chore: 纳入本地 Dify 1.19.0 源码及品牌与免登改动
演示目录只保留可重建的源码。运行密钥、数据卷、依赖和带口令的本地配置不进入仓库。 Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
1 parent
60158935f4
commit
6519682885
15831 files changed
+3063428
No files matched your search
@@ -0,0 +1,105 @@
|
||||
import logging
|
||||
import time
|
||||
|
||||
import click
|
||||
from celery import shared_task
|
||||
from sqlalchemy import select, update
|
||||
|
||||
from core.db.session_factory import session_factory
|
||||
from core.rag.index_processor.index_processor import IndexProcessorFactory
|
||||
from extensions.ext_redis import redis_client
|
||||
from models.dataset import Dataset, DocumentSegment, SegmentAttachmentBinding
|
||||
from models.dataset import Document as DatasetDocument
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
|
||||
@shared_task(queue="dataset")
|
||||
def disable_segments_from_index_task(segment_ids: list, dataset_id: str, document_id: str):
|
||||
"""
|
||||
Async disable segments from index
|
||||
:param segment_ids: list of segment ids
|
||||
:param dataset_id: dataset id
|
||||
:param document_id: document id
|
||||
|
||||
Usage: disable_segments_from_index_task.delay(segment_ids, dataset_id, document_id)
|
||||
"""
|
||||
start_at = time.perf_counter()
|
||||
|
||||
with session_factory.create_session() as session:
|
||||
dataset = session.scalar(select(Dataset).where(Dataset.id == dataset_id).limit(1))
|
||||
if not dataset:
|
||||
logger.info(click.style(f"Dataset {dataset_id} not found, pass.", fg="cyan"))
|
||||
return
|
||||
|
||||
dataset_document = session.scalar(select(DatasetDocument).where(DatasetDocument.id == document_id).limit(1))
|
||||
|
||||
if not dataset_document:
|
||||
logger.info(click.style(f"Document {document_id} not found, pass.", fg="cyan"))
|
||||
return
|
||||
if not dataset_document.enabled or dataset_document.archived or dataset_document.indexing_status != "completed":
|
||||
logger.info(click.style(f"Document {document_id} status is invalid, pass.", fg="cyan"))
|
||||
return
|
||||
# sync index processor
|
||||
index_processor = IndexProcessorFactory(dataset_document.doc_form).init_index_processor()
|
||||
|
||||
segments = session.scalars(
|
||||
select(DocumentSegment).where(
|
||||
DocumentSegment.id.in_(segment_ids),
|
||||
DocumentSegment.dataset_id == dataset_id,
|
||||
DocumentSegment.document_id == document_id,
|
||||
)
|
||||
).all()
|
||||
|
||||
if not segments:
|
||||
return
|
||||
|
||||
try:
|
||||
index_node_ids = [segment.index_node_id for segment in segments if segment.index_node_id]
|
||||
if dataset.is_multimodal:
|
||||
segment_ids = [segment.id for segment in segments]
|
||||
segment_attachment_bindings = session.scalars(
|
||||
select(SegmentAttachmentBinding).where(SegmentAttachmentBinding.segment_id.in_(segment_ids))
|
||||
).all()
|
||||
if segment_attachment_bindings:
|
||||
attachment_ids = [binding.attachment_id for binding in segment_attachment_bindings]
|
||||
index_node_ids.extend(attachment_ids)
|
||||
index_processor.clean(
|
||||
dataset, index_node_ids, with_keywords=True, delete_child_chunks=False, session=session
|
||||
)
|
||||
session.commit()
|
||||
|
||||
# Disable summary indexes for these segments
|
||||
from services.knowledge.summaries.adapters import SummaryIndexAdapter
|
||||
|
||||
segment_ids_list = [segment.id for segment in segments]
|
||||
try:
|
||||
# Get disabled_by from first segment (they should all have the same disabled_by)
|
||||
disabled_by = segments[0].disabled_by if segments else None
|
||||
SummaryIndexAdapter.disable_summaries_for_segments(
|
||||
dataset=dataset,
|
||||
segment_ids=segment_ids_list,
|
||||
disabled_by=disabled_by,
|
||||
)
|
||||
except Exception:
|
||||
logger.warning("Failed to disable summaries for segments", exc_info=True)
|
||||
|
||||
end_at = time.perf_counter()
|
||||
logger.info(click.style(f"Segments removed from index latency: {end_at - start_at}", fg="green"))
|
||||
except Exception:
|
||||
# update segment error msg
|
||||
session.rollback()
|
||||
session.execute(
|
||||
update(DocumentSegment)
|
||||
.where(
|
||||
DocumentSegment.id.in_(segment_ids),
|
||||
DocumentSegment.dataset_id == dataset_id,
|
||||
DocumentSegment.document_id == document_id,
|
||||
)
|
||||
.values(disabled_at=None, disabled_by=None, enabled=True)
|
||||
)
|
||||
session.commit()
|
||||
finally:
|
||||
for segment in segments:
|
||||
indexing_cache_key = f"segment_{segment.id}_indexing"
|
||||
redis_client.delete(indexing_cache_key)
|
||||
Reference in new issue
Block a user