Files
mayan-edms/mayan/apps/ocr/tasks.py
Roberto Rosario 0699ad0556 Add support for new document page structure
Documents now have their own dedicated DocumentPage
submodel. The old DocumentPage is now called DocumentVersionPage.
This allows mappings between document pages and document version
pages, allowing renumbering, appending pages.
DocumentPages have a content_object to map them to any other
object. For now they only map to DocumentVersionPages.
New option added to the version upload form to append the
pages of the new version.
A new view was added to just append new pages with wraps the
new document version upload form and hides the append pages
checkbox set to True.
Add a new action, reset_pages to reset the pages of the
document to those of the latest version.

Missing: appending tests, checks for proper content_object in OCR and
document parsing.

Author: Roberto Rosario <roberto.rosario@mayan-edms.com>
Date:   Thu Oct 11 12:00:25 2019 -0400
2019-10-10 11:55:42 -04:00

55 lines
1.9 KiB
Python

from __future__ import unicode_literals
import logging
from django.apps import apps
from django.db import OperationalError
from mayan.apps.lock_manager.exceptions import LockError
from mayan.apps.lock_manager.runtime import locking_backend
from mayan.celery import app
from .literals import DO_OCR_RETRY_DELAY, LOCK_EXPIRE
logger = logging.getLogger(__name__)
@app.task(bind=True, default_retry_delay=DO_OCR_RETRY_DELAY, ignore_result=True)
def task_do_ocr(self, document_version_pk):
DocumentVersion = apps.get_model(
app_label='documents', model_name='DocumentVersion'
)
DocumentVersionPageOCRContent = apps.get_model(
app_label='ocr', model_name='DocumentVersionPageOCRContent'
)
lock_id = 'task_do_ocr_doc_version-%d' % document_version_pk
try:
logger.debug('trying to acquire lock: %s', lock_id)
# Acquire lock to avoid doing OCR on the same document version more
# than once concurrently
lock = locking_backend.acquire_lock(name=lock_id, timeout=LOCK_EXPIRE)
logger.debug('acquired lock: %s', lock_id)
document_version = None
try:
document_version = DocumentVersion.objects.get(
pk=document_version_pk
)
logger.info(
'Starting document OCR for document version: %s',
document_version
)
DocumentVersionPageOCRContent.objects.process_document_version(
document_version=document_version
)
except OperationalError as exception:
logger.warning(
'OCR error for document version: %d; %s. Retrying.',
document_version_pk, exception
)
raise self.retry(exc=exception)
finally:
lock.release()
except LockError:
logger.debug('unable to obtain lock: %s' % lock_id)