matthias/mayan-edms
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Initial changes for the new queue based OCR processing

Browse Source
This commit is contained in:
Roberto Rosario
2012-07-29 05:33:04 -04:00
parent d97b3f344d
commit d2e6df4dde
14 changed files with 280 additions and 180 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
apps/ocr/__init__.py
View File

@@ -17,47 +17,43 @@ from project_tools.api import register_tool
from acls.api import class_permissions
from scheduler.api import register_interval_job
from statistics.api import register_statistics
from queue_manager.models import Queue
from .conf.settings import (AUTOMATIC_OCR, QUEUE_PROCESSING_INTERVAL)
from .models import DocumentQueue, QueueTransformation
from .models import OCRProcessingSingleton
from .tasks import task_process_document_queues
from .permissions import PERMISSION_OCR_DOCUMENT
from .exceptions import AlreadyQueued
from . import models as ocr_models
from .statistics import get_statistics
from .literals import OCR_QUEUE_NAME
logger = logging.getLogger(__name__)
from .links import (submit_document, re_queue_multiple_document,
queue_document_multiple_delete, document_queue_disable,
document_queue_enable, all_document_ocr_cleanup, queue_document_list,
ocr_tool_link, setup_queue_transformation_list,
setup_queue_transformation_create, setup_queue_transformation_edit,
setup_queue_transformation_delete, submit_document_multiple)
queue_document_multiple_delete, ocr_disable,
ocr_enable, all_document_ocr_cleanup, ocr_log,
ocr_tool_link, submit_document_multiple)
bind_links([Document], [submit_document])
bind_links([DocumentQueue], [document_queue_disable, document_queue_enable, setup_queue_transformation_list])
bind_links([QueueTransformation], [setup_queue_transformation_edit, setup_queue_transformation_delete])
bind_links([OCRProcessingSingleton], [ocr_disable, ocr_enable])
#bind_links([QueueTransformation], [setup_queue_transformation_edit, setup_queue_transformation_delete])
register_multi_item_links(['queue_document_list'], [re_queue_multiple_document, queue_document_multiple_delete])
#register_multi_item_links(['queue_document_list'], [re_queue_multiple_document, queue_document_multiple_delete])
bind_links(['setup_queue_transformation_create', 'setup_queue_transformation_edit', 'setup_queue_transformation_delete', 'document_queue_disable', 'document_queue_enable', 'queue_document_list', 'setup_queue_transformation_list'], [queue_document_list], menu_name='secondary_menu')
bind_links(['setup_queue_transformation_edit', 'setup_queue_transformation_delete', 'setup_queue_transformation_list', 'setup_queue_transformation_create'], [setup_queue_transformation_create], menu_name='sidebar')
#bind_links(['setup_queue_transformation_create', 'setup_queue_transformation_edit', 'setup_queue_transformation_delete', 'document_queue_disable', 'document_queue_enable', 'queue_document_list', 'setup_queue_transformation_list'], [queue_document_list], menu_name='secondary_menu')
#bind_links(['setup_queue_transformation_edit', 'setup_queue_transformation_delete', 'setup_queue_transformation_list', 'setup_queue_transformation_create'], [setup_queue_transformation_create], menu_name='sidebar')
register_maintenance_links([all_document_ocr_cleanup], namespace='ocr', title=_(u'OCR'))
register_multi_item_links(['folder_view', 'search', 'results', 'index_instance_node_view', 'document_find_duplicates', 'document_type_document_list', 'document_group_view', 'document_list', 'document_list_recent'], [submit_document_multiple])
#register_multi_item_links(['folder_view', 'search', 'results', 'index_instance_node_view', 'document_find_duplicates', 'document_type_document_list', 'document_group_view', 'document_list', 'document_list_recent'], [submit_document_multiple])
@transaction.commit_on_success
def create_default_queue():
def create_ocr_queue():
try:
default_queue, created = DocumentQueue.objects.get_or_create(name='default')
queue, created = Queue.objects.get_or_create(name=OCR_QUEUE_NAME, defaults={'label': _('OCR'), 'unique_names': True})
except DatabaseError:
transaction.rollback()
else:
if created:
default_queue.label = ugettext(u'Default')
default_queue.save()
@receiver(post_save, dispatch_uid='document_post_save', sender=DocumentVersion)
@@ -81,11 +77,10 @@ def document_post_save(sender, instance, **kwargs):
# task_process_document_queues()
@receiver(post_syncdb, dispatch_uid='create_default_queue', sender=ocr_models)
def create_default_queue_signal_handler(sender, **kwargs):
create_default_queue()
#@receiver(post_syncdb, dispatch_uid='create_ocr_queue_on_syncdb', sender=ocr_models)
#def create_ocr_queue_on_syncdb(sender, **kwargs):
register_interval_job('task_process_document_queues', _(u'Checks the OCR queue for pending documents.'), task_process_document_queues, seconds=QUEUE_PROCESSING_INTERVAL)
#register_interval_job('task_process_document_queues', _(u'Checks the OCR queue for pending documents.'), task_process_document_queues, seconds=QUEUE_PROCESSING_INTERVAL)
register_tool(ocr_tool_link)
@@ -93,4 +88,5 @@ class_permissions(Document, [
PERMISSION_OCR_DOCUMENT,
])
register_statistics(get_statistics)
#register_statistics(get_statistics)
create_ocr_queue()

