delete_segment_from_index_task.py 1.5 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243
  1. import logging
  2. import time
  3. import click
  4. from celery import shared_task # type: ignore
  5. from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
  6. from extensions.ext_database import db
  7. from models.dataset import Dataset, Document
  8. @shared_task(queue="dataset")
  9. def delete_segment_from_index_task(index_node_ids: list, dataset_id: str, document_id: str):
  10. """
  11. Async Remove segment from index
  12. :param index_node_ids:
  13. :param dataset_id:
  14. :param document_id:
  15. Usage: delete_segment_from_index_task.delay(segment_ids)
  16. """
  17. logging.info(click.style("Start delete segment from index", fg="green"))
  18. start_at = time.perf_counter()
  19. try:
  20. dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
  21. if not dataset:
  22. return
  23. dataset_document = db.session.query(Document).filter(Document.id == document_id).first()
  24. if not dataset_document:
  25. return
  26. if not dataset_document.enabled or dataset_document.archived or dataset_document.indexing_status != "completed":
  27. return
  28. index_type = dataset_document.doc_form
  29. index_processor = IndexProcessorFactory(index_type).init_index_processor()
  30. index_processor.clean(dataset, index_node_ids, with_keywords=True, delete_child_chunks=True)
  31. end_at = time.perf_counter()
  32. logging.info(click.style("Segment deleted from index latency: {}".format(end_at - start_at), fg="green"))
  33. except Exception:
  34. logging.exception("delete segment from index failed")