optimize the unused dataset clean logic

api/.env.example

@@ -256,3 +256,7 @@ WORKFLOW_CALL_MAX_DEPTH=5
 # App configuration
 APP_MAX_EXECUTION_TIME=1200
 APP_MAX_ACTIVE_REQUESTS=0
+
+
+# Celery beat configuration
+CELERY_BEAT_SCHEDULER_TIME=1

api/configs/feature/__init__.py

@@ -23,6 +23,7 @@ class SecurityConfig(BaseSettings):
         default=24,
     )
 
+
 class AppExecutionConfig(BaseSettings):
     """
     App Execution configs
@@ -435,6 +436,13 @@ class ImageFormatConfig(BaseSettings):
     )
 
 
+class CeleryBeatConfig(BaseSettings):
+    CELERY_BEAT_SCHEDULER_TIME: int = Field(
+        description='the time of the celery scheduler, default to 1 day',
+        default=1,
+    )
+
+
 class FeatureConfig(
     # place the configs in alphabet order
     AppExecutionConfig,
@@ -462,5 +470,6 @@ class FeatureConfig(
 
     # hosted services config
     HostedServiceConfig,
+    CeleryBeatConfig,
 ):
     pass

api/extensions/ext_celery.py

@@ -43,15 +43,15 @@ def init_app(app: Flask) -> Celery:
         "schedule.clean_embedding_cache_task",
         "schedule.clean_unused_datasets_task",
     ]
-
+    day = app.config["CELERY_BEAT_SCHEDULER_TIME"]
     beat_schedule = {
         'clean_embedding_cache_task': {
             'task': 'schedule.clean_embedding_cache_task.clean_embedding_cache_task',
-            'schedule': timedelta(days=1),
+            'schedule': timedelta(days=day),
         },
         'clean_unused_datasets_task': {
             'task': 'schedule.clean_unused_datasets_task.clean_unused_datasets_task',
-            'schedule': timedelta(minutes=3),
+            'schedule': timedelta(minutes=1),
         }
     }
     celery_app.conf.update(

api/schedule/clean_unused_datasets_task.py

@@ -9,7 +9,7 @@ from configs import dify_config
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
 from extensions.ext_database import db
 from models.dataset import Dataset, DatasetQuery, Document
-
+from sqlalchemy import func
 
 @app.celery.task(queue='dataset')
 def clean_unused_datasets_task():
@@ -20,10 +20,46 @@ def clean_unused_datasets_task():
     page = 1
     while True:
         try:
-            datasets = db.session.query(Dataset).filter(Dataset.created_at < thirty_days_ago) \
+            # Subquery for counting new documents
-                .order_by(Dataset.created_at.desc()).paginate(page=page, per_page=50)
+            document_subquery_new = db.session.query(
+                Document.dataset_id,
+                func.count(Document.id).label('document_count')
+            ).filter(
+                Document.indexing_status == 'completed',
+                Document.enabled == True,
+                Document.archived == False,
+                Document.updated_at > thirty_days_ago
+            ).group_by(Document.dataset_id).subquery()
+
+            # Subquery for counting old documents
+            document_subquery_old = db.session.query(
+                Document.dataset_id,
+                func.count(Document.id).label('document_count')
+            ).filter(
+                Document.indexing_status == 'completed',
+                Document.enabled == True,
+                Document.archived == False,
+                Document.updated_at < thirty_days_ago
+            ).group_by(Document.dataset_id).subquery()
+
+            # Main query with join and filter
+            datasets = (db.session.query(Dataset)
+                        .outerjoin(
+                document_subquery_new, Dataset.id == document_subquery_new.c.dataset_id
+            ).outerjoin(
+                document_subquery_old, Dataset.id == document_subquery_old.c.dataset_id
+            ).filter(
+                Dataset.created_at < thirty_days_ago,
+                func.coalesce(document_subquery_new.c.document_count, 0) == 0,
+                func.coalesce(document_subquery_old.c.document_count, 0) > 0
+            ).order_by(
+                Dataset.created_at.desc()
+            ).paginate(page=page, per_page=50))
+
         except NotFound:
             break
+        if datasets.items is None or len(datasets.items) == 0:
+            break
         page += 1
         for dataset in datasets:
             dataset_query = db.session.query(DatasetQuery).filter(
@@ -31,14 +67,6 @@ def clean_unused_datasets_task():
                 DatasetQuery.dataset_id == dataset.id
             ).all()
             if not dataset_query or len(dataset_query) == 0:
-                documents = db.session.query(Document).filter(
-                    Document.dataset_id == dataset.id,
-                    Document.indexing_status == 'completed',
-                    Document.enabled == True,
-                    Document.archived == False,
-                    Document.updated_at > thirty_days_ago
-                ).all()
-                if not documents or len(documents) == 0:
                 try:
                     # remove index
                     index_processor = IndexProcessorFactory(dataset.doc_form).init_index_processor()