2
apps/ocr/admin.py
View File

@@ -1,3 +1,4 @@
"""
from __future__ import absolute_import
from django.contrib import admin
@@ -18,3 +19,4 @@ class DocumentQueueAdmin(admin.ModelAdmin):
admin.site.register(DocumentQueue, DocumentQueueAdmin)
"""

2
apps/ocr/api.py
View File

@@ -87,7 +87,7 @@ def do_document_ocr(queue_document):
parser, if the parser fails or if there is no parser registered for
the document mimetype do a visual OCR by calling tesseract
"""
for document_page in queue_document.document.pages.all():
for document_page in queue_document.document_version.pages.all():
try:
# Try to extract text by means of a parser
parse_document_page(document_page)

8
apps/ocr/exceptions.py
View File

@@ -21,3 +21,11 @@ class UnpaperError(Exception):
class ReQueueError(Exception):
pass
class OCRProcessingAlreadyDisabled(Exception):
pass
class OCRProcessingAlreadyEnabled(Exception):
pass

2
apps/ocr/forms.py
View File

@@ -1,3 +1,4 @@
"""
from __future__ import absolute_import
from django import forms
@@ -19,3 +20,4 @@ class QueueTransformationForm_create(forms.ModelForm):
class Meta:
model = QueueTransformation
exclude = ('content_type', 'object_id')
"""

