deal db session in celery worker (#17549)

api/tasks/add_document_to_index_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.index_processor.constant.index_type import IndexType
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
@@ -28,7 +27,9 @@ def add_document_to_index_task(dataset_document_id: str):
 
     dataset_document = db.session.query(DatasetDocument).filter(DatasetDocument.id == dataset_document_id).first()
     if not dataset_document:
-        raise NotFound("Document not found")
+        logging.info(click.style("Document not found: {}".format(dataset_document_id), fg="red"))
+        db.session.close()
+        return
 
     if dataset_document.indexing_status != "completed":
         return

api/tasks/annotation/add_annotation_to_index_task.py

@@ -6,6 +6,7 @@ from celery import shared_task  # type: ignore
 
 from core.rag.datasource.vdb.vector_factory import Vector
 from core.rag.models.document import Document
+from extensions.ext_database import db
 from models.dataset import Dataset
 from services.dataset_service import DatasetCollectionBindingService
 
@@ -55,3 +56,5 @@ def add_annotation_to_index_task(
         )
     except Exception:
         logging.exception("Build index for annotation failed")
+    finally:
+        db.session.close()

api/tasks/annotation/batch_import_annotations_task.py

@@ -88,3 +88,5 @@ def batch_import_annotations_task(job_id: str, content_list: list[dict], app_id:
             indexing_error_msg_key = "app_annotation_batch_import_error_msg_{}".format(str(job_id))
             redis_client.setex(indexing_error_msg_key, 600, str(e))
             logging.exception("Build index for batch import annotations failed")
+        finally:
+            db.session.close()

api/tasks/annotation/delete_annotation_index_task.py

@@ -5,6 +5,7 @@ import click
 from celery import shared_task  # type: ignore
 
 from core.rag.datasource.vdb.vector_factory import Vector
+from extensions.ext_database import db
 from models.dataset import Dataset
 from services.dataset_service import DatasetCollectionBindingService
 
@@ -39,3 +40,5 @@ def delete_annotation_index_task(annotation_id: str, app_id: str, tenant_id: str
         )
     except Exception as e:
         logging.exception("Annotation deleted index failed")
+    finally:
+        db.session.close()

api/tasks/annotation/disable_annotation_reply_task.py

@@ -3,7 +3,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.datasource.vdb.vector_factory import Vector
 from extensions.ext_database import db
@@ -23,14 +22,18 @@ def disable_annotation_reply_task(job_id: str, app_id: str, tenant_id: str):
     app = db.session.query(App).filter(App.id == app_id, App.tenant_id == tenant_id, App.status == "normal").first()
     annotations_count = db.session.query(MessageAnnotation).filter(MessageAnnotation.app_id == app_id).count()
     if not app:
-        raise NotFound("App not found")
+        logging.info(click.style("App not found: {}".format(app_id), fg="red"))
+        db.session.close()
+        return
 
     app_annotation_setting = (
         db.session.query(AppAnnotationSetting).filter(AppAnnotationSetting.app_id == app_id).first()
     )
 
     if not app_annotation_setting:
-        raise NotFound("App annotation setting not found")
+        logging.info(click.style("App annotation setting not found: {}".format(app_id), fg="red"))
+        db.session.close()
+        return
 
     disable_app_annotation_key = "disable_app_annotation_{}".format(str(app_id))
     disable_app_annotation_job_key = "disable_app_annotation_job_{}".format(str(job_id))
@@ -66,3 +69,4 @@ def disable_annotation_reply_task(job_id: str, app_id: str, tenant_id: str):
         redis_client.setex(disable_app_annotation_error_key, 600, str(e))
     finally:
         redis_client.delete(disable_app_annotation_key)
+        db.session.close()

api/tasks/annotation/enable_annotation_reply_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.datasource.vdb.vector_factory import Vector
 from core.rag.models.document import Document
@@ -34,7 +33,9 @@ def enable_annotation_reply_task(
     app = db.session.query(App).filter(App.id == app_id, App.tenant_id == tenant_id, App.status == "normal").first()
 
     if not app:
-        raise NotFound("App not found")
+        logging.info(click.style("App not found: {}".format(app_id), fg="red"))
+        db.session.close()
+        return
 
     annotations = db.session.query(MessageAnnotation).filter(MessageAnnotation.app_id == app_id).all()
     enable_app_annotation_key = "enable_app_annotation_{}".format(str(app_id))
@@ -121,3 +122,4 @@ def enable_annotation_reply_task(
         db.session.rollback()
     finally:
         redis_client.delete(enable_app_annotation_key)
+        db.session.close()

api/tasks/annotation/update_annotation_to_index_task.py

@@ -6,6 +6,7 @@ from celery import shared_task  # type: ignore
 
 from core.rag.datasource.vdb.vector_factory import Vector
 from core.rag.models.document import Document
+from extensions.ext_database import db
 from models.dataset import Dataset
 from services.dataset_service import DatasetCollectionBindingService
 
