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

Add support for new document page structure

Browse Source 
Documents now have their own dedicated DocumentPage
submodel. The old DocumentPage is now called DocumentVersionPage.
This allows mappings between document pages and document version
pages, allowing renumbering, appending pages.
DocumentPages have a content_object to map them to any other
object. For now they only map to DocumentVersionPages.
New option added to the version upload form to append the
pages of the new version.
A new view was added to just append new pages with wraps the
new document version upload form and hides the append pages
checkbox set to True.
Add a new action, reset_pages to reset the pages of the
document to those of the latest version.

Missing: appending tests, checks for proper content_object in OCR and
document parsing.

Author: Roberto Rosario <roberto.rosario@mayan-edms.com>
Date:   Thu Oct 11 12:00:25 2019 -0400
This commit is contained in:
Roberto Rosario
2019-10-10 11:55:42 -04:00
parent 4a99a9df3e
commit 0699ad0556
87 changed files with 1973 additions and 707 deletions
Download Patch File Download Diff File Expand all files Collapse all files
mayan/apps/documents/models/document_page_models.py
+72 -46
View File
@@ -4,14 +4,16 @@ import logging
from furl import furl
from django.contrib.contenttypes.fields import GenericForeignKey
from django.contrib.contenttypes.models import ContentType
from django.db import models
from django.db.models import Max
from django.urls import reverse
from django.utils.encoding import force_text, python_2_unicode_compatible
from django.utils.functional import cached_property
from django.utils.translation import ugettext_lazy as _
from mayan.apps.converter.literals import DEFAULT_ZOOM_LEVEL, DEFAULT_ROTATION
from mayan.apps.converter.models import LayerTransformation
from mayan.apps.converter.transformations import (
BaseTransformation, TransformationResize, TransformationRotate,
@@ -26,7 +28,7 @@ from ..settings import (
setting_zoom_min_level
)
from .document_version_models import DocumentVersion
from .document_models import Document
__all__ = ('DocumentPage', 'DocumentPageResult')
logger = logging.getLogger(__name__)
@@ -35,16 +37,22 @@ logger = logging.getLogger(__name__)
@python_2_unicode_compatible
class DocumentPage(models.Model):
"""
Model that describes a document version page
Model that describes a document page
"""
document_version = models.ForeignKey(
on_delete=models.CASCADE, related_name='version_pages', to=DocumentVersion,
verbose_name=_('Document version')
document = models.ForeignKey(
on_delete=models.CASCADE, related_name='pages', to=Document,
verbose_name=_('Document')
)
enabled = models.BooleanField(default=True, verbose_name=_('Enabled'))
page_number = models.PositiveIntegerField(
db_index=True, default=1, editable=False,
verbose_name=_('Page number')
db_index=True, blank=True, null=True, verbose_name=_('Page number')
)
content_type = models.ForeignKey(
on_delete=models.CASCADE, to=ContentType
)
object_id = models.PositiveIntegerField()
content_object = GenericForeignKey(
ct_field='content_type', fk_field='object_id'
)
objects = DocumentPageManager()
@@ -52,6 +60,7 @@ class DocumentPage(models.Model):
class Meta:
ordering = ('page_number',)
unique_together = ('document', 'page_number')
verbose_name = _('Document page')
verbose_name_plural = _('Document pages')
@@ -60,7 +69,7 @@ class DocumentPage(models.Model):
@cached_property
def cache_partition(self):
partition, created = self.document_version.cache.partitions.get_or_create(
partition, created = self.document.cache.partitions.get_or_create(
name=self.uuid
)
return partition
@@ -69,19 +78,15 @@ class DocumentPage(models.Model):
self.cache_partition.delete()
super(DocumentPage, self).delete(*args, **kwargs)
def detect_orientation(self):
with self.document_version.open() as file_object:
converter = get_converter_class()(
file_object=file_object,
mime_type=self.document_version.mimetype
)
return converter.detect_orientation(
page_number=self.page_number
)
@property
def document(self):
return self.document_version.document
#def detect_orientation(self):
# with self.document_version.open() as file_object:
# converter = get_converter_class()(
# file_object=file_object,
# mime_type=self.document_version.mimetype
# )
# return converter.detect_orientation(
# page_number=self.page_number
# )
def generate_image(self, user=None, **kwargs):
