src/picture/models.py

   1 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
   2 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
   3 #
   4 from django.db import models, transaction
   5 import catalogue.models
   6 from sorl.thumbnail import ImageField
   7 from django.conf import settings
   8 from django.contrib.contenttypes.fields import GenericRelation
   9 from django.core.files.storage import FileSystemStorage
  10 from django.urls import reverse
  11 from slugify import slugify
  12 from ssify import flush_ssi_includes
  13
  14 from catalogue.models.tag import prefetched_relations
  15 from catalogue.utils import split_tags
  16 from picture import tasks
  17 from io import BytesIO
  18 import jsonfield
  19 import itertools
  20 import logging
  21 import re
  22
  23 from PIL import Image
  24
  25 from django.utils.translation import ugettext_lazy as _
  26 from newtagging import managers
  27 from os import path
  28
  29
  30 picture_storage = FileSystemStorage(location=path.join(
  31         settings.MEDIA_ROOT, 'pictures'),
  32         base_url=settings.MEDIA_URL + "pictures/")
  33
  34
  35 class PictureArea(models.Model):
  36     picture = models.ForeignKey('picture.Picture', models.CASCADE, related_name='areas')
  37     area = jsonfield.JSONField(_('area'), default={}, editable=False)
  38     kind = models.CharField(
  39         _('kind'), max_length=10, blank=False, null=False, db_index=True,
  40         choices=(('thing', _('thing')), ('theme', _('theme'))))
  41
  42     objects = models.Manager()
  43     tagged = managers.ModelTaggedItemManager(catalogue.models.Tag)
  44     tags = managers.TagDescriptor(catalogue.models.Tag)
  45     tag_relations = GenericRelation(catalogue.models.Tag.intermediary_table_model)
  46
  47     short_html_url_name = 'picture_area_short'
  48
  49     @classmethod
  50     def rectangle(cls, picture, kind, coords):
  51         pa = PictureArea()
  52         pa.picture = picture
  53         pa.kind = kind
  54         pa.area = coords
  55         return pa
  56
  57     def flush_includes(self, languages=True):
  58         if not languages:
  59             return
  60         if languages is True:
  61             languages = [lc for (lc, _ln) in settings.LANGUAGES]
  62         flush_ssi_includes([
  63             template % (self.pk, lang)
  64             for template in [
  65                 '/katalog/pa/%d/short.%s.html',
  66                 ]
  67             for lang in languages
  68             ])
  69
  70
  71 class Picture(models.Model):
  72     """
  73     Picture resource.
  74
  75     """
  76     title = models.CharField(_('title'), max_length=32767)
  77     slug = models.SlugField(_('slug'), max_length=120, db_index=True, unique=True)
  78     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True, editable=False)
  79     sort_key_author = models.CharField(
  80         _('sort key by author'), max_length=120, db_index=True, editable=False, default=u'')
  81     created_at = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
  82     changed_at = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
  83     xml_file = models.FileField(_('xml file'), upload_to="xml", storage=picture_storage)
  84     image_file = ImageField(_('image file'), upload_to="images", storage=picture_storage)
  85     html_file = models.FileField(_('html file'), upload_to="html", storage=picture_storage)
  86     areas_json = jsonfield.JSONField(_('picture areas JSON'), default={}, editable=False)
  87     extra_info = jsonfield.JSONField(_('extra information'), default={})
  88     culturepl_link = models.CharField(blank=True, max_length=240)
  89     wiki_link = models.CharField(blank=True, max_length=240)
  90
  91     width = models.IntegerField(null=True)
  92     height = models.IntegerField(null=True)
  93
  94     objects = models.Manager()
  95     tagged = managers.ModelTaggedItemManager(catalogue.models.Tag)
  96     tags = managers.TagDescriptor(catalogue.models.Tag)
  97     tag_relations = GenericRelation(catalogue.models.Tag.intermediary_table_model)
  98
  99     short_html_url_name = 'picture_short'
 100
 101     class AlreadyExists(Exception):
 102         pass
 103
 104     class Meta:
 105         ordering = ('sort_key_author', 'sort_key')
 106
 107         verbose_name = _('picture')
 108         verbose_name_plural = _('pictures')
 109
 110     def save(self, force_insert=False, force_update=False, **kwargs):
 111         from sortify import sortify
 112
 113         self.sort_key = sortify(self.title)[:120]
 114
 115         try:
 116             author = self.authors().first().sort_key
 117         except AttributeError:
 118             author = u''
 119         self.sort_key_author = author
 120
 121         ret = super(Picture, self).save(force_insert, force_update)
 122
 123         return ret
 124
 125     def __str__(self):
 126         return self.title
 127
 128     def authors(self):
 129         return self.tags.filter(category='author')
 130
 131     def tag_unicode(self, category):
 132         relations = prefetched_relations(self, category)
 133         if relations:
 134             return ', '.join(rel.tag.name for rel in relations)
 135         else:
 136             return ', '.join(self.tags.filter(category=category).values_list('name', flat=True))
 137
 138     def author_unicode(self):
 139         return self.tag_unicode('author')
 140
 141     def tags_by_category(self):
 142         return split_tags(self.tags)
 143
 144     def get_absolute_url(self):
 145         return reverse('picture_detail', args=[self.slug])
 146
 147     def get_initial(self):
 148         try:
 149             return re.search(r'\w', self.title, re.U).group(0)
 150         except AttributeError:
 151             return ''
 152
 153     def get_next(self):
 154         try:
 155             return type(self).objects.filter(sort_key__gt=self.sort_key)[0]
 156         except IndexError:
 157             return None
 158
 159     def get_previous(self):
 160         try:
 161             return type(self).objects.filter(sort_key__lt=self.sort_key).order_by('-sort_key')[0]
 162         except IndexError:
 163             return None
 164
 165     @classmethod
 166     def from_xml_file(cls, xml_file, image_file=None, image_store=None, overwrite=False, search_index=True):
 167         """
 168         Import xml and it's accompanying image file.
 169         If image file is missing, it will be fetched by librarian.picture.ImageStore
 170         which looks for an image file in the same directory the xml is, with extension matching
 171         its mime type.
 172         """
 173         from sortify import sortify
 174         from django.core.files import File
 175         from librarian.picture import WLPicture, ImageStore
 176         close_xml_file = False
 177         close_image_file = False
 178
 179         if image_file is not None and not isinstance(image_file, File):
 180             image_file = File(open(image_file, 'rb'))
 181             close_image_file = True
 182
 183         if not isinstance(xml_file, File):
 184             xml_file = File(open(xml_file))
 185             close_xml_file = True
 186
 187         with transaction.atomic():
 188             # use librarian to parse meta-data
 189             if image_store is None:
 190                 image_store = ImageStore(picture_storage.path('images'))
 191             picture_xml = WLPicture.from_file(xml_file, image_store=image_store)
 192
 193             picture, created = Picture.objects.get_or_create(slug=picture_xml.slug[:120])
 194             if not created and not overwrite:
 195                 raise Picture.AlreadyExists('Picture %s already exists' % picture_xml.slug)
 196
 197             picture.areas.all().delete()
 198             picture.title = str(picture_xml.picture_info.title)
 199             picture.extra_info = picture_xml.picture_info.to_dict()
 200
 201             picture_tags = set(catalogue.models.Tag.tags_from_info(picture_xml.picture_info))
 202             for tag in picture_tags:
 203                 if not tag.for_pictures:
 204                     tag.for_pictures = True
 205                     tag.save()
 206
 207             area_data = {'themes': {}, 'things': {}}
 208
 209             # Treat all names in picture XML as in default language.
 210             lang = settings.LANGUAGE_CODE
 211
 212             for part in picture_xml.partiter():
 213                 if picture_xml.frame:
 214                     c = picture_xml.frame[0]
 215                     part['coords'] = [[p[0] - c[0], p[1] - c[1]] for p in part['coords']]
 216                 if part.get('object', None) is not None:
 217                     _tags = set()
 218                     for objname in part['object'].split(','):
 219                         objname = objname.strip()
 220                         assert objname, 'Empty object name'
 221                         # str.capitalize() is wrong, because it also lowers letters
 222                         objname = objname[0].upper() + objname[1:]
 223                         tag, created = catalogue.models.Tag.objects.get_or_create(
 224                             slug=slugify(objname), category='thing')
 225                         if created:
 226                             tag.name = objname
 227                             setattr(tag, 'name_%s' % lang, tag.name)
 228                             tag.sort_key = sortify(tag.name)
 229                             tag.for_pictures = True
 230                             tag.save()
 231                         area_data['things'][tag.slug] = {
 232                             'object': objname,
 233                             'coords': part['coords'],
 234                             }
 235
 236                         _tags.add(tag)
 237                         if not tag.for_pictures:
 238                             tag.for_pictures = True
 239                             tag.save()
 240                     area = PictureArea.rectangle(picture, 'thing', part['coords'])
 241                     area.save()
 242                     # WTF thing area does not inherit tags from picture and theme area does, is it intentional?
 243                     area.tags = _tags
 244                 else:
 245                     _tags = set()
 246                     for motifs in part['themes']:
 247                         for motif in motifs.split(','):
 248                             tag, created = catalogue.models.Tag.objects.get_or_create(
 249                                 slug=slugify(motif), category='theme')
 250                             if created:
 251                                 tag.name = motif
 252                                 tag.sort_key = sortify(tag.name)
 253                                 tag.for_pictures = True
 254                                 tag.save()
 255                             # motif_tags.add(tag)
 256                             _tags.add(tag)
 257                             if not tag.for_pictures:
 258                                 tag.for_pictures = True
 259                                 tag.save()
 260                             area_data['themes'][tag.slug] = {
 261                                 'theme': motif,
 262                                 'coords': part['coords']
 263                                 }
 264
 265                     logging.debug("coords for theme: %s" % part['coords'])
 266                     area = PictureArea.rectangle(picture, 'theme', part['coords'])
 267                     area.save()
 268                     area.tags = _tags.union(picture_tags)
 269
 270             picture.tags = picture_tags
 271             picture.areas_json = area_data
 272
 273             if image_file is not None:
 274                 img = image_file
 275             else:
 276                 img = picture_xml.image_file()
 277
 278             modified = cls.crop_to_frame(picture_xml, img)
 279             modified = cls.add_source_note(picture_xml, modified)
 280
 281             picture.width, picture.height = modified.size
 282
 283             modified_file = BytesIO()
 284             modified.save(modified_file, format='JPEG', quality=95)
 285             # FIXME: hardcoded extension - detect from DC format or orginal filename
 286             picture.image_file.save(path.basename(picture_xml.image_path), File(modified_file))
 287
 288             picture.xml_file.save("%s.xml" % picture.slug, File(xml_file))
 289             picture.save()
 290             tasks.generate_picture_html(picture.id)
 291             if not settings.NO_SEARCH_INDEX and search_index:
 292                 tasks.index_picture.delay(picture.id, picture_info=picture_xml.picture_info)
 293
 294         if close_xml_file:
 295             xml_file.close()
 296         if close_image_file:
 297             image_file.close()
 298
 299         return picture
 300
 301     @classmethod
 302     def crop_to_frame(cls, wlpic, image_file):
 303         img = Image.open(image_file)
 304         if wlpic.frame is None or wlpic.frame == [[0, 0], [-1, -1]]:
 305             return img
 306         img = img.crop(itertools.chain(*wlpic.frame))
 307         return img
 308
 309     @staticmethod
 310     def add_source_note(wlpic, img):
 311         from PIL import ImageDraw, ImageFont
 312         from librarian import get_resource
 313
 314         annotated = Image.new(img.mode, (img.size[0], img.size[1] + 40), (255, 255, 255))
 315         annotated.paste(img, (0, 0))
 316         annotation = Image.new('RGB', (img.size[0] * 3, 120), (255, 255, 255))
 317         ImageDraw.Draw(annotation).text(
 318             (30, 15),
 319             wlpic.picture_info.source_name,
 320             (0, 0, 0),
 321             font=ImageFont.truetype(get_resource("fonts/DejaVuSerif.ttf"), 75)
 322         )
 323         annotated.paste(annotation.resize((img.size[0], 40), Image.ANTIALIAS), (0, img.size[1]))
 324         return annotated
 325
 326     @property
 327     def info(self):
 328         if not hasattr(self, '_info'):
 329             from librarian import dcparser
 330             from librarian import picture
 331             info = dcparser.parse(self.xml_file.path, picture.PictureInfo)
 332             self._info = info
 333         return self._info
 334
 335     def pretty_title(self, html_links=False):
 336         names = [(tag.name, tag.get_absolute_url()) for tag in self.authors().only('name', 'category', 'slug')]
 337         names.append((self.title, self.get_absolute_url()))
 338
 339         if html_links:
 340             names = ['<a href="%s">%s</a>' % (tag[1], tag[0]) for tag in names]
 341         else:
 342             names = [tag[0] for tag in names]
 343         return ', '.join(names)
 344
 345     def related_themes(self):
 346         return catalogue.models.Tag.objects.usage_for_queryset(
 347             self.areas.all(), counts=True).filter(category__in=('theme', 'thing'))
 348
 349     def flush_includes(self, languages=True):
 350         if not languages:
 351             return
 352         if languages is True:
 353             languages = [lc for (lc, _ln) in settings.LANGUAGES]
 354         flush_ssi_includes([
 355             template % (self.pk, lang)
 356             for template in [
 357                 '/katalog/p/%d/short.%s.html',
 358                 '/katalog/p/%d/mini.%s.html',
 359                 ]
 360             for lang in languages
 361             ])
 362
 363     def search_index(self, picture_info=None, index=None, index_tags=True, commit=True):
 364         if index is None:
 365             from search.index import Index
 366             index = Index()
 367         try:
 368             index.index_picture(self, picture_info)
 369             if index_tags:
 370                 index.index_tags()
 371             if commit:
 372                 index.index.commit()
 373         except Exception as e:
 374             index.index.rollback()
 375             raise e