24
apps/ocr/links.py
View File

@@ -7,7 +7,18 @@ from navigation.api import Link
from .permissions import (PERMISSION_OCR_DOCUMENT,
PERMISSION_OCR_DOCUMENT_DELETE, PERMISSION_OCR_QUEUE_ENABLE_DISABLE,
PERMISSION_OCR_CLEAN_ALL_PAGES)
from .models import OCRProcessingSingleton
def is_enabled(context):
return OCRProcessingSingleton.get().is_enabled()
def is_disabled(context):
return not OCRProcessingSingleton.get().is_enabled()
ocr_log = Link(text=_(u'queue document list'), view='ocr_log', sprite='text', permissions=[PERMISSION_OCR_DOCUMENT])
ocr_disable = Link(text=_(u'disable OCR processing'), view='ocr_disable', sprite='control_stop_blue', permissions=[PERMISSION_OCR_QUEUE_ENABLE_DISABLE], conditional_disable=is_disabled)
ocr_enable = Link(text=_(u'enable OCR processing'), view='ocr_enable', sprite='control_play_blue', permissions=[PERMISSION_OCR_QUEUE_ENABLE_DISABLE], conditional_disable=is_enabled)
submit_document = Link(text=_('submit to OCR queue'), view='submit_document', args='object.id', sprite='hourglass_add', permissions=[PERMISSION_OCR_DOCUMENT])
submit_document_multiple = Link(text=_('submit to OCR queue'), view='submit_document_multiple', sprite='hourglass_add', permissions=[PERMISSION_OCR_DOCUMENT])
re_queue_document = Link(text=_('re-queue'), view='re_queue_document', args='object.id', sprite='hourglass_add', permissions=[PERMISSION_OCR_DOCUMENT])
@@ -15,15 +26,12 @@ re_queue_multiple_document = Link(text=_('re-queue'), view='re_queue_multiple_do
queue_document_delete = Link(text=_(u'delete'), view='queue_document_delete', args='object.id', sprite='hourglass_delete', permissions=[PERMISSION_OCR_DOCUMENT_DELETE])
queue_document_multiple_delete = Link(text=_(u'delete'), view='queue_document_multiple_delete', sprite='hourglass_delete', permissions=[PERMISSION_OCR_DOCUMENT_DELETE])
document_queue_disable = Link(text=_(u'stop queue'), view='document_queue_disable', args='queue.id', sprite='control_stop_blue', permissions=[PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
document_queue_enable = Link(text=_(u'activate queue'), view='document_queue_enable', args='queue.id', sprite='control_play_blue', permissions=[PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
all_document_ocr_cleanup = Link(text=_(u'clean up pages content'), view='all_document_ocr_cleanup', sprite='text_strikethrough', permissions=[PERMISSION_OCR_CLEAN_ALL_PAGES], description=_(u'Runs a language filter to remove common OCR mistakes from document pages content.'))
queue_document_list = Link(text=_(u'queue document list'), view='queue_document_list', sprite='hourglass', permissions=[PERMISSION_OCR_DOCUMENT])
ocr_tool_link = Link(text=_(u'OCR'), view='queue_document_list', sprite='hourglass', icon='text.png', permissions=[PERMISSION_OCR_DOCUMENT], children_view_regex=[r'queue_', r'document_queue'])
ocr_tool_link = Link(text=_(u'OCR'), view='ocr_log', sprite='hourglass', icon='text.png', permissions=[PERMISSION_OCR_DOCUMENT]) # children_view_regex=[r'queue_', r'document_queue'])
setup_queue_transformation_list = Link(text=_(u'transformations'), view='setup_queue_transformation_list', args='queue.pk', sprite='shape_move_front')
setup_queue_transformation_create = Link(text=_(u'add transformation'), view='setup_queue_transformation_create', args='queue.pk', sprite='shape_square_add')
setup_queue_transformation_edit = Link(text=_(u'edit'), view='setup_queue_transformation_edit', args='transformation.pk', sprite='shape_square_edit')
setup_queue_transformation_delete = Link(text=_(u'delete'), view='setup_queue_transformation_delete', args='transformation.pk', sprite='shape_square_delete')
#setup_queue_transformation_list = Link(text=_(u'transformations'), view='setup_queue_transformation_list', args='queue.pk', sprite='shape_move_front')
#setup_queue_transformation_create = Link(text=_(u'add transformation'), view='setup_queue_transformation_create', args='queue.pk', sprite='shape_square_add')
#setup_queue_transformation_edit = Link(text=_(u'edit'), view='setup_queue_transformation_edit', args='transformation.pk', sprite='shape_square_edit')
#setup_queue_transformation_delete = Link(text=_(u'delete'), view='setup_queue_transformation_delete', args='transformation.pk', sprite='shape_square_delete')

