apps/catalogue/models.py

   1 # -*- coding: utf-8 -*-
   2 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
   3 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
   4 #
   5 from django.db import models
   6 from django.db.models import permalink, Q
   7 from django.utils.translation import ugettext_lazy as _
   8 from django.contrib.auth.models import User
   9 from django.core.files import File
  10 from django.template.loader import render_to_string
  11 from django.utils.safestring import mark_safe
  12 from django.utils.translation import get_language
  13 from django.core.urlresolvers import reverse
  14 from django.db.models.signals import post_save, m2m_changed, pre_delete
  15
  16 from django.conf import settings
  17
  18 from newtagging.models import TagBase, tags_updated
  19 from newtagging import managers
  20 from catalogue.fields import JSONField
  21
  22 from librarian import dcparser, html, epub, NoDublinCore
  23 from mutagen import id3
  24 from slughifi import slughifi
  25
  26
  27 TAG_CATEGORIES = (
  28     ('author', _('author')),
  29     ('epoch', _('epoch')),
  30     ('kind', _('kind')),
  31     ('genre', _('genre')),
  32     ('theme', _('theme')),
  33     ('set', _('set')),
  34     ('book', _('book')),
  35 )
  36
  37 MEDIA_FORMATS = (
  38     ('odt', _('ODT file')),
  39     ('mp3', _('MP3 file')),
  40     ('ogg', _('OGG file')),
  41     ('daisy', _('DAISY file')),
  42 )
  43
  44 class TagSubcategoryManager(models.Manager):
  45     def __init__(self, subcategory):
  46         super(TagSubcategoryManager, self).__init__()
  47         self.subcategory = subcategory
  48
  49     def get_query_set(self):
  50         return super(TagSubcategoryManager, self).get_query_set().filter(category=self.subcategory)
  51
  52
  53 class Tag(TagBase):
  54     name = models.CharField(_('name'), max_length=50, db_index=True)
  55     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
  56     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True)
  57     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
  58         db_index=True, choices=TAG_CATEGORIES)
  59     description = models.TextField(_('description'), blank=True)
  60     main_page = models.BooleanField(_('main page'), default=False, db_index=True, help_text=_('Show tag on main page'))
  61
  62     user = models.ForeignKey(User, blank=True, null=True)
  63     book_count = models.IntegerField(_('book count'), blank=True, null=True)
  64     gazeta_link = models.CharField(blank=True, max_length=240)
  65     wiki_link = models.CharField(blank=True, max_length=240)
  66
  67     categories_rev = {
  68         'autor': 'author',
  69         'epoka': 'epoch',
  70         'rodzaj': 'kind',
  71         'gatunek': 'genre',
  72         'motyw': 'theme',
  73         'polka': 'set',
  74     }
  75     categories_dict = dict((item[::-1] for item in categories_rev.iteritems()))
  76
  77     class Meta:
  78         ordering = ('sort_key',)
  79         verbose_name = _('tag')
  80         verbose_name_plural = _('tags')
  81         unique_together = (("slug", "category"),)
  82
  83     def __unicode__(self):
  84         return self.name
  85
  86     def __repr__(self):
  87         return "Tag(slug=%r)" % self.slug
  88
  89     @permalink
  90     def get_absolute_url(self):
  91         return ('catalogue.views.tagged_object_list', [self.url_chunk])
  92
  93     def has_description(self):
  94         return len(self.description) > 0
  95     has_description.short_description = _('description')
  96     has_description.boolean = True
  97
  98     def get_count(self):
  99         """ returns global book count for book tags, fragment count for themes """
 100
 101         if self.book_count is None:
 102             if self.category == 'book':
 103                 # never used
 104                 objects = Book.objects.none()
 105             elif self.category == 'theme':
 106                 objects = Fragment.tagged.with_all((self,))
 107             else:
 108                 objects = Book.tagged.with_all((self,)).order_by()
 109                 if self.category != 'set':
 110                     # eliminate descendants
 111                     l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects])
