fix: merge main branch and renumber migration 027→037

Resolve 3 merge conflicts and renumber the automation_hooks migration
to follow main's migration chain (036_add_document_translation_fields).

Conflicts resolved:
- app/api/__init__.py: add automation_router alongside main's new routers
- app/utils/settings_service.py: add automation_hooks_enabled alongside compliance_enabled
- tests/conftest.py: add AutomationHook alongside AuditLog/ComplianceTemplate imports

Migration renumbered:
- 027_add_automation_hooks → 037_add_automation_hooks
- down_revision: 026_add_scheduled_jobs → 036_add_document_translation_fields

Co-authored-by: christianlouis <361235+christianlouis@users.noreply.github.com>
This commit is contained in:
copilot-swe-agent[bot]
2026-03-16 22:40:15 +00:00
parent 6a83d51d88
commit 204000aabc
318 changed files with 317516 additions and 2858 deletions
+69 -2
View File
@@ -1,10 +1,15 @@
# app/celery_app.py
import logging
import os
from celery import Celery
from celery.signals import task_failure, worker_ready
from app.config import settings
logger = logging.getLogger(__name__)
celery = Celery(
"document_processor",
broker=settings.redis_url,
@@ -21,6 +26,64 @@ celery.conf.task_routes = {
"app.tasks.*": {"queue": "document_processor"},
}
# Mapping of document pipeline task names to the positional index of ``file_id``
# in their ``args`` tuple. These indices correspond to the task signatures:
# process_with_ocr(filename, file_id, ...) → index 1
# extract_metadata_with_gpt(filename, text, file_id) → index 2
# embed_metadata_into_pdf(path, text, metadata, file_id) → index 3
# Tasks that always pass ``file_id`` as a keyword argument
# (e.g. ``process_document``, ``finalize_document_storage``) are not listed
# here — their ``file_id`` is found via ``kwargs`` instead.
_FILE_ID_ARG_INDEX: dict[str, int] = {
"app.tasks.process_with_ocr.process_with_ocr": 1,
"app.tasks.extract_metadata_with_gpt.extract_metadata_with_gpt": 2,
"app.tasks.embed_metadata_into_pdf.embed_metadata_into_pdf": 3,
}
def _dispatch_user_failure_notification(sender, exception, args: list | None, kwargs: dict | None) -> None:
"""Best-effort per-user failure notification for document pipeline tasks.
Extracts ``file_id`` from the failed task's arguments, looks up the owning
user from the database, and dispatches a ``document.failed`` notification.
"""
from app.database import SessionLocal
from app.models import FileRecord
from app.utils.user_notification import notify_user_document_failed
task_name = sender.name if sender else ""
if not task_name.startswith("app.tasks."):
return
# 1. Resolve file_id from kwargs or positional args
file_id = (kwargs or {}).get("file_id")
if file_id is None:
idx = _FILE_ID_ARG_INDEX.get(task_name)
if idx is not None and args and len(args) > idx:
val = args[idx]
if isinstance(val, int):
file_id = val
if file_id is None:
return
# 2. Look up owner from the database
with SessionLocal() as db:
record = db.query(FileRecord).filter(FileRecord.id == file_id).first()
if not record or not record.owner_id:
return
owner_id = record.owner_id
filename = record.original_filename or record.local_filename or "unknown"
# 3. Dispatch per-user notification
error_msg = f"{type(exception).__name__}: {exception}" if exception else "Unknown error"
notify_user_document_failed(
owner_id=owner_id,
filename=os.path.basename(filename),
error=error_msg,
file_id=file_id,
)
@worker_ready.connect
def init_sentry_on_worker_ready(**kwargs):
@@ -48,6 +111,10 @@ def task_failure_handler(
kwargs=kwargs or {},
)
except Exception as e:
import logging
logger.exception(f"Failed to send task failure notification: {e}")
logging.exception(f"Failed to send task failure notification: {e}")
# Also dispatch a per-user failure notification for document pipeline tasks
try:
_dispatch_user_failure_notification(sender, exception, args, kwargs)
except Exception:
logger.warning("Could not dispatch per-user failure notification", exc_info=True)