28
apps/ocr/literals.py
View File

@@ -1,25 +1,27 @@
from django.utils.translation import ugettext_lazy as _
DOCUMENTQUEUE_STATE_STOPPED = 's'
DOCUMENTQUEUE_STATE_ACTIVE = 'a'
OCR_STATE_DISABLED = 'd'
OCR_STATE_ENABLED = 'e'
DOCUMENTQUEUE_STATE_CHOICES = (
(DOCUMENTQUEUE_STATE_STOPPED, _(u'stopped')),
(DOCUMENTQUEUE_STATE_ACTIVE, _(u'active')),
OCR_STATE_CHOICES = (
(OCR_STATE_DISABLED, _(u'disabled')),
(OCR_STATE_ENABLED, _(u'enabled')),
)
QUEUEDOCUMENT_STATE_PENDING = 'p'
QUEUEDOCUMENT_STATE_PROCESSING = 'i'
QUEUEDOCUMENT_STATE_ERROR = 'e'
#QUEUEDOCUMENT_STATE_PENDING = 'p'
#QUEUEDOCUMENT_STATE_PROCESSING = 'i'
#QUEUEDOCUMENT_STATE_ERROR = 'e'
QUEUEDOCUMENT_STATE_CHOICES = (
(QUEUEDOCUMENT_STATE_PENDING, _(u'pending')),
(QUEUEDOCUMENT_STATE_PROCESSING, _(u'processing')),
(QUEUEDOCUMENT_STATE_ERROR, _(u'error')),
)
#QUEUEDOCUMENT_STATE_CHOICES = (
# (QUEUEDOCUMENT_STATE_PENDING, _(u'pending')),
# (QUEUEDOCUMENT_STATE_PROCESSING, _(u'processing')),
# (QUEUEDOCUMENT_STATE_ERROR, _(u'error')),
#)
DEFAULT_OCR_FILE_FORMAT = u'tiff'
DEFAULT_OCR_FILE_EXTENSION = u'tif'
UNPAPER_FILE_FORMAT = u'ppm'
OCR_QUEUE_NAME = 'ocr'

24
apps/ocr/managers.py
View File

@@ -2,19 +2,19 @@ from __future__ import absolute_import
from django.db import models
from .exceptions import AlreadyQueued
#from .exceptions import AlreadyQueued
class DocumentQueueManager(models.Manager):
'''
Module manager class to handle adding documents to an OCR document
queue
'''
def queue_document(self, document, queue_name='default'):
document_queue = self.model.objects.get(name=queue_name)
if document_queue.queuedocument_set.filter(document=document):
raise AlreadyQueued
class OCRProcessingManager(models.Manager):
"""
Module manager class to handle adding documents to an OCR queue
"""
def queue_document(self, document):
pass
#document_queue = self.model.objects.get(name=queue_name)
#if document_queue.queuedocument_set.filter(document_version=document.latest_version):
# raise AlreadyQueued
document_queue.queuedocument_set.create(document=document, delay=True)
#document_queue.queuedocument_set.create(document_version=document.latest_version, delay=True)
return document_queue
#return document_queue

