refactor: centralize scheduler initialization and improve logging in job files

apps/common/job/clean_chat_job.py

@@ -2,20 +2,16 @@
 
 import datetime
 
-from apscheduler.schedulers.background import BackgroundScheduler
 from django.db import transaction
 from django.db.models import Q, Max
 from django.utils import timezone
-from django_apscheduler.jobstores import DjangoJobStore
 
 from application.models import Application, Chat, ChatRecord
+from common.job.scheduler import scheduler
 from common.utils.lock import try_lock, un_lock, lock
 from common.utils.logger import maxkb_logger
 from knowledge.models import File
 
-scheduler = BackgroundScheduler()
-scheduler.add_jobstore(DjangoJobStore(), "default")
-
 
 def clean_chat_log_job():
     clean_chat_log_job_lock()
@@ -76,7 +72,8 @@ def clean_chat_log_job_lock():
 def run():
     if try_lock('clean_chat_log_job', 30 * 30):
         try:
-            scheduler.start()
+            maxkb_logger.info('get lock clean_chat_log_job')
+
             existing_job = scheduler.get_job(job_id='clean_chat_log')
             if existing_job is not None:
                 existing_job.remove()

apps/common/job/clean_debug_file_job.py

@@ -1,19 +1,14 @@
 # coding=utf-8
-import time
 from datetime import timedelta
 
-from apscheduler.schedulers.background import BackgroundScheduler
 from django.db.models import Q
 from django.utils import timezone
-from django_apscheduler.jobstores import DjangoJobStore
 
+from common.job.scheduler import scheduler
 from common.utils.lock import un_lock, try_lock, lock
 from common.utils.logger import maxkb_logger
 from knowledge.models import File, FileSourceType
 
-scheduler = BackgroundScheduler()
-scheduler.add_jobstore(DjangoJobStore(), "default")
-
 
 def clean_debug_file():
     clean_debug_file_lock()
@@ -22,7 +17,7 @@ def clean_debug_file():
 @lock(lock_key='clean_debug_file_execute', timeout=30)
 def clean_debug_file_lock():
     from django.utils.translation import gettext_lazy as _
-    maxkb_logger.debug(_('start clean debug file'))
+    maxkb_logger.info(_('start clean debug file'))
     minutes_30_ago = timezone.now() - timedelta(minutes=30)
     two_hours_ago = timezone.now() - timedelta(hours=2)
     one_days_ago = timezone.now() - timedelta(hours=24)
@@ -31,17 +26,18 @@ def clean_debug_file_lock():
         Q(create_time__lt=one_days_ago, source_type=FileSourceType.TEMPORARY_1_DAY.value) |
         Q(create_time__lt=two_hours_ago, source_type=FileSourceType.TEMPORARY_120_MINUTE.value) |
         Q(create_time__lt=minutes_30_ago, source_type=FileSourceType.TEMPORARY_30_MINUTE.value)).delete()
-    maxkb_logger.debug(_('end clean debug file'))
-    time.sleep(2)
+    maxkb_logger.info(_('end clean debug file'))
+    # time.sleep(2)
 
 
 def run():
     if try_lock('clean_debug_file', 30 * 30):
         try:
-            scheduler.start()
+            maxkb_logger.info('get lock clean_debug_file')
+
             clean_debug_file_job = scheduler.get_job(job_id='clean_debug_file')
             if clean_debug_file_job is not None:
                 clean_debug_file_job.remove()
-            scheduler.add_job(clean_debug_file, 'cron', hour='*', minute='*/30', second='0', id='clean_debug_file')
+            scheduler.add_job(clean_debug_file, 'cron', hour='*', minute='*/1', second='0', id='clean_debug_file')
         finally:
             un_lock('clean_debug_file')

apps/common/job/client_access_num_job.py

@@ -7,17 +7,13 @@
     @desc:
 """
 
-from apscheduler.schedulers.background import BackgroundScheduler
 from django.db.models import QuerySet
-from django_apscheduler.jobstores import DjangoJobStore
 
 from application.models import ApplicationChatUserStats
+from common.job.scheduler import scheduler
 from common.utils.lock import try_lock, un_lock, lock
 from common.utils.logger import maxkb_logger
 
-scheduler = BackgroundScheduler()
-scheduler.add_jobstore(DjangoJobStore(), "default")
-
 
 def client_access_num_reset_job():
     client_access_num_reset_job_lock()
@@ -34,7 +30,8 @@ def client_access_num_reset_job_lock():
 def run():
     if try_lock('access_num_reset', 30 * 30):
         try:
-            scheduler.start()
+            maxkb_logger.info('get lock access_num_reset')
+
             access_num_reset = scheduler.get_job(job_id='access_num_reset')
             if access_num_reset is not None:
                 access_num_reset.remove()

apps/common/job/scheduler.py

@@ -0,0 +1,12 @@
+from apscheduler.schedulers.background import BackgroundScheduler
+from django_apscheduler.jobstores import DjangoJobStore
+
+scheduler = BackgroundScheduler()
+scheduler.add_jobstore(DjangoJobStore(), "default")
+
+try:
+    scheduler.start()
+except Exception as e:
+    from common.utils.logger import maxkb_logger
+
+    maxkb_logger.error(f"Failed to start scheduler: {e}")