Underground-Digital/Workflow-Engine
0
1import logging2import time3 4import click5from celery import shared_task6 7from core.rag.index_processor.index_processor_factory import IndexProcessorFactory8from extensions.ext_database import db9from models.dataset import Dataset, Document, DocumentSegment10 11 12@shared_task(queue="dataset")13def clean_notion_document_task(document_ids: list[str], dataset_id: str):14 """15 Clean document when document deleted.16 :param document_ids: document ids17 :param dataset_id: dataset id18 19 Usage: clean_notion_document_task.delay(document_ids, dataset_id)20 """21 logging.info(22 click.style("Start clean document when import form notion document deleted: {}".format(dataset_id), fg="green")23 )24 start_at = time.perf_counter()25 26 try:27 dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()28 29 if not dataset:30 raise Exception("Document has no dataset")31 index_type = dataset.doc_form32 index_processor = IndexProcessorFactory(index_type).init_index_processor()33 for document_id in document_ids:34 document = db.session.query(Document).filter(Document.id == document_id).first()35 db.session.delete(document)36 37 segments = db.session.query(DocumentSegment).filter(DocumentSegment.document_id == document_id).all()38 index_node_ids = [segment.index_node_id for segment in segments]39 40 index_processor.clean(dataset, index_node_ids)41 42 for segment in segments:43 db.session.delete(segment)44 db.session.commit()45 end_at = time.perf_counter()46 logging.info(47 click.style(48 "Clean document when import form notion document deleted end :: {} latency: {}".format(49 dataset_id, end_at - start_at50 ),51 fg="green",52 )53 )54 except Exception:55 logging.exception("Cleaned document when import form notion document deleted failed")56 