 112                     descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
 113                     if descendants_keys:
 114                         objects = objects.exclude(pk__in=descendants_keys)
 115             self.book_count = objects.count()
 116             self.save()
 117         return self.book_count
 118
 119     @staticmethod
 120     def get_tag_list(tags):
 121         if isinstance(tags, basestring):
 122             real_tags = []
 123             ambiguous_slugs = []
 124             category = None
 125             tags_splitted = tags.split('/')
 126             for index, name in enumerate(tags_splitted):
 127                 if name in Tag.categories_rev:
 128                     category = Tag.categories_rev[name]
 129                 else:
 130                     if category:
 131                         real_tags.append(Tag.objects.get(slug=name, category=category))
 132                         category = None
 133                     else:
 134                         try:
 135                             real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
 136                         except Tag.MultipleObjectsReturned, e:
 137                             ambiguous_slugs.append(name)
 138
 139             if category:
 140                 # something strange left off
 141                 raise Tag.DoesNotExist()
 142             if ambiguous_slugs:
 143                 # some tags should be qualified
 144                 e = Tag.MultipleObjectsReturned()
 145                 e.tags = real_tags
 146                 e.ambiguous_slugs = ambiguous_slugs
 147                 raise e
 148             else:
 149                 return real_tags
 150         else:
 151             return TagBase.get_tag_list(tags)
 152
 153     @property
 154     def url_chunk(self):
 155         return '/'.join((Tag.categories_dict[self.category], self.slug))
 156
 157
 158 # TODO: why is this hard-coded ?
 159 def book_upload_path(ext=None, maxlen=100):
 160     def get_dynamic_path(media, filename, ext=ext):
 161         # how to put related book's slug here?
 162         if not ext:
 163             if media.type == 'daisy':
 164                 ext = 'daisy.zip'
 165             else:
 166                 ext = media.type
 167         if not media.name:
 168             name = slughifi(filename.split(".")[0])
 169         else:
 170             name = slughifi(media.name)
 171         return 'lektura/%s.%s' % (name[:maxlen-len('lektura/.%s' % ext)-4], ext)
 172     return get_dynamic_path
 173
 174
 175 class BookMedia(models.Model):
 176     type        = models.CharField(_('type'), choices=MEDIA_FORMATS, max_length="100")
 177     name        = models.CharField(_('name'), max_length="100")
 178     file        = models.FileField(_('file'), upload_to=book_upload_path())
 179     uploaded_at = models.DateTimeField(_('creation date'), auto_now_add=True, editable=False)
 180     extra_info  = JSONField(_('extra information'), default='{}')
 181
 182     def book_count(self):
 183         return self.book_set.count()
 184     book_count.short_description = _('book count')
 185
 186     def books(self):
 187         return mark_safe('<br/>'.join("<a href='%s'>%s</a>" % (reverse('admin:catalogue_book_change', args=[b.id]), b.title) for b in self.book_set.all()))
 188     books.short_description = _('books')
 189
 190     def __unicode__(self):
 191         return "%s (%s)" % (self.name, self.file.name.split("/")[-1])
 192
 193     class Meta:
 194         ordering            = ('type', 'name')
 195         verbose_name        = _('book media')
 196         verbose_name_plural = _('book media')
 197
 198     def save(self, force_insert=False, force_update=False, **kwargs):
 199         media = super(BookMedia, self).save(force_insert, force_update, **kwargs)
 200         if self.type == 'mp3':
 201             file = self.file
 202             extra_info = self.get_extra_info_value()
 203             extra_info.update(self.get_mp3_info())
 204             self.set_extra_info_value(extra_info)
 205             media = super(BookMedia, self).save(force_insert, force_update, **kwargs)
 206         return media
 207
 208     def get_mp3_info(self):
 209         """Retrieves artist and director names from audio ID3 tags."""
