Added distributed OCR queue support

apps/ocr/__init__.py

@@ -25,11 +25,10 @@ submit_document = {'text':_('submit to OCR queue'), 'view':'submit_document', 'a
 register_links(Document, [submit_document], menu_name='sidebar')
 
 #Menus
-#register_menu([
-#    {'text':_('OCR'), 'view':'ocr_queue', 'links':[
-#        ocr_queue
-#    ],'famfam':'hourglass','position':5}])
-
+register_menu([
+    {'text':_('OCR'), 'view':'queue_document_list', 'links':[
+        #ocr_queue
+    ],'famfam':'hourglass','position':4}])
 
 try:
     default_queue, created = DocumentQueue.objects.get_or_create(name='default')

apps/ocr/api.py

@@ -15,11 +15,6 @@ from converter.api import convert_document_for_ocr
 
 from ocr.conf.settings import TESSERACT_PATH
 
-#from literals import QUEUEDOCUMENT_STATE_PROCESSING, \
-#    QUEUEDOCUMENT_STATE_ERROR, QUEUEDOCUMENT_STATE_PENDING
-
-#from models import DocumentQueue
-
 
 def cleanup(filename):
     ''' tries to remove the given filename. Ignores non-existent files '''

apps/ocr/models.py

@@ -6,6 +6,13 @@ from documents.models import Document
 from literals import DOCUMENTQUEUE_STATE_STOPPED,\
     DOCUMENTQUEUE_STATE_CHOICES, QUEUEDOCUMENT_STATE_PENDING,\
     QUEUEDOCUMENT_STATE_CHOICES
+
+
+def add_document_to_queue(document, queue_name='default'):
+    document_queue = DocumentQueue.objects.get(name=queue_name)
+    queue_document = QueueDocument(document_queue=document_queue, document=document)
+    queue_document.save()
+    return document_queue
     
 
 class DocumentQueue(models.Model):
@@ -23,10 +30,9 @@ class DocumentQueue(models.Model):
     def __unicode__(self):
         return self.label
 
-#    def add_document(self, document):
-#        queue_document = QueueDocument(document_queue=self, document=document)
-#        queue_document.save()
-#        queue_dict[self.name].put(queue_document)
+    def add_document(self, document):
+        queue_document = QueueDocument(document_queue=self, document=document)
+        queue_document.save()
         
 
 class QueueDocument(models.Model):
@@ -38,7 +44,6 @@ class QueueDocument(models.Model):
         default=QUEUEDOCUMENT_STATE_PENDING,
         verbose_name=_(u'state'))
     result = models.TextField(blank=True, null=True, verbose_name=_(u'result'))
-    pid = models.PositiveIntegerField(blank=True, null=True, verbose_name=_(u'process id'))
     
     class Meta:
         verbose_name = _(u'queue document')

apps/ocr/urls.py

@@ -3,4 +3,5 @@ from django.utils.translation import ugettext_lazy as _
 
 urlpatterns = patterns('ocr.views',
     url(r'^(?P<document_id>\d+)/submit/$', 'submit_document', (), 'submit_document'),
+    url(r'^ocr/queue/document/list/$', 'queue_document_list', (), 'queue_document_list'),
 )

apps/ocr/views.py

@@ -13,10 +13,36 @@ from documents.models import Document
 
 from ocr import PERMISSION_OCR_DOCUMENT
 
-from models import DocumentQueue, QueueDocument
+from models import DocumentQueue, QueueDocument, add_document_to_queue
 
 from tasks import do_document_ocr_task
 
+
+def queue_document_list(request, queue_name='default'):
+    permissions = [PERMISSION_OCR_DOCUMENT]
+    try:
+        check_permissions(request.user, 'ocr', permissions)
+    except Unauthorized, e:
+        raise Http404(e)
+        
+    document_queue = get_object_or_404(DocumentQueue, name=queue_name)
+            
+    return object_list(
+        request,
+        queryset=document_queue.queuedocument_set.all(),
+        template_name='generic_list.html',
+        extra_context={
+            'title':_(u'queued documents'),
+            'hide_object':True,
+            'extra_columns':[
+                {'name':'document', 'attribute': 'document'},
+                {'name':'submitted', 'attribute': lambda x: unicode(x.datetime_submitted).split('.')[0]},
+                {'name':'state', 'attribute': lambda x: x.get_state_display()},
+            ],
+        },
+    )    
+        
+
 def submit_document(request, document_id, queue_name='default'):
     permissions = [PERMISSION_OCR_DOCUMENT]
     try:
@@ -27,13 +53,7 @@ def submit_document(request, document_id, queue_name='default'):
     document = get_object_or_404(Document, pk=document_id)
     
     document_queue = get_object_or_404(DocumentQueue, name=queue_name)
-    do_document_ocr_task.delay(document.id)
-    ##document_queue.add_document(document)
-    #queue_document = QueueDocument(document_queue=document_queue, document=document)
-    #queue_document.save()
-
-
-    #add.delay(1,2)
+    add_document_to_queue(document, document_queue.name)
 
     messages.success(request, _(u'Document: %s was added to the OCR queue: %s.') % (document, document_queue.label))
     return HttpResponseRedirect(request.META['HTTP_REFERER'])