transformation_list = self.get_combined_transformation_list(user=user, **kwargs)
@@ -90,7 +95,7 @@ class DocumentPage(models.Model):
# Check is transformed image is available
logger.debug('transformations cache filename: %s', combined_cache_filename)
if not setting_disable_transformed_image_cache.value and self.cache_partition.get_file(filename=combined_cache_filename):
if self.cache_partition.get_file(filename=combined_cache_filename):
logger.debug(
'transformations cache file "%s" found', combined_cache_filename
)
@@ -128,8 +133,7 @@ class DocumentPage(models.Model):
final_url.args = kwargs
final_url.path = reverse(
viewname='rest_api:documentpage-image', kwargs={
'pk': self.document.pk, 'version_pk': self.document_version.pk,
'page_pk': self.pk
'pk': self.document.pk, 'page_pk': self.pk
}
)
final_url.args['_hash'] = transformations_hash
@@ -190,12 +194,12 @@ class DocumentPage(models.Model):
return transformation_list
def get_image(self, transformations=None):
cache_filename = 'base_image'
cache_filename = 'document_page'
logger.debug('Page cache filename: %s', cache_filename)
cache_file = self.cache_partition.get_file(filename=cache_filename)
if not setting_disable_base_image_cache.value and cache_file:
if cache_file:
logger.debug('Page cache file "%s" found', cache_filename)
with cache_file.open() as file_object:
@@ -216,14 +220,25 @@ class DocumentPage(models.Model):
logger.debug('Page cache file "%s" not found', cache_filename)
try:
with self.document_version.get_intermediate_file() as file_object:
#with self.document_version.get_intermediate_file() as file_object:
#Render or get cached document version page
#self.content_object.generate_image()
self.content_object.get_image()
cache_filename = 'base_image'
cache_file = self.content_object.cache_partition.get_file(filename=cache_filename)
with cache_file.open() as file_object:
converter = get_converter_class()(
file_object=file_object
)
converter.seek_page(page_number=self.page_number - 1)
converter.seek_page(page_number=0)
#self.page_number - 1)
page_image = converter.get_page()
cache_filename = 'document_page'
# Since open "wb+" doesn't create files, create it explicitly
with self.cache_partition.create_file(filename=cache_filename) as file_object:
file_object.write(page_image.getvalue())
@@ -241,28 +256,39 @@ class DocumentPage(models.Model):
)
raise
def get_label(self):
return _(
'Page %(page_number)d out of %(total_pages)d of %(document)s'
) % {
'document': force_text(self.document),
'page_number': self.page_number,
'total_pages': self.document.pages_all.count()
}
get_label.short_description = _('Label')
@property
def is_in_trash(self):
return self.document.is_in_trash
def get_label(self):
return _(
'Page %(page_num)d out of %(total_pages)d of %(document)s'
) % {
'document': force_text(self.document),
'page_num': self.page_number,
'total_pages': self.document_version.pages_all.count()
}
get_label.short_description = _('Label')
def natural_key(self):
return (self.page_number, self.document_version.natural_key())
natural_key.dependencies = ['documents.DocumentVersion']
return (self.page_number, self.document.natural_key())
natural_key.dependencies = ['documents.Document']
def save(self, *args, **kwargs):
if not self.page_number:
last_page_number = DocumentPage.objects.filter(
document=self.document
).aggregate(Max('page_number'))['page_number__max']
if last_page_number is not None:
self.page_number = last_page_number + 1
else:
self.page_number = 1
super(DocumentPage, self).save(*args, **kwargs)
@property
def siblings(self):
return DocumentPage.objects.filter(
document_version=self.document_version
document=self.document
)
@property
@@ -271,12 +297,12 @@ class DocumentPage(models.Model):
Make cache UUID a mix of version ID and page ID to avoid using stale
images
"""
return '{}-{}'.format(self.document_version.uuid, self.pk)
return '{}-{}'.format(self.document.uuid, self.pk)
class DocumentPageResult(DocumentPage):
class Meta:
ordering = ('document_version__document', 'page_number')
ordering = ('document', 'page_number')
proxy = True
verbose_name = _('Document page')
verbose_name_plural = _('Document pages')
verbose_name = _('Document page result')
verbose_name_plural = _('Document pages result')