 210         try:
 211             audio = id3.ID3(self.file.path)
 212             artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
 213             director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
 214         except:
 215             artist_name = director_name = ''
 216         return {'artist_name': artist_name, 'director_name': director_name}
 217
 218
 219 class Book(models.Model):
 220     title         = models.CharField(_('title'), max_length=120)
 221     slug          = models.SlugField(_('slug'), max_length=120, unique=True, db_index=True)
 222     description   = models.TextField(_('description'), blank=True)
 223     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True)
 224     _short_html   = models.TextField(_('short HTML'), editable=False)
 225     parent_number = models.IntegerField(_('parent number'), default=0)
 226     extra_info    = JSONField(_('extra information'))
 227     gazeta_link   = models.CharField(blank=True, max_length=240)
 228     wiki_link     = models.CharField(blank=True, max_length=240)
 229     # files generated during publication
 230     xml_file      = models.FileField(_('XML file'), upload_to=book_upload_path('xml'), blank=True)
 231     html_file     = models.FileField(_('HTML file'), upload_to=book_upload_path('html'), blank=True)
 232     pdf_file      = models.FileField(_('PDF file'), upload_to=book_upload_path('pdf'), blank=True)
 233     epub_file     = models.FileField(_('EPUB file'), upload_to=book_upload_path('epub'), blank=True)
 234     txt_file      = models.FileField(_('TXT file'), upload_to=book_upload_path('txt'), blank=True)
 235     # other files
 236     medias        = models.ManyToManyField(BookMedia, blank=True)
 237
 238     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
 239     objects  = models.Manager()
 240     tagged   = managers.ModelTaggedItemManager(Tag)
 241     tags     = managers.TagDescriptor(Tag)
 242
 243     _tag_counter = JSONField(null=True, editable=False)
 244     _theme_counter = JSONField(null=True, editable=False)
 245
 246     class AlreadyExists(Exception):
 247         pass
 248
 249     class Meta:
 250         ordering = ('title',)
 251         verbose_name = _('book')
 252         verbose_name_plural = _('books')
 253
 254     def __unicode__(self):
 255         return self.title
 256
 257     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
 258         if reset_short_html:
 259             # Reset _short_html during save
 260             update = {}
 261             for key in filter(lambda x: x.startswith('_short_html'), self.__dict__):
 262                 update[key] = ''
 263                 self.__setattr__(key, '')
 264             # Fragment.short_html relies on book's tags, so reset it here too
 265             self.fragments.all().update(**update)
 266
 267         return super(Book, self).save(force_insert, force_update)
 268
 269     @permalink
 270     def get_absolute_url(self):
 271         return ('catalogue.views.book_detail', [self.slug])
 272
 273     @property
 274     def name(self):
 275         return self.title
 276
 277     def book_tag_slug(self):
 278         return ('l-' + self.slug)[:120]
 279
 280     def book_tag(self):
 281         slug = self.book_tag_slug()
 282         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
 283         if created:
 284             book_tag.name = self.title[:50]
 285             book_tag.sort_key = self.title.lower()
 286             book_tag.save()
 287         return book_tag
 288
 289     def has_media(self, type):
 290         if   type == 'xml':
 291             if self.xml_file:
 292                 return True
 293             else:
 294                 return False
 295         elif type == 'html':
 296             if self.html_file:
 297                 return True
 298             else:
 299                 return False
 300         elif type == 'txt':
 301             if self.txt_file:
 302                 return True
 303             else:
 304                 return False
 305         elif type == 'pdf':
 306             if self.pdf_file:
 307                 return True
 308             else:
 309                 return False
 310         elif type == 'epub':
 311             if self.epub_file:
 312                 return True
 313             else:
 314                 return False
 315         else:
 316             if self.medias.filter(book=self, type=type).count() > 0:
 317                 return True
 318             else:
 319                 return False
 320