111
apps/ocr/models.py
View File

@@ -1,7 +1,7 @@
from __future__ import absolute_import
from ast import literal_eval
from datetime import datetime
import datetime
from django.db import models
from django.utils.translation import ugettext_lazy as _
@@ -11,35 +11,109 @@ from django.contrib.contenttypes.models import ContentType
from django.contrib.contenttypes import generic
from django.core.exceptions import ValidationError
from documents.models import Document
from common.models import Singleton
from documents.models import Document, DocumentVersion
from converter.api import get_available_transformations_choices
from sources.managers import SourceTransformationManager
from .literals import (DOCUMENTQUEUE_STATE_CHOICES,
QUEUEDOCUMENT_STATE_PENDING, QUEUEDOCUMENT_STATE_CHOICES,
QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_ACTIVE)
from .managers import DocumentQueueManager
from .exceptions import ReQueueError
from .literals import (OCR_STATE_CHOICES, OCR_STATE_ENABLED,
OCR_STATE_DISABLED)
from .managers import OCRProcessingManager
from .exceptions import (ReQueueError, OCRProcessingAlreadyDisabled,
OCRProcessingAlreadyEnabled)
class DocumentQueue(models.Model):
name = models.CharField(max_length=64, unique=True, verbose_name=_(u'name'))
label = models.CharField(max_length=64, verbose_name=_(u'label'))
class OCRProcessingSingleton(Singleton):
state = models.CharField(max_length=4,
choices=DOCUMENTQUEUE_STATE_CHOICES,
default=DOCUMENTQUEUE_STATE_ACTIVE,
choices=OCR_STATE_CHOICES,
default=OCR_STATE_ENABLED,
verbose_name=_(u'state'))
objects = DocumentQueueManager()
class Meta:
verbose_name = _(u'document queue')
verbose_name_plural = _(u'document queues')
#objects = AnonymousUserSingletonManager()
def __unicode__(self):
return self.label
return ugettext('OCR processing')
def disable(self):
if self.state == OCR_STATE_DISABLED:
raise OCRProcessingAlreadyDisabled
self.state = OCR_STATE_DISABLED
self.save()
def enable(self):
if self.state == OCR_STATE_ENABLED:
raise OCRProcessingAlreadyEnabled
self.state = OCR_STATE_ENABLED
self.save()
def is_enabled(self):
return self.state == OCR_STATE_ENABLED
class Meta:
verbose_name = verbose_name_plural = _(u'OCR processing properties')
"""
class OCRLog(models.Model):
#queue = models.ForeignKey(Queue, verbose_name=_(u'queue'))
document_version = models.ForeignKey(DocumentVersion, verbose_name=_(u'document version'))
datetime = models.DateTimeField(verbose_name=_(u'date time'), default=lambda: datetime.datetime.now(), db_index=True)
delay = models.BooleanField(verbose_name=_(u'delay OCR'), default=False)
#state = models.CharField(max_length=4,
# choices=QUEUEDOCUMENT_STATE_CHOICES,
# default=QUEUEDOCUMENT_STATE_PENDING,
# verbose_name=_(u'state'))
result = models.TextField(blank=True, null=True, verbose_name=_(u'result'))
#node_name = models.CharField(max_length=32, verbose_name=_(u'node name'), blank=True, null=True)
class Meta:
ordering = ('datetime',)
verbose_name = _(u'OCR log entry')
verbose_name_plural = _(u'OCR log entries')
#def get_transformation_list(self):
# return QueueTransformation.transformations.get_for_object_as_list(self)
def requeue(self):
pass
#if self.state == QUEUEDOCUMENT_STATE_PROCESSING:
# raise ReQueueError
#else:
# self.datetime_submitted = datetime.now()
# self.state = QUEUEDOCUMENT_STATE_PENDING
# self.delay = False
# self.result = None
# self.node_name = None
# self.save()
def __unicode__(self):
try:
return unicode(self.document)
except ObjectDoesNotExist:
return ugettext(u'Missing document.')
"""
#class DocumentQueue(models.Model):
# name = models.CharField(max_length=64, unique=True, verbose_name=_(u'name'))
# label = models.CharField(max_length=64, verbose_name=_(u'label'))
# state = models.CharField(max_length=4,
# choices=DOCUMENTQUEUE_STATE_CHOICES,
# default=DOCUMENTQUEUE_STATE_ACTIVE,
# verbose_name=_(u'state'))
#
# objects = DocumentQueueManager()#
#
# class Meta:
# verbose_name = _(u'document queue')
# verbose_name_plural = _(u'document queues')#
#
# def __unicode__(self):
# return self.label
"""
class QueueDocument(models.Model):
document_queue = models.ForeignKey(DocumentQueue, verbose_name=_(u'document queue'))
document = models.ForeignKey(Document, verbose_name=_(u'document'))
@@ -121,3 +195,4 @@ class QueueTransformation(models.Model):
ordering = ('order',)
verbose_name = _(u'document queue transformation')
verbose_name_plural = _(u'document queue transformations')
"""

2
apps/ocr/permissions.py
View File

@@ -7,6 +7,6 @@ from permissions.models import Permission, PermissionNamespace
ocr_namespace = PermissionNamespace('ocr', _(u'OCR'))
PERMISSION_OCR_DOCUMENT = Permission.objects.register(ocr_namespace, 'ocr_document', _(u'Submit documents for OCR'))
PERMISSION_OCR_DOCUMENT_DELETE = Permission.objects.register(ocr_namespace, 'ocr_document_delete', _(u'Delete documents from OCR queue'))
PERMISSION_OCR_QUEUE_ENABLE_DISABLE = Permission.objects.register(ocr_namespace, 'ocr_queue_enable_disable', _(u'Can enable/disable the OCR queue'))
PERMISSION_OCR_QUEUE_ENABLE_DISABLE = Permission.objects.register(ocr_namespace, 'ocr_queue_enable_disable', _(u'Can enable/disable the OCR processing'))
PERMISSION_OCR_CLEAN_ALL_PAGES = Permission.objects.register(ocr_namespace, 'ocr_clean_all_pages', _(u'Can execute the OCR clean up on all document pages'))
PERMISSION_OCR_QUEUE_EDIT = Permission.objects.register(ocr_namespace, 'ocr_queue_edit', _(u'Can edit an OCR queue properties'))