@@ -56,3 +57,5 @@ def update_annotation_to_index_task(
         )
     except Exception:
         logging.exception("Build index for annotation failed")
+    finally:
+        db.session.close()

api/tasks/batch_clean_document_task.py

@@ -74,3 +74,5 @@ def batch_clean_document_task(document_ids: list[str], dataset_id: str, doc_form
         )
     except Exception:
         logging.exception("Cleaned documents when documents deleted failed")
+    finally:
+        db.session.close()

api/tasks/batch_create_segment_to_index_task.py

@@ -127,3 +127,5 @@ def batch_create_segment_to_index_task(
     except Exception:
         logging.exception("Segments batch created index failed")
         redis_client.setex(indexing_cache_key, 600, "error")
+    finally:
+        db.session.close()

api/tasks/clean_dataset_task.py

@@ -121,3 +121,5 @@ def clean_dataset_task(
         )
     except Exception:
         logging.exception("Cleaned dataset when dataset deleted failed")
+    finally:
+        db.session.close()

api/tasks/clean_document_task.py

@@ -82,3 +82,5 @@ def clean_document_task(document_id: str, dataset_id: str, doc_form: str, file_i
         )
     except Exception:
         logging.exception("Cleaned document when document deleted failed")
+    finally:
+        db.session.close()

api/tasks/clean_notion_document_task.py

@@ -53,3 +53,5 @@ def clean_notion_document_task(document_ids: list[str], dataset_id: str):
         )
     except Exception:
         logging.exception("Cleaned document when import form notion document deleted  failed")
+    finally:
+        db.session.close()

api/tasks/create_segment_to_index_task.py

@@ -5,7 +5,6 @@ from typing import Optional
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
 from core.rag.models.document import Document
@@ -27,7 +26,9 @@ def create_segment_to_index_task(segment_id: str, keywords: Optional[list[str]]
 
     segment = db.session.query(DocumentSegment).filter(DocumentSegment.id == segment_id).first()
     if not segment:
-        raise NotFound("Segment not found")
+        logging.info(click.style("Segment not found: {}".format(segment_id), fg="red"))
+        db.session.close()
+        return
 
     if segment.status != "waiting":
         return
@@ -93,3 +94,4 @@ def create_segment_to_index_task(segment_id: str, keywords: Optional[list[str]]
         db.session.commit()
     finally:
         redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/deal_dataset_vector_index_task.py

@@ -167,3 +167,5 @@ def deal_dataset_vector_index_task(dataset_id: str, action: str):
         )
     except Exception:
         logging.exception("Deal dataset vector index failed")
+    finally:
+        db.session.close()

api/tasks/delete_segment_from_index_task.py

@@ -41,3 +41,5 @@ def delete_segment_from_index_task(index_node_ids: list, dataset_id: str, docume
         logging.info(click.style("Segment deleted from index latency: {}".format(end_at - start_at), fg="green"))
     except Exception:
         logging.exception("delete segment from index failed")
+    finally:
+        db.session.close()

api/tasks/disable_segment_from_index_task.py

@@ -3,7 +3,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
 from extensions.ext_database import db
@@ -24,10 +23,14 @@ def disable_segment_from_index_task(segment_id: str):
 
     segment = db.session.query(DocumentSegment).filter(DocumentSegment.id == segment_id).first()
     if not segment:
-        raise NotFound("Segment not found")
+        logging.info(click.style("Segment not found: {}".format(segment_id), fg="red"))
+        db.session.close()
+        return
 
     if segment.status != "completed":
-        raise NotFound("Segment is not completed , disable action is not allowed.")
+        logging.info(click.style("Segment is not completed, disable is not allowed: {}".format(segment_id), fg="red"))
+        db.session.close()
+        return
 
     indexing_cache_key = "segment_{}_indexing".format(segment.id)
 
@@ -62,3 +65,4 @@ def disable_segment_from_index_task(segment_id: str):
         db.session.commit()
     finally:
         redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/disable_segments_from_index_task.py