 321     def get_media(self, type):
 322         if self.has_media(type):
 323             if   type == "xml":
 324                 return self.xml_file
 325             elif type == "html":
 326                 return self.html_file
 327             elif type == "epub":
 328                 return self.epub_file
 329             elif type == "txt":
 330                 return self.txt_file
 331             elif type == "pdf":
 332                 return self.pdf_file
 333             else:
 334                 return self.medias.filter(book=self, type=type)
 335         else:
 336             return None
 337
 338     def get_mp3(self):
 339         return self.get_media("mp3")
 340     def get_odt(self):
 341         return self.get_media("odt")
 342     def get_ogg(self):
 343         return self.get_media("ogg")
 344     def get_daisy(self):
 345         return self.get_media("daisy")
 346
 347     def short_html(self):
 348         key = '_short_html_%s' % get_language()
 349         short_html = getattr(self, key)
 350
 351         if short_html and len(short_html):
 352             return mark_safe(short_html)
 353         else:
 354             tags = self.tags.filter(~Q(category__in=('set', 'theme', 'book')))
 355             tags = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name)) for tag in tags]
 356
 357             formats = []
 358             # files generated during publication
 359             if self.has_media("html"):
 360                 formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', kwargs={'slug': self.slug}), _('Read online')))
 361             if self.has_media("pdf"):
 362                 formats.append(u'<a href="%s">PDF</a>' % self.get_media('pdf').url)
 363             if self.root_ancestor.has_media("epub"):
 364                 formats.append(u'<a href="%s">EPUB</a>' % self.root_ancestor.get_media('epub').url)
 365             if self.has_media("txt"):
 366                 formats.append(u'<a href="%s">TXT</a>' % self.get_media('txt').url)
 367             # other files
 368             for m in self.medias.order_by('type'):
 369                 formats.append(u'<a href="%s">%s</a>' % (m.file.url, m.type.upper()))
 370
 371             formats = [mark_safe(format) for format in formats]
 372
 373             setattr(self, key, unicode(render_to_string('catalogue/book_short.html',
 374                 {'book': self, 'tags': tags, 'formats': formats})))
 375             self.save(reset_short_html=False)
 376             return mark_safe(getattr(self, key))
 377
 378
 379     @property
 380     def root_ancestor(self):
 381         """ returns the oldest ancestor """
 382
 383         if not hasattr(self, '_root_ancestor'):
 384             book = self
 385             while book.parent:
 386                 book = book.parent
 387             self._root_ancestor = book
 388         return self._root_ancestor
 389
 390
 391     def has_description(self):
 392         return len(self.description) > 0
 393     has_description.short_description = _('description')
 394     has_description.boolean = True
 395
 396     # ugly ugly ugly
 397     def has_pdf_file(self):
 398         return bool(self.pdf_file)
 399     has_pdf_file.short_description = 'PDF'
 400     has_pdf_file.boolean = True
 401
 402     def has_epub_file(self):
 403         return bool(self.epub_file)
 404     has_epub_file.short_description = 'EPUB'
 405     has_epub_file.boolean = True
 406
 407     def has_txt_file(self):
 408         return bool(self.txt_file)
 409     has_txt_file.short_description = 'HTML'
 410     has_txt_file.boolean = True
 411
 412     def has_html_file(self):
 413         return bool(self.html_file)
 414     has_html_file.short_description = 'HTML'
 415     has_html_file.boolean = True
 416
 417     def has_odt_file(self):
 418         return bool(self.has_media("odt"))
 419     has_odt_file.short_description = 'ODT'
 420     has_odt_file.boolean = True
 421
 422     def has_mp3_file(self):
 423         return bool(self.has_media("mp3"))
 424     has_mp3_file.short_description = 'MP3'
 425     has_mp3_file.boolean = True
 426
 427     def has_ogg_file(self):
 428         return bool(self.has_media("ogg"))
 429     has_ogg_file.short_description = 'OGG'
 430     has_ogg_file.boolean = True
 431
 432     def has_daisy_file(self):
 433         return bool(self.has_media("daisy"))
 434     has_daisy_file.short_description = 'DAISY'
 435     has_daisy_file.boolean = True
 436
 437     def build_epub(self, remove_descendants=True):
 438         """ (Re)builds the epub file.
