Feat/delete file when clean document (#5882)

This commit is contained in:
Jyong
2024-07-15 19:57:05 +08:00
committed by GitHub
parent b47fa27a35
commit d320d1468d
4 changed files with 52 additions and 7 deletions

View File

@@ -1,21 +1,25 @@
import logging
import time
from typing import Optional
import click
from celery import shared_task
from core.rag.index_processor.index_processor_factory import IndexProcessorFactory
from extensions.ext_database import db
from extensions.ext_storage import storage
from models.dataset import Dataset, DocumentSegment
from models.model import UploadFile
@shared_task(queue='dataset')
def clean_document_task(document_id: str, dataset_id: str, doc_form: str):
def clean_document_task(document_id: str, dataset_id: str, doc_form: str, file_id: Optional[str]):
"""
Clean document when document deleted.
:param document_id: document id
:param dataset_id: dataset id
:param doc_form: doc_form
:param file_id: file id
Usage: clean_document_task.delay(document_id, dataset_id)
"""
@@ -39,8 +43,20 @@ def clean_document_task(document_id: str, dataset_id: str, doc_form: str):
db.session.delete(segment)
db.session.commit()
end_at = time.perf_counter()
logging.info(
click.style('Cleaned document when document deleted: {} latency: {}'.format(document_id, end_at - start_at), fg='green'))
if file_id:
file = db.session.query(UploadFile).filter(
UploadFile.id == file_id
).first()
if file:
try:
storage.delete(file.key)
except Exception:
logging.exception("Delete file failed when document deleted, file_id: {}".format(file_id))
db.session.delete(file)
db.session.commit()
end_at = time.perf_counter()
logging.info(
click.style('Cleaned document when document deleted: {} latency: {}'.format(document_id, end_at - start_at), fg='green'))
except Exception:
logging.exception("Cleaned document when document deleted failed")