@@ -26,15 +26,18 @@ def disable_segments_from_index_task(segment_ids: list, dataset_id: str, documen
     dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
     if not dataset:
         logging.info(click.style("Dataset {} not found, pass.".format(dataset_id), fg="cyan"))
+        db.session.close()
         return
 
     dataset_document = db.session.query(DatasetDocument).filter(DatasetDocument.id == document_id).first()
 
     if not dataset_document:
         logging.info(click.style("Document {} not found, pass.".format(document_id), fg="cyan"))
+        db.session.close()
         return
     if not dataset_document.enabled or dataset_document.archived or dataset_document.indexing_status != "completed":
         logging.info(click.style("Document {} status is invalid, pass.".format(document_id), fg="cyan"))
+        db.session.close()
         return
     # sync index processor
     index_processor = IndexProcessorFactory(dataset_document.doc_form).init_index_processor()
@@ -50,6 +53,7 @@ def disable_segments_from_index_task(segment_ids: list, dataset_id: str, documen
     )
 
     if not segments:
+        db.session.close()
         return
 
     try:
@@ -76,3 +80,4 @@ def disable_segments_from_index_task(segment_ids: list, dataset_id: str, documen
         for segment in segments:
             indexing_cache_key = "segment_{}_indexing".format(segment.id)
             redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/document_indexing_sync_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.indexing_runner import DocumentIsPausedError, IndexingRunner
 from core.rag.extractor.notion_extractor import NotionExtractor
@@ -29,7 +28,9 @@ def document_indexing_sync_task(dataset_id: str, document_id: str):
     document = db.session.query(Document).filter(Document.id == document_id, Document.dataset_id == dataset_id).first()
 
     if not document:
-        raise NotFound("Document not found")
+        logging.info(click.style("Document not found: {}".format(document_id), fg="red"))
+        db.session.close()
+        return
 
     data_source_info = document.data_source_info_dict
     if document.data_source_type == "notion_import":

api/tasks/document_indexing_task.py

@@ -27,6 +27,7 @@ def document_indexing_task(dataset_id: str, document_ids: list):
     dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
     if not dataset:
         logging.info(click.style("Dataset is not found: {}".format(dataset_id), fg="yellow"))
+        db.session.close()
         return
     # check document limit
     features = FeatureService.get_features(dataset.tenant_id)
@@ -55,6 +56,7 @@ def document_indexing_task(dataset_id: str, document_ids: list):
                 document.stopped_at = datetime.datetime.now(datetime.UTC).replace(tzinfo=None)
                 db.session.add(document)
         db.session.commit()
+        db.session.close()
         return
 
     for document_id in document_ids:
@@ -80,3 +82,5 @@ def document_indexing_task(dataset_id: str, document_ids: list):
         logging.info(click.style(str(ex), fg="yellow"))
     except Exception:
         pass
+    finally:
+        db.session.close()

api/tasks/document_indexing_update_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.indexing_runner import DocumentIsPausedError, IndexingRunner
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
@@ -27,7 +26,9 @@ def document_indexing_update_task(dataset_id: str, document_id: str):
     document = db.session.query(Document).filter(Document.id == document_id, Document.dataset_id == dataset_id).first()
 
     if not document:
-        raise NotFound("Document not found")
+        logging.info(click.style("Document not found: {}".format(document_id), fg="red"))
+        db.session.close()
+        return
 
     document.indexing_status = "parsing"
     document.processing_started_at = datetime.datetime.now(datetime.UTC).replace(tzinfo=None)
@@ -73,3 +74,5 @@ def document_indexing_update_task(dataset_id: str, document_id: str):
         logging.info(click.style(str(ex), fg="yellow"))
     except Exception:
         pass
+    finally:
+        db.session.close()

api/tasks/duplicate_document_indexing_task.py

@@ -27,7 +27,9 @@ def duplicate_document_indexing_task(dataset_id: str, document_ids: list):
 
     dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
     if dataset is None:
-        raise ValueError("Dataset not found")
+        logging.info(click.style("Dataset not found: {}".format(dataset_id), fg="red"))
+        db.session.close()
+        return
 
     # check document limit
     features = FeatureService.get_features(dataset.tenant_id)
@@ -57,6 +59,8 @@ def duplicate_document_indexing_task(dataset_id: str, document_ids: list):
                 db.session.add(document)
         db.session.commit()
         return
+    finally:
+        db.session.close()
 
     for document_id in document_ids:
         logging.info(click.style("Start process document: {}".format(document_id), fg="green"))
@@ -96,3 +100,5 @@ def duplicate_document_indexing_task(dataset_id: str, document_ids: list):
         logging.info(click.style(str(ex), fg="yellow"))
     except Exception:
         pass
+    finally:
+        db.session.close()

api/tasks/enable_segment_to_index_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.index_processor.constant.index_type import IndexType
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
@@ -27,10 +26,14 @@ def enable_segment_to_index_task(segment_id: str):
 
     segment = db.session.query(DocumentSegment).filter(DocumentSegment.id == segment_id).first()
     if not segment:
-        raise NotFound("Segment not found")
+        logging.info(click.style("Segment not found: {}".format(segment_id), fg="red"))
+        db.session.close()
+        return
 
     if segment.status != "completed":
-        raise NotFound("Segment is not completed, enable action is not allowed.")
+        logging.info(click.style("Segment is not completed, enable is not allowed: {}".format(segment_id), fg="red"))
+        db.session.close()
+        return
 
     indexing_cache_key = "segment_{}_indexing".format(segment.id)
 
