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

Clean up (unused imports, PEP8, etc)

Browse Source
This commit is contained in:
Roberto Rosario
2012-06-08 16:43:54 -04:00
parent c24a3f935e
commit 58f027db60
34 changed files with 121 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files
apps/ocr/__init__.py
+16 -15
View File
@@ -29,25 +29,25 @@ from . import models as ocr_models
logger = logging.getLogger(__name__)
#Links
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])
re_queue_multiple_document = Link(text=_('re-queue'),view='re_queue_multiple_document',sprite='hourglass_add',permissions=[PERMISSION_OCR_DOCUMENT])
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])
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])
re_queue_multiple_document = Link(text=_('re-queue'), view='re_queue_multiple_document', sprite='hourglass_add', permissions=[PERMISSION_OCR_DOCUMENT])
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])
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.'))
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'])
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'])
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')
bind_links([Document], [submit_document])
bind_links([DocumentQueue], [document_queue_disable, document_queue_enable, setup_queue_transformation_list])
@@ -93,6 +93,7 @@ def document_post_save(sender, instance, **kwargs):
# logger.debug('got call_queue signal: %s' % 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()
apps/ocr/conf/settings.py
-1
View File
@@ -2,7 +2,6 @@
from django.utils.translation import ugettext_lazy as _
<<<<<<< HEAD
from smart_settings.api import Setting, SettingNamespace
namespace = SettingNamespace('ocr', _(u'OCR'), module='ocr.conf.settings')
apps/ocr/parsers/__init__.py
+23 -23
View File
@@ -52,7 +52,7 @@ def parse_document_page(document_page, descriptor=None, mimetype=None):
# If parser was successfull there is no need to try
# others in the list for this mimetype
return
raise ParserError('Parser list exhausted')
except KeyError:
raise ParserUnknownFile
@@ -113,7 +113,7 @@ class OfficeParser(Parser):
except OfficeConversionError, msg:
logger.error(msg)
raise ParserError
class PopplerParser(Parser):
"""
@@ -123,11 +123,11 @@ class PopplerParser(Parser):
self.pdftotext_path = PDFTOTEXT_PATH if PDFTOTEXT_PATH else u'/usr/bin/pdftotext'
if not os.path.exists(self.pdftotext_path):
raise ParserError('cannot find pdftotext executable')
logger.debug('self.pdftotext_path: %s' % self.pdftotext_path)
def parse(self, document_page, descriptor=None):
logger.debug('parsing PDF with PopplerParser')
pagenum = str(document_page.page_number)
logger.debug('self.pdftotext_path: %s' % self.pdftotext_path)
def parse(self, document_page, descriptor=None):
logger.debug('parsing PDF with PopplerParser')
pagenum = str(document_page.page_number)
if descriptor:
destination_descriptor, temp_filepath = tempfile.mkstemp(dir=TEMPORARY_DIRECTORY)
@@ -135,23 +135,23 @@ class PopplerParser(Parser):
document_file = temp_filepath
else:
document_file = document_save_to_temp_dir(document_page.document, document_page.document.checksum)
logger.debug('document_file: %s', document_file)
logger.debug('parsing PDF page %s' % pagenum)
logger.debug('parsing PDF page %s' % pagenum)
command = []
command.append(self.pdftotext_path)
command.append('-f')
command.append(pagenum)
command.append('-l')
command.append(pagenum)
command.append(document_file)
command.append('-')
command = []
command.append(self.pdftotext_path)
command.append('-f')
command.append(pagenum)
command.append('-l')
command.append(pagenum)
command.append(document_file)
command.append('-')
proc = subprocess.Popen(command, close_fds=True, stderr=subprocess.PIPE, stdout=subprocess.PIPE)
return_code = proc.wait()
if return_code != 0:
proc = subprocess.Popen(command, close_fds=True, stderr=subprocess.PIPE, stdout=subprocess.PIPE)
return_code = proc.wait()
if return_code != 0:
logger.error(proc.stderr.readline())
raise ParserError
@@ -160,9 +160,9 @@ class PopplerParser(Parser):
logger.debug('Parser didn\'t any output')
raise ParserError('No output')
document_page.content = output
document_page.page_label = _(u'Text extracted from PDF')
document_page.save()
document_page.content = output
document_page.page_label = _(u'Text extracted from PDF')
document_page.save()
register_parser(mimetypes=[u'application/pdf'], parsers=[PopplerParser, SlateParser])
apps/ocr/tasks.py
+1 -2
View File
@@ -14,8 +14,7 @@ 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,
QUEUE_PROCESSING_INTERVAL)
from .conf.settings import NODE_CONCURRENT_EXECUTION, REPLICATION_DELAY
LOCK_EXPIRE = 60 * 10 # Lock expires in 10 minutes
# TODO: Tie LOCK_EXPIRATION with hard task timeout
apps/ocr/views.py
+2 -3
View File
@@ -19,9 +19,8 @@ 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 .literals import (QUEUEDOCUMENT_STATE_PROCESSING,
DOCUMENTQUEUE_STATE_ACTIVE, DOCUMENTQUEUE_STATE_STOPPED)
from .exceptions import AlreadyQueued, ReQueueError
from .api import clean_pages
from .forms import QueueTransformationForm, QueueTransformationForm_create