Use the new post version upload signal to ensure OCR starts only when all of a new

document version pages have been created

mayan/apps/documents/models.py

@@ -37,6 +37,7 @@ from .managers import (DocumentManager, DocumentPageTransformationManager,
 from .runtime import storage_backend
 from .settings import (CACHE_PATH, CHECKSUM_FUNCTION, DISPLAY_SIZE, LANGUAGE,
                        UUID_FUNCTION, ZOOM_MAX_LEVEL, ZOOM_MIN_LEVEL)
+from .signals import post_version_upload
 
 HASH_FUNCTION = lambda x: hashlib.sha256(x).hexdigest()  # document image cache name hash function
 logger = logging.getLogger(__name__)
@@ -365,7 +366,7 @@ class DocumentVersion(models.Model):
         Overloaded save method that updates the document version's checksum,
         mimetype, page count and transformation when created
         """
-        new_document = not self.pk
+        new_document_version = not self.pk
 
         # Only do this for new documents
         transformations = kwargs.pop('transformations', None)
@@ -374,7 +375,7 @@ class DocumentVersion(models.Model):
         for key in sorted(DocumentVersion._post_save_hooks):
             DocumentVersion._post_save_hooks[key](self)
 
-        if new_document:
+        if new_document_version:
             # Only do this for new documents
             self.update_checksum(save=False)
             self.update_mimetype(save=False)
@@ -384,6 +385,8 @@ class DocumentVersion(models.Model):
             if transformations:
                 self.apply_default_transformations(transformations)
 
+            post_version_upload.send(sender=self.__class__, instance=self)
+
     def update_checksum(self, save=True):
         """
         Open a document version's file and update the checksum field using the

mayan/apps/ocr/__init__.py

@@ -2,7 +2,6 @@ from __future__ import absolute_import
 
 import logging
 
-from django.db.models.signals import post_save
 from django.dispatch import receiver
 from django.utils.translation import ugettext_lazy as _
 
@@ -10,6 +9,7 @@ from south.signals import post_migrate
 
 from acls.api import class_permissions
 from documents.models import Document, DocumentVersion
+from documents.signals import post_version_upload
 from main.api import register_maintenance_links
 from navigation.api import register_links, register_multi_item_links
 from project_tools.api import register_tool
@@ -36,13 +36,12 @@ def document_ocr_submit(self):
     task_do_ocr.apply_async(args=[self.pk], queue='ocr')
 
 
-@receiver(post_save, dispatch_uid='document_post_save', sender=DocumentVersion)
-def document_post_save(sender, instance, **kwargs):
-    logger.debug('received post save signal')
+@receiver(post_version_upload, dispatch_uid='post_version_upload_ocr', sender=DocumentVersion)
+def post_version_upload_ocr(sender, instance, **kwargs):
+    logger.debug('received post_version_upload')
     logger.debug('instance: %s' % instance)
-    if kwargs.get('created', False):
-        if instance.document.document_type.ocr:
-            instance.document.submit_for_ocr()
+    if instance.document.document_type.ocr:
+        instance.document.submit_for_ocr()
 
 
 @receiver(post_migrate, dispatch_uid='create_default_queue')