2
apps/ocr/statistics.py
View File

@@ -2,7 +2,7 @@ from __future__ import absolute_import
from django.utils.translation import ugettext as _
from .models import DocumentQueue, QueueDocument
#from .models import DocumentQueue, QueueDocument
def get_statistics():

8
apps/ocr/tasks.py
View File

@@ -10,10 +10,10 @@ from job_processor.api import process_job
from lock_manager import Lock, LockError
from .api import do_document_ocr
from .literals import (QUEUEDOCUMENT_STATE_PENDING,
QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_ACTIVE,
QUEUEDOCUMENT_STATE_ERROR)
from .models import QueueDocument, DocumentQueue
#from .literals import (QUEUEDOCUMENT_STATE_PENDING,
# QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_ACTIVE,
# QUEUEDOCUMENT_STATE_ERROR)
#from .models import QueueDocument, DocumentQueue
from .conf.settings import NODE_CONCURRENT_EXECUTION, REPLICATION_DELAY
LOCK_EXPIRE = 60 * 10 # Lock expires in 10 minutes

8
apps/ocr/urls.py
View File

@@ -1,16 +1,18 @@
from django.conf.urls.defaults import patterns, url
urlpatterns = patterns('ocr.views',
url(r'^log/$', 'ocr_log', (), 'ocr_log'),
url(r'^processing/enable/$', 'ocr_enable', (), 'ocr_enable'),
url(r'^processing/disable/$', 'ocr_disable', (), 'ocr_disable'),
url(r'^document/(?P<document_id>\d+)/submit/$', 'submit_document', (), 'submit_document'),
url(r'^document/multiple/submit/$', 'submit_document_multiple', (), 'submit_document_multiple'),
url(r'^queue/document/list/$', 'queue_document_list', (), 'queue_document_list'),
url(r'^queue/document/(?P<queue_document_id>\d+)/delete/$', 'queue_document_delete', (), 'queue_document_delete'),
url(r'^queue/document/multiple/delete/$', 'queue_document_multiple_delete', (), 'queue_document_multiple_delete'),
url(r'^queue/document/(?P<queue_document_id>\d+)/re-queue/$', 're_queue_document', (), 're_queue_document'),
url(r'^queue/document/multiple/re-queue/$', 're_queue_multiple_document', (), 're_queue_multiple_document'),
url(r'^queue/(?P<document_queue_id>\d+)/enable/$', 'document_queue_enable', (), 'document_queue_enable'),
url(r'^queue/(?P<document_queue_id>\d+)/disable/$', 'document_queue_disable', (), 'document_queue_disable'),
url(r'^document/all/clean_up/$', 'all_document_ocr_cleanup', (), 'all_document_ocr_cleanup'),

