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

Cleanups, permissions separation into explicit module, absolute import update

Browse Source
This commit is contained in:
Roberto Rosario
2012-01-02 03:48:26 -04:00
parent 2781a211ec
commit 34311fb17e
107 changed files with 693 additions and 468 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
apps/ocr/__init__.py
View File

@@ -1,3 +1,5 @@
from __future__ import absolute_import
import logging
from django.db import transaction
@@ -14,21 +16,15 @@ from project_tools.api import register_tool
from scheduler.api import register_interval_job
from ocr.conf.settings import AUTOMATIC_OCR
from ocr.conf.settings import QUEUE_PROCESSING_INTERVAL
from ocr.models import DocumentQueue, QueueTransformation, QueueDocument
from ocr.tasks import task_process_document_queues
from .conf.settings import (AUTOMATIC_OCR, QUEUE_PROCESSING_INTERVAL)
from .models import DocumentQueue, QueueTransformation, QueueDocument
from .tasks import task_process_document_queues
from .permissions import (PERMISSION_OCR_DOCUMENT,
PERMISSION_OCR_DOCUMENT_DELETE, PERMISSION_OCR_QUEUE_ENABLE_DISABLE,
PERMISSION_OCR_CLEAN_ALL_PAGES, PERMISSION_OCR_QUEUE_EDIT)
logger = logging.getLogger(__name__)
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_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'))
#Links
submit_document = {'text': _('submit to OCR queue'), 'view': 'submit_document', 'args': 'object.id', 'famfam': 'hourglass_add', 'permissions': [PERMISSION_OCR_DOCUMENT]}
submit_document_multiple = {'text': _('submit to OCR queue'), 'view': 'submit_document_multiple', 'famfam': 'hourglass_add', 'permissions': [PERMISSION_OCR_DOCUMENT]}

15
apps/ocr/api.py
View File

@@ -1,4 +1,5 @@
#Some code from http://wiki.github.com/hoffstaetter/python-tesseract
from __future__ import absolute_import
import codecs
import os
@@ -13,14 +14,12 @@ from common.conf.settings import TEMPORARY_DIRECTORY
from converter.api import convert
from documents.models import DocumentPage
from ocr.conf.settings import TESSERACT_PATH
from ocr.conf.settings import TESSERACT_LANGUAGE
from ocr.exceptions import TesseractError, UnpaperError
from ocr.conf.settings import UNPAPER_PATH
from ocr.parsers import parse_document_page
from ocr.parsers.exceptions import ParserError, ParserUnknownFile
from ocr.literals import DEFAULT_OCR_FILE_FORMAT, UNPAPER_FILE_FORMAT, \
DEFAULT_OCR_FILE_EXTENSION
from .conf.settings import (TESSERACT_PATH, TESSERACT_LANGUAGE, UNPAPER_PATH)
from .exceptions import TesseractError, UnpaperError
from .parsers import parse_document_page
from .parsers.exceptions import ParserError, ParserUnknownFile
from .literals import (DEFAULT_OCR_FILE_FORMAT, UNPAPER_FILE_FORMAT,
DEFAULT_OCR_FILE_EXTENSION)
def get_language_backend():

4
apps/ocr/managers.py
View File

@@ -1,6 +1,8 @@
from __future__ import absolute_import
from django.db import models
from ocr.exceptions import AlreadyQueued
from .exceptions import AlreadyQueued
class DocumentQueueManager(models.Manager):

12
apps/ocr/models.py
View File

@@ -1,3 +1,5 @@
from __future__ import absolute_import
from ast import literal_eval
from datetime import datetime
@@ -13,11 +15,11 @@ from documents.models import Document
from converter.api import get_available_transformations_choices
from sources.managers import SourceTransformationManager
from ocr.literals import DOCUMENTQUEUE_STATE_STOPPED, \
DOCUMENTQUEUE_STATE_CHOICES, QUEUEDOCUMENT_STATE_PENDING, \
QUEUEDOCUMENT_STATE_CHOICES, QUEUEDOCUMENT_STATE_PROCESSING
from ocr.managers import DocumentQueueManager
from ocr.exceptions import ReQueueError
from .literals import (DOCUMENTQUEUE_STATE_STOPPED,
DOCUMENTQUEUE_STATE_CHOICES, QUEUEDOCUMENT_STATE_PENDING,
QUEUEDOCUMENT_STATE_CHOICES, QUEUEDOCUMENT_STATE_PROCESSING)
from .managers import DocumentQueueManager
from .exceptions import ReQueueError
class DocumentQueue(models.Model):

13
apps/ocr/permissions.py Normal file
View File

@@ -0,0 +1,13 @@
from __future__ import absolute_import
from django.utils.translation import ugettext_lazy as _
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_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'))

18
apps/ocr/tasks.py
View File

@@ -1,3 +1,5 @@
from __future__ import absolute_import
from datetime import timedelta, datetime
import platform
from time import sleep
@@ -9,15 +11,13 @@ from django.db.models import Q
from job_processor.api import process_job
from lock_manager import Lock, LockError
from ocr.api import do_document_ocr
from ocr.literals import QUEUEDOCUMENT_STATE_PENDING, \
QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_ACTIVE, \
QUEUEDOCUMENT_STATE_ERROR
from ocr.models import QueueDocument, DocumentQueue
from ocr.conf.settings import NODE_CONCURRENT_EXECUTION
from ocr.conf.settings import REPLICATION_DELAY
from ocr.conf.settings import CACHE_URI
from ocr.conf.settings import QUEUE_PROCESSING_INTERVAL
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 .conf.settings import (NODE_CONCURRENT_EXECUTION, REPLICATION_DELAY,
CACHE_URI, QUEUE_PROCESSING_INTERVAL)
LOCK_EXPIRE = 60 * 10 # Lock expires in 10 minutes
# TODO: Tie LOCK_EXPIRATION with hard task timeout

23
apps/ocr/views.py
View File

@@ -1,3 +1,5 @@
from __future__ import absolute_import
import socket
from django.http import HttpResponseRedirect
@@ -14,17 +16,16 @@ from documents.models import Document
from documents.widgets import document_link, document_thumbnail
from common.utils import encapsulate
from ocr import PERMISSION_OCR_DOCUMENT, PERMISSION_OCR_DOCUMENT_DELETE, \
PERMISSION_OCR_QUEUE_ENABLE_DISABLE, PERMISSION_OCR_CLEAN_ALL_PAGES, \
PERMISSION_OCR_QUEUE_EDIT
from ocr.models import DocumentQueue, QueueDocument, QueueTransformation
from ocr.literals import QUEUEDOCUMENT_STATE_PENDING, \
QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_STOPPED, \
DOCUMENTQUEUE_STATE_ACTIVE
from ocr.exceptions import AlreadyQueued, ReQueueError
from ocr.api import clean_pages
from ocr.forms import QueueTransformationForm, QueueTransformationForm_create
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_PENDING,
QUEUEDOCUMENT_STATE_PROCESSING, DOCUMENTQUEUE_STATE_STOPPED,
DOCUMENTQUEUE_STATE_ACTIVE)
from .exceptions import AlreadyQueued, ReQueueError
from .api import clean_pages
from .forms import QueueTransformationForm, QueueTransformationForm_create
def queue_document_list(request, queue_name='default'):