2023-05-15 08:51:32 +08:00
|
|
|
import datetime
|
|
|
|
import logging
|
|
|
|
import time
|
|
|
|
|
|
|
|
import click
|
|
|
|
from celery import shared_task
|
2024-02-22 17:16:22 +08:00
|
|
|
from flask import current_app
|
2024-02-06 13:21:13 +08:00
|
|
|
|
2024-01-12 12:34:01 +08:00
|
|
|
from core.indexing_runner import DocumentIsPausedException, IndexingRunner
|
2023-05-15 08:51:32 +08:00
|
|
|
from extensions.ext_database import db
|
2024-02-22 17:16:22 +08:00
|
|
|
from models.dataset import Dataset, Document
|
|
|
|
from services.feature_service import FeatureService
|
2023-05-15 08:51:32 +08:00
|
|
|
|
|
|
|
|
2023-07-31 13:13:08 +08:00
|
|
|
@shared_task(queue='dataset')
|
2023-06-16 21:47:51 +08:00
|
|
|
def document_indexing_task(dataset_id: str, document_ids: list):
|
2023-05-15 08:51:32 +08:00
|
|
|
"""
|
|
|
|
Async process document
|
|
|
|
:param dataset_id:
|
2023-06-16 21:47:51 +08:00
|
|
|
:param document_ids:
|
2023-05-15 08:51:32 +08:00
|
|
|
|
|
|
|
Usage: document_indexing_task.delay(dataset_id, document_id)
|
|
|
|
"""
|
2023-06-16 21:47:51 +08:00
|
|
|
documents = []
|
2023-06-25 16:49:14 +08:00
|
|
|
start_at = time.perf_counter()
|
2024-02-22 17:16:22 +08:00
|
|
|
|
|
|
|
dataset = db.session.query(Dataset).filter(Dataset.id == dataset_id).first()
|
|
|
|
|
|
|
|
# check document limit
|
|
|
|
features = FeatureService.get_features(dataset.tenant_id)
|
|
|
|
try:
|
|
|
|
if features.billing.enabled:
|
|
|
|
vector_space = features.vector_space
|
|
|
|
count = len(document_ids)
|
|
|
|
batch_upload_limit = int(current_app.config['BATCH_UPLOAD_LIMIT'])
|
|
|
|
if count > batch_upload_limit:
|
|
|
|
raise ValueError(f"You have reached the batch upload limit of {batch_upload_limit}.")
|
|
|
|
if 0 < vector_space.limit <= vector_space.size:
|
|
|
|
raise ValueError("Your total number of documents plus the number of uploads have over the limit of "
|
|
|
|
"your subscription.")
|
|
|
|
except Exception as e:
|
|
|
|
for document_id in document_ids:
|
|
|
|
document = db.session.query(Document).filter(
|
|
|
|
Document.id == document_id,
|
|
|
|
Document.dataset_id == dataset_id
|
|
|
|
).first()
|
|
|
|
if document:
|
|
|
|
document.indexing_status = 'error'
|
|
|
|
document.error = str(e)
|
2024-04-12 16:22:24 +08:00
|
|
|
document.stopped_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
|
2024-02-22 17:16:22 +08:00
|
|
|
db.session.add(document)
|
|
|
|
db.session.commit()
|
|
|
|
return
|
|
|
|
|
2023-06-16 21:47:51 +08:00
|
|
|
for document_id in document_ids:
|
|
|
|
logging.info(click.style('Start process document: {}'.format(document_id), fg='green'))
|
|
|
|
|
|
|
|
document = db.session.query(Document).filter(
|
|
|
|
Document.id == document_id,
|
|
|
|
Document.dataset_id == dataset_id
|
|
|
|
).first()
|
|
|
|
|
2023-10-12 13:30:44 +08:00
|
|
|
if document:
|
|
|
|
document.indexing_status = 'parsing'
|
2024-04-12 16:22:24 +08:00
|
|
|
document.processing_started_at = datetime.datetime.now(datetime.timezone.utc).replace(tzinfo=None)
|
2023-10-12 13:30:44 +08:00
|
|
|
documents.append(document)
|
|
|
|
db.session.add(document)
|
2023-05-15 08:51:32 +08:00
|
|
|
db.session.commit()
|
|
|
|
|
|
|
|
try:
|
|
|
|
indexing_runner = IndexingRunner()
|
2023-06-16 21:47:51 +08:00
|
|
|
indexing_runner.run(documents)
|
2023-05-15 08:51:32 +08:00
|
|
|
end_at = time.perf_counter()
|
2023-06-25 16:49:14 +08:00
|
|
|
logging.info(click.style('Processed dataset: {} latency: {}'.format(dataset_id, end_at - start_at), fg='green'))
|
|
|
|
except DocumentIsPausedException as ex:
|
|
|
|
logging.info(click.style(str(ex), fg='yellow'))
|
|
|
|
except Exception:
|
|
|
|
pass
|