get_tld(self.url)
soup = BeautifulSoup(html, 'lxml')
a_docs = soup.find_all("a")
a_set = set()
a_param = {}
out_json_srt = ''
status = 0
host = hu.get_url_host(self.url)
for a in a_docs:
a_href = self.get_format_url(a,host)
a_title = a.get_text().strip()
if a_href == '' or a_title == '':
continue
if a_set.__contains__(a_href):
continue
a_set.add(a_href)
req = urllib2.Request(url=a_href)
a_host = req.get_host() if req.get_host() is not None else ''
a_md5 = u.get_md5(a_href)
if a_title != '':
a_param['title'] = a_title
out_json_srt = json.dumps(a_param,ensure_ascii=False)
a_xpath = hu.get_dom_parent_xpath_js(a)
insert_values = (main_md5,domain,host,a_md5,a_host,a_xpath,create_time,create_day,create_hour,update_time,status,
MySQLdb.escape_string(self.url),
MySQLdb.escape_string(a_href),
MySQLdb.escape_string(a_title),
out_json_srt)
if a_host.__contains__(domain):
in_values.append(insert_values)
else:
ex_values.append(insert_values)
in_table = 'hainiu_web_seed_internally'
ex_table = 'hainiu_web_seed_externally'
insert_sql = """
insert into <table> (md5,domain,host,a_md5,a_host,a_xpath,create_time,create_day,create_hour,update_time,status,url,a_url,a_title,param)
values (%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s,%s) ON DUPLICATE KEY UPDATE update_time=update_time;
"""
try:
d = DBUtil(config._HAINIU_DB)
d.execute_no_commit("set NAMES utf8mb4;")
if in_values.__len__() != 0:
sql = insert_sql.replace('<table>',in_table)
d.executemany_no_commit(sql,in_values)
if ex_values.__len__() != 0:
sql = insert_sql.replace('<table>',ex_table)
d.executemany_no_commit(sql,ex_values)
d.commit()
except:
is_success = False
self.rl.exception()
self.rl.error(sql)
d.rollback()
finally:
d.close()
except:
is_success = False
self.rl.exception()
finally:
r.close_phandomjs()
return super(self.__class__, self).result(is_success, [main_md5,self.url,a_href,in_values.__len__(),ex_values.__len__(),self.queue_id])
def put_queue(page_show_num):
db_util = DBUtil(_ZZ_DB)
# 统计queue符合条件的记录数
count_queue_sql = '''
select count(*) from web_queue where is_work=%s and fail_times < %s;
'''
# 统计web_seed表的符合条件的总记录数
count_seed_sql = '''
select count(*) from web_seed where status=0;
'''
# 分页查询web_seed 表的记录
select_seed_limit_sql = '''
select id,url,category from web_seed where status=0 limit %s,%s;
'''
# 插入queue表记录
insert_queue_sql = '''
insert into web_queue (type,action,params) values(%s,%s,%s);
'''
# 更新web_seed表中的 status
update_sql = '''
update web_seed set status=1 where id in(%s);
'''
try:
sql_params = [0, _QUEUE_ZZ["MAX_FAIL_TIMES"]]
res1 = db_util.read_one(count_queue_sql, sql_params)
total_num1 = res1[0]
if total_num1 != 0:
print "queue has %d records,not insert!" % total_num1
return None
start_time = time.time()
# 统计web_seed 表符合条件的总记录数
res2 = db_util.read_one(count_seed_sql)
total_num2 = res2[0]
# 计算分多少页查询
page_num = total_num2 / page_show_num if total_num2 % page_show_num == 0 else total_num2 / page_show_num + 1
# 分页查询
ids = []
for i in range(0, page_num):
sql_params = [i * page_show_num, page_show_num]
print sql_params
res3 = db_util.read_dict(select_seed_limit_sql, sql_params)
list1 = []
for row in res3:
id = row["id"]
ids.append(str(id))
action = row["url"]
params = row["category"]
type = 1
list1.append((type, action, params))
# 批量插入queue
db_util.executemany(insert_queue_sql, list1)
# 更新 status=1
db_util.execute_no_commit(update_sql % ",".join(ids))
db_util.commit()
end_time = time.time()
run_time = end_time - start_time
print "total_num:%d, run_time:%.2f" % (total_num2, run_time)
except Exception, err:
db_util.rollback()
traceback.print_exc(err)
#-*- encoding: utf-8 -*-
'''
db_test.py
Created on 2019/6/25 11:14
Copyright (c) 2019/6/25, 海牛学院版权所有.
@author: 潘牛
'''
from commons.util.db_util import DBUtil
from configs.config import _HAINIU_DB
db_util = DBUtil(_HAINIU_DB)
# 设置字符集是utf8mb4
db_util.execute_no_commit("set NAMES utf8mb4;")
# 测试 execute(self,sql,params = None):
# sql = """
# insert into hainiu_queue (type,action,params) values (1, 'www.hainiubl.com', 'aa');
# """
# db_util.execute(sql)
# 字符串拼接（不推荐用）
# sql = """
# insert into hainiu_queue (type,action,params) values (%d, '%s', '%s');
# """ % (1, 'www.hainiubl.com', 'aa')
# db_util.execute(sql)
# -------------------------------------
# 测试 execute(self,sql,params != None):
# sql占位符（推荐用法）
# sql = """