175
apps/ocr/views.py
View File

@@ -18,38 +18,35 @@ from acls.models import AccessEntry
from .permissions import (PERMISSION_OCR_DOCUMENT,
PERMISSION_OCR_DOCUMENT_DELETE, PERMISSION_OCR_QUEUE_ENABLE_DISABLE,
PERMISSION_OCR_CLEAN_ALL_PAGES, PERMISSION_OCR_QUEUE_EDIT)
from .models import DocumentQueue, QueueDocument, QueueTransformation
from .literals import (QUEUEDOCUMENT_STATE_PROCESSING,
DOCUMENTQUEUE_STATE_ACTIVE, DOCUMENTQUEUE_STATE_STOPPED)
from .exceptions import AlreadyQueued, ReQueueError
from .models import OCRProcessingSingleton
#from .literals import (QUEUEDOCUMENT_STATE_PROCESSING,
# DOCUMENTQUEUE_STATE_ACTIVE, DOCUMENTQUEUE_STATE_STOPPED)
from .exceptions import (AlreadyQueued, ReQueueError, OCRProcessingAlreadyDisabled,
OCRProcessingAlreadyEnabled)
from .api import clean_pages
from .forms import QueueTransformationForm, QueueTransformationForm_create
#from .forms import QueueTransformationForm, QueueTransformationForm_create
def queue_document_list(request, queue_name='default'):
def ocr_log(request):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_DOCUMENT])
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'documents in queue: %s') % document_queue,
'hide_object': True,
'queue': document_queue,
'object_name': _(u'document queue'),
context = {
'queue': OCRProcessingSingleton.get(),
'object_name': _(u'OCR processing'), # TODO fix, not working
'navigation_object_name': 'queue',
'list_object_variable_name': 'queue_document',
'object_list': [],
'title': _(u'OCR log items'),
#'hide_object': True,
#'hide_link': True,
'extra_columns': [
{'name': 'document', 'attribute': encapsulate(lambda x: document_link(x.document) if hasattr(x, 'document') else _(u'Missing document.'))},
{'name': _(u'thumbnail'), 'attribute': encapsulate(lambda x: document_thumbnail(x.document))},
{'name': 'submitted', 'attribute': encapsulate(lambda x: unicode(x.datetime_submitted).split('.')[0]), 'keep_together':True},
{'name': 'delay', 'attribute': 'delay'},
{'name': 'state', 'attribute': encapsulate(lambda x: x.get_state_display())},
{'name': 'node', 'attribute': 'node_name'},
{'name': 'result', 'attribute': 'result'},
{'name': _(u'document'), 'attribute': encapsulate(lambda x: document_link(x.document_version.document) if hasattr(x, 'document_version') else _(u'Missing document.'))},
{'name': _(u'version'), 'attribute': 'document_version'},
{'name': _(u'thumbnail'), 'attribute': encapsulate(lambda x: document_thumbnail(x.document_version.document))},
{'name': _('submitted'), 'attribute': encapsulate(lambda x: unicode(x.datetime_submitted).split('.')[0]), 'keep_together':True},
#{'name': _('delay'), 'attribute': 'delay'},
#{'name': _('state'), 'attribute': encapsulate(lambda x: x.get_state_display())},
#{'name': _('node'), 'attribute': 'node_name'},
{'name': _('result'), 'attribute': 'result'},
],
'multi_select_as_buttons': True,
'sidebar_subtemplates_list': [
@@ -57,13 +54,75 @@ def queue_document_list(request, queue_name='default'):
'name': 'generic_subtemplate.html',
'context': {
'side_bar': True,
'title': _(u'document queue properties'),
'content': _(u'Current state: %s') % document_queue.get_state_display(),
'title': _(u'OCR processing properties'),
'content': _(u'Current state: %s') % OCRProcessingSingleton.get().get_state_display(),
}
}
]
},
)
}
return render_to_response('generic_list.html', context,
context_instance=RequestContext(request))
# 'queue': document_queue,
# 'object_name': _(u'document queue'),
# 'navigation_object_name': 'queue',
# 'list_object_variable_name': 'queue_document',
# },
#)
def ocr_disable(request):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
next = request.POST.get('next', request.GET.get('next', request.META.get('HTTP_REFERER', None)))
previous = request.POST.get('previous', request.GET.get('previous', request.META.get('HTTP_REFERER', None)))
if request.method == 'POST':
try:
OCRProcessingSingleton.get().disable()
except OCRProcessingAlreadyDisabled:
messages.warning(request, _(u'OCR processing already disabled.'))
return HttpResponseRedirect(previous)
else:
messages.success(request, _(u'OCR processing disabled successfully.'))
return HttpResponseRedirect(next)
return render_to_response('generic_confirm.html', {
'queue': OCRProcessingSingleton.get(),
'navigation_object_name': 'queue',
'title': _(u'Are you sure you wish to disable OCR processing?'),
'next': next,
'previous': previous,
'form_icon': u'control_stop_blue.png',
}, context_instance=RequestContext(request))
def ocr_enable(request):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
next = request.POST.get('next', request.GET.get('next', request.META.get('HTTP_REFERER', None)))
previous = request.POST.get('previous', request.GET.get('previous', request.META.get('HTTP_REFERER', None)))
if request.method == 'POST':
try:
OCRProcessingSingleton.get().enable()
except OCRProcessingAlreadyDisabled:
messages.warning(request, _(u'OCR processing already enabled.'))
return HttpResponseRedirect(previous)
else:
messages.success(request, _(u'OCR processing enabled successfully.'))
return HttpResponseRedirect(next)
return render_to_response('generic_confirm.html', {
'queue': OCRProcessingSingleton.get(),
'navigation_object_name': 'queue',
'title': _(u'Are you sure you wish to enable OCR processing?'),
'next': next,
'previous': previous,
'form_icon': u'control_play_blue.png',
}, context_instance=RequestContext(request))
def queue_document_delete(request, queue_document_id=None, queue_document_id_list=None):
@@ -175,12 +234,12 @@ def re_queue_document(request, queue_document_id=None, queue_document_id_list=No
messages.success(
request,
_(u'Document: %(document)s was re-queued to the OCR queue: %(queue)s') % {
'document': queue_document.document,
'document': queue_document.document_version.document,
'queue': queue_document.document_queue.label
}
)
except Document.DoesNotExist:
messages.error(request, _(u'Document id#: %d, no longer exists.') % queue_document.document_id)
messages.error(request, _(u'Document no longer in queue.'))
except ReQueueError:
messages.warning(
request,
@@ -208,60 +267,6 @@ def re_queue_multiple_document(request):
return re_queue_document(request, queue_document_id_list=request.GET.get('id_list', []))
def document_queue_disable(request, document_queue_id):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
next = request.POST.get('next', request.GET.get('next', request.META.get('HTTP_REFERER', None)))
previous = request.POST.get('previous', request.GET.get('previous', request.META.get('HTTP_REFERER', None)))
document_queue = get_object_or_404(DocumentQueue, pk=document_queue_id)
if document_queue.state == DOCUMENTQUEUE_STATE_STOPPED:
messages.warning(request, _(u'Document queue: %s, already stopped.') % document_queue)
return HttpResponseRedirect(previous)
if request.method == 'POST':
document_queue.state = DOCUMENTQUEUE_STATE_STOPPED
document_queue.save()
messages.success(request, _(u'Document queue: %s, stopped successfully.') % document_queue)
return HttpResponseRedirect(next)
return render_to_response('generic_confirm.html', {
'queue': document_queue,
'navigation_object_name': 'queue',
'title': _(u'Are you sure you wish to disable document queue: %s') % document_queue,
'next': next,
'previous': previous,
'form_icon': u'control_stop_blue.png',
}, context_instance=RequestContext(request))
def document_queue_enable(request, document_queue_id):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_QUEUE_ENABLE_DISABLE])
next = request.POST.get('next', request.GET.get('next', request.META.get('HTTP_REFERER', None)))
previous = request.POST.get('previous', request.GET.get('previous', request.META.get('HTTP_REFERER', None)))
document_queue = get_object_or_404(DocumentQueue, pk=document_queue_id)
if document_queue.state == DOCUMENTQUEUE_STATE_ACTIVE:
messages.warning(request, _(u'Document queue: %s, already active.') % document_queue)
return HttpResponseRedirect(previous)
if request.method == 'POST':
document_queue.state = DOCUMENTQUEUE_STATE_ACTIVE
document_queue.save()
messages.success(request, _(u'Document queue: %s, activated successfully.') % document_queue)
return HttpResponseRedirect(next)
return render_to_response('generic_confirm.html', {
'queue': document_queue,
'navigation_object_name': 'queue',
'title': _(u'Are you sure you wish to activate document queue: %s') % document_queue,
'next': next,
'previous': previous,
'form_icon': u'control_play_blue.png',
}, context_instance=RequestContext(request))
def all_document_ocr_cleanup(request):
Permission.objects.check_permissions(request.user, [PERMISSION_OCR_CLEAN_ALL_PAGES])