@@ -94,3 +97,4 @@ def enable_segment_to_index_task(segment_id: str):
         db.session.commit()
     finally:
         redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/enable_segments_to_index_task.py

@@ -34,9 +34,11 @@ def enable_segments_to_index_task(segment_ids: list, dataset_id: str, document_i
 
     if not dataset_document:
         logging.info(click.style("Document {} not found, pass.".format(document_id), fg="cyan"))
+        db.session.close()
         return
     if not dataset_document.enabled or dataset_document.archived or dataset_document.indexing_status != "completed":
         logging.info(click.style("Document {} status is invalid, pass.".format(document_id), fg="cyan"))
+        db.session.close()
         return
     # sync index processor
     index_processor = IndexProcessorFactory(dataset_document.doc_form).init_index_processor()
@@ -51,6 +53,8 @@ def enable_segments_to_index_task(segment_ids: list, dataset_id: str, document_i
         .all()
     )
     if not segments:
+        logging.info(click.style("Segments not found: {}".format(segment_ids), fg="cyan"))
+        db.session.close()
         return
 
     try:
@@ -108,3 +112,4 @@ def enable_segments_to_index_task(segment_ids: list, dataset_id: str, document_i
         for segment in segments:
             indexing_cache_key = "segment_{}_indexing".format(segment.id)
             redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/recover_document_indexing_task.py

@@ -3,7 +3,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.indexing_runner import DocumentIsPausedError, IndexingRunner
 from extensions.ext_database import db
@@ -25,7 +24,9 @@ def recover_document_indexing_task(dataset_id: str, document_id: str):
     document = db.session.query(Document).filter(Document.id == document_id, Document.dataset_id == dataset_id).first()
 
     if not document:
-        raise NotFound("Document not found")
+        logging.info(click.style("Document not found: {}".format(document_id), fg="red"))
+        db.session.close()
+        return
 
     try:
         indexing_runner = IndexingRunner()
@@ -43,3 +44,5 @@ def recover_document_indexing_task(dataset_id: str, document_id: str):
         logging.info(click.style(str(ex), fg="yellow"))
     except Exception:
         pass
+    finally:
+        db.session.close()

api/tasks/remove_document_from_index_task.py

@@ -4,7 +4,6 @@ import time
 
 import click
 from celery import shared_task  # type: ignore
-from werkzeug.exceptions import NotFound
 
 from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
 from extensions.ext_database import db
@@ -25,9 +24,13 @@ def remove_document_from_index_task(document_id: str):
 
     document = db.session.query(Document).filter(Document.id == document_id).first()
     if not document:
-        raise NotFound("Document not found")
+        logging.info(click.style("Document not found: {}".format(document_id), fg="red"))
+        db.session.close()
+        return
 
     if document.indexing_status != "completed":
+        logging.info(click.style("Document is not completed, remove is not allowed: {}".format(document_id), fg="red"))
+        db.session.close()
         return
 
     indexing_cache_key = "document_{}_indexing".format(document.id)
@@ -71,3 +74,4 @@ def remove_document_from_index_task(document_id: str):
             db.session.commit()
     finally:
         redis_client.delete(indexing_cache_key)
+        db.session.close()

api/tasks/retry_document_indexing_task.py

@@ -27,7 +27,9 @@ def retry_document_indexing_task(dataset_id: str, document_ids: list[str]):
 
     dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
     if not dataset:
-        raise ValueError("Dataset not found")
+        logging.info(click.style("Dataset not found: {}".format(dataset_id), fg="red"))
+        db.session.close()
+        return
 
     for document_id in document_ids:
         retry_indexing_cache_key = "document_{}_is_retried".format(document_id)
@@ -52,6 +54,7 @@ def retry_document_indexing_task(dataset_id: str, document_ids: list[str]):
                 db.session.add(document)
                 db.session.commit()
             redis_client.delete(retry_indexing_cache_key)
+            db.session.close()
             return
 
         logging.info(click.style("Start retry document: {}".format(document_id), fg="green"))
@@ -60,6 +63,7 @@ def retry_document_indexing_task(dataset_id: str, document_ids: list[str]):
         )
         if not document:
             logging.info(click.style("Document not found: {}".format(document_id), fg="yellow"))
+            db.session.close()
             return
         try:
             # clean old data
@@ -92,5 +96,7 @@ def retry_document_indexing_task(dataset_id: str, document_ids: list[str]):
             logging.info(click.style(str(ex), fg="yellow"))
             redis_client.delete(retry_indexing_cache_key)
             pass
+        finally:
+            db.session.close()
     end_at = time.perf_counter()
     logging.info(click.style("Retry dataset: {} latency: {}".format(dataset_id, end_at - start_at), fg="green"))