 439             If book has a parent, does nothing.
 440             Unless remove_descendants is False, descendants' epubs are removed.
 441         """
 442
 443         from StringIO import StringIO
 444         from hashlib import sha1
 445         from django.core.files.base import ContentFile
 446         from librarian import DocProvider
 447
 448         class BookImportDocProvider(DocProvider):
 449             """ used for joined EPUBs """
 450
 451             def __init__(self, book):
 452                 self.book = book
 453
 454             def by_slug(self, slug):
 455                 if slug == self.book.slug:
 456                     return self.book.xml_file
 457                 else:
 458                     return Book.objects.get(slug=slug).xml_file
 459
 460         if self.parent:
 461             # don't need an epub
 462             return
 463
 464         epub_file = StringIO()
 465         try:
 466             epub.transform(BookImportDocProvider(self), self.slug, output_file=epub_file)
 467             self.epub_file.save('%s.epub' % self.slug, ContentFile(epub_file.getvalue()))
 468             FileRecord(slug=self.slug, type='epub', sha1=sha1(epub_file.getvalue()).hexdigest()).save()
 469         except NoDublinCore:
 470             pass
 471
 472         book_descendants = list(self.children.all())
 473         while len(book_descendants) > 0:
 474             child_book = book_descendants.pop(0)
 475             if remove_descendants and child_book.has_epub_file():
 476                 child_book.epub_file.delete()
 477             # save anyway, to refresh short_html
 478             child_book.save()
 479             book_descendants += list(child_book.children.all())
 480
 481     def build_txt(self):
 482         from StringIO import StringIO
 483         from django.core.files.base import ContentFile
 484         from librarian import text
 485
 486         out = StringIO()
 487         text.transform(open(self.xml_file.path), out)
 488         self.txt_file.save('%s.txt' % self.slug, ContentFile(out.getvalue()))
 489         self.save()
 490
 491
 492     @classmethod
 493     def from_xml_file(cls, xml_file, **kwargs):
 494         # use librarian to parse meta-data
 495         book_info = dcparser.parse(xml_file)
 496
 497         if not isinstance(xml_file, File):
 498             xml_file = File(open(xml_file))
 499
 500         try:
 501             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
 502         finally:
 503             xml_file.close()
 504
 505     @classmethod
 506     def from_text_and_meta(cls, raw_file, book_info, overwrite=False, build_epub=True, build_txt=True):
 507         import re
 508         from tempfile import NamedTemporaryFile
 509         from markupstring import MarkupString
 510         from django.core.files.storage import default_storage
 511
 512         # check for parts before we do anything
 513         children = []
 514         if hasattr(book_info, 'parts'):
 515             for part_url in book_info.parts:
 516                 base, slug = part_url.rsplit('/', 1)
 517                 try:
 518                     children.append(Book.objects.get(slug=slug))
 519                 except Book.DoesNotExist, e:
 520                     raise Book.DoesNotExist(_('Book with slug = "%s" does not exist.') % slug)
 521
 522
 523         # Read book metadata
 524         book_base, book_slug = book_info.url.rsplit('/', 1)
 525         if re.search(r'[^a-zA-Z0-9-]', book_slug):
 526             raise ValueError('Invalid characters in slug')
 527         book, created = Book.objects.get_or_create(slug=book_slug)
 528
 529         if created:
 530             book_shelves = []
 531         else:
 532             if not overwrite:
 533                 raise Book.AlreadyExists(_('Book %s already exists') % book_slug)
 534             # Save shelves for this book
 535             book_shelves = list(book.tags.filter(category='set'))
 536
 537         book.title = book_info.title
 538         book.set_extra_info_value(book_info.to_dict())
 539         book._short_html = ''
 540         book.save()
 541
 542         book_tags = []
 543         categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
 544         for field_name, category in categories:
 545             try:
 546                 tag_names = getattr(book_info, field_name)
 547             except:
 548                 tag_names = [getattr(book_info, category)]
 549             for tag_name in tag_names:
 550                 tag_sort_key = tag_name
 551                 if category == 'author':
