apunkt/SurfSense
1
0
Fork 0
mirror of https://github.com/MODSetter/SurfSense.git synced 2026-06-06 20:15:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix bugs in indexing pipeline exception handling

Browse source
This commit is contained in:
CREDO23 2026-02-25 16:27:12 +02:00
parent b6c25628c8
commit 66d7d3da8a
3 changed files with 25 additions and 16 deletions
Download patch file Download diff file Expand all files Collapse all files

22
surfsense_backend/app/indexing_pipeline/document_persistence.py
View file

@ -10,12 +10,22 @@ from app.db import Document, DocumentStatus
async def rollback_and_persist_failure(
session: AsyncSession, document: Document, message: str
) -> None:
"""Roll back the current transaction, refresh the document, and persist a failed status."""
await session.rollback()
await session.refresh(document)
document.updated_at = datetime.now(UTC)
document.status = DocumentStatus.failed(message)
await session.commit()
"""Roll back the current transaction and best-effort persist a failed status.
Called exclusively from except blocks — must never raise, or the new exception
would chain with the original and mask it entirely.
"""
try:
await session.rollback()
except Exception:
return # Session is completely dead; nothing further we can do.
try:
await session.refresh(document)
document.updated_at = datetime.now(UTC)
document.status = DocumentStatus.failed(message)
await session.commit()
except Exception:
pass # Best-effort; document will be retried on the next sync.
def attach_chunks_to_document(document: Document, chunks: list) -> None:

9
surfsense_backend/app/indexing_pipeline/indexing_pipeline_service.py
View file

@ -135,7 +135,6 @@ class IndexingPipelineService:
return []
except TRANSIENT_DB_ERRORS as e:
log_doc_skipped_db(ctx, e)
await self.session.rollback()
except Exception as e:
log_doc_skipped_unknown(ctx, e)
continue
@ -207,14 +206,14 @@ class IndexingPipelineService:
log_permanent_llm_error(ctx, e)
await rollback_and_persist_failure(self.session, document, llm_permanent_message(e))
except EMBEDDING_ERRORS as e:
log_embedding_error(ctx, e)
await rollback_and_persist_failure(self.session, document, embedding_message(e))
except RecursionError as e:
log_chunking_overflow(ctx, e)
await rollback_and_persist_failure(self.session, document, PipelineMessages.CHUNKING_OVERFLOW)
except EMBEDDING_ERRORS as e:
log_embedding_error(ctx, e)
await rollback_and_persist_failure(self.session, document, embedding_message(e))
except FATAL_DB_ERRORS as e:
log_db_fatal_error(ctx, e)
raise

10
surfsense_backend/app/indexing_pipeline/pipeline_logger.py
View file

@ -88,7 +88,7 @@ def log_doc_skipped_unknown(ctx: PipelineLogContext, exc: Exception) -> None:
def log_batch_aborted(ctx: PipelineLogContext, exc: Exception) -> None:
_safe_log(logger.critical, LogMessages.BATCH_ABORTED, ctx, error=exc)
_safe_log(logger.critical, LogMessages.BATCH_ABORTED, ctx, exc_info=exc, error=exc)
def log_race_condition(ctx: PipelineLogContext) -> None:
@ -110,15 +110,15 @@ def log_retryable_llm_error(ctx: PipelineLogContext, exc: Exception) -> None:
def log_permanent_llm_error(ctx: PipelineLogContext, exc: Exception) -> None:
_safe_log(logger.error, LogMessages.LLM_PERMANENT, ctx, error=exc)
_safe_log(logger.error, LogMessages.LLM_PERMANENT, ctx, exc_info=exc, error=exc)
def log_embedding_error(ctx: PipelineLogContext, exc: Exception) -> None:
_safe_log(logger.error, LogMessages.EMBEDDING_FAILED, ctx, error=exc)
_safe_log(logger.error, LogMessages.EMBEDDING_FAILED, ctx, exc_info=exc, error=exc)
def log_chunking_overflow(ctx: PipelineLogContext, exc: Exception) -> None:
_safe_log(logger.error, LogMessages.CHUNKING_OVERFLOW, ctx, error=exc)
_safe_log(logger.error, LogMessages.CHUNKING_OVERFLOW, ctx, exc_info=exc, error=exc)
def log_db_transient_error(ctx: PipelineLogContext, exc: Exception) -> None:
@ -126,7 +126,7 @@ def log_db_transient_error(ctx: PipelineLogContext, exc: Exception) -> None:
def log_db_fatal_error(ctx: PipelineLogContext, exc: Exception) -> None:
_safe_log(logger.critical, LogMessages.DB_FATAL, ctx, error=exc)
_safe_log(logger.critical, LogMessages.DB_FATAL, ctx, exc_info=exc, error=exc)
def log_unexpected_error(ctx: PipelineLogContext, exc: Exception) -> None: