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

Update and re-enable ocr app

Browse Source
This commit is contained in:
Roberto Rosario
2012-09-16 03:30:32 -04:00
parent a4bbc65508
commit 6f585a2836
14 changed files with 109 additions and 268 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
apps/ocr/registry.py
View File

@@ -2,60 +2,52 @@ from __future__ import absolute_import
from django.utils.translation import ugettext_lazy as _
from smart_settings import LocalScope
from smart_settings import LocalScope, ClusterScope
from .icons import icon_submit_document
from .literals import (DEFAULT_TESSERACT_PATH, DEFAULT_TESSERACT_LANGUAGE,
DEFAULT_REPLICATION_DELAY, DEFAULT_UNPAPER_PATH, DEFAULT_PDFTOTEXT_PATH)
from .links import all_document_ocr_cleanup
label = _(u'OCR')
description = _(u'Handles optical character recognition.')
icon = icon_submit_document
dependencies = ['app_registry', 'icons', 'navigation']
#maintenance_links = [all_document_ocr_cleanup]
settings = [
{
'name': 'AUTOMATIC_OCR',
'default': True,
'description': _(u'Automatically queue newly created documents for OCR.'),
'scopes': [ClusterScope()]
},
{
'name': 'TESSERACT_PATH',
'default': u'/usr/bin/tesseract',
'default': DEFAULT_TESSERACT_PATH,
'exists': True,
'scopes': [LocalScope()]
},
{
'name': 'TESSERACT_LANGUAGE',
'default': u'eng',
'scopes': [LocalScope()]
'default': DEFAULT_TESSERACT_LANGUAGE,
'scopes': [ClusterScope()]
},
{
'name': 'REPLICATION_DELAY',
'default': 0,
'default': DEFAULT_REPLICATION_DELAY,
'description': _(u'Amount of seconds to delay OCR of documents to allow for the node\'s storage replication overhead.'),
'scopes': [LocalScope()]
},
{
'name': 'NODE_CONCURRENT_EXECUTION',
'default': 1,
'description': _(u'Maximum amount of concurrent document OCRs a node can perform.'),
'scopes': [LocalScope()]
},
{
'name': 'AUTOMATIC_OCR',
'default': True,
'description': _(u'Automatically queue newly created documents for OCR.'),
'scopes': [LocalScope()]
},
{
'name': 'QUEUE_PROCESSING_INTERVAL',
'default': 10,
'description': _(u'Automatically queue newly created documents for OCR.'),
'scopes': [LocalScope()]
},
{
'name': 'UNPAPER_PATH',
'default': u'/usr/bin/unpaper',
'default': DEFAULT_UNPAPER_PATH,
'description': _(u'File path to unpaper program.'),
'exists': True,
'scopes': [LocalScope()]
},
{
'name': 'PDFTOTEXT_PATH',
'default': u'/usr/bin/pdftotext',
'default': DEFAULT_PDFTOTEXT_PATH,
'description': _(u'File path to poppler\'s pdftotext program used to extract text from PDF files.'),
'exists': True,
'scopes': [LocalScope()]