 552                     tag_sort_key = tag_name.last_name
 553                     tag_name = ' '.join(tag_name.first_names) + ' ' + tag_name.last_name
 554                 tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
 555                 if created:
 556                     tag.name = tag_name
 557                     tag.sort_key = tag_sort_key.lower()
 558                     tag.save()
 559                 book_tags.append(tag)
 560
 561         book.tags = set(book_tags + book_shelves)
 562
 563         book_tag = book.book_tag()
 564
 565         for n, child_book in enumerate(children):
 566             child_book.parent = book
 567             child_book.parent_number = n
 568             child_book.save()
 569
 570         # Save XML and HTML files
 571         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
 572
 573         # delete old fragments when overwriting
 574         book.fragments.all().delete()
 575
 576         html_file = NamedTemporaryFile()
 577         if html.transform(book.xml_file.path, html_file, parse_dublincore=False):
 578             book.html_file.save('%s.html' % book.slug, File(html_file), save=False)
 579
 580             # get ancestor l-tags for adding to new fragments
 581             ancestor_tags = []
 582             p = book.parent
 583             while p:
 584                 ancestor_tags.append(p.book_tag())
 585                 p = p.parent
 586
 587             # Extract fragments
 588             closed_fragments, open_fragments = html.extract_fragments(book.html_file.path)
 589             for fragment in closed_fragments.values():
 590                 try:
 591                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 592                 except AttributeError:
 593                     continue
 594                 themes = []
 595                 for theme_name in theme_names:
 596                     if not theme_name:
 597                         continue
 598                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
 599                     if created:
 600                         tag.name = theme_name
 601                         tag.sort_key = theme_name.lower()
 602                         tag.save()
 603                     themes.append(tag)
 604                 if not themes:
 605                     continue
 606
 607                 text = fragment.to_string()
 608                 short_text = ''
 609                 if (len(MarkupString(text)) > 240):
 610                     short_text = unicode(MarkupString(text)[:160])
 611                 new_fragment, created = Fragment.objects.get_or_create(anchor=fragment.id, book=book,
 612                     defaults={'text': text, 'short_text': short_text})
 613
 614                 new_fragment.save()
 615                 new_fragment.tags = set(book_tags + themes + [book_tag] + ancestor_tags)
 616
 617         if not settings.NO_BUILD_TXT and build_txt:
 618             book.build_txt()
 619
 620         if not settings.NO_BUILD_EPUB and build_epub:
 621             book.root_ancestor.build_epub()
 622
 623         book_descendants = list(book.children.all())
 624         # add l-tag to descendants and their fragments
 625         # delete unnecessary EPUB files
 626         while len(book_descendants) > 0:
 627             child_book = book_descendants.pop(0)
 628             child_book.tags = list(child_book.tags) + [book_tag]
 629             child_book.save()
 630             for fragment in child_book.fragments.all():
 631                 fragment.tags = set(list(fragment.tags) + [book_tag])
 632             book_descendants += list(child_book.children.all())
 633
 634         # refresh cache
 635         book.reset_tag_counter()
 636         book.reset_theme_counter()
 637
 638         book.save()
 639         return book
 640
 641
 642     def refresh_tag_counter(self):
 643         tags = {}
 644         for child in self.children.all().order_by():
 645             for tag_pk, value in child.tag_counter.iteritems():
 646                 tags[tag_pk] = tags.get(tag_pk, 0) + value
 647         for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
 648             tags[tag.pk] = 1
 649         self.set__tag_counter_value(tags)
 650         self.save(reset_short_html=False)
 651         return tags
 652
 653     def reset_tag_counter(self):
 654         self._tag_counter = None
 655         self.save(reset_short_html=False)
 656         if self.parent:
 657             self.parent.reset_tag_counter()
 658
 659     @property
 660     def tag_counter(self):
 661         if self._tag_counter is None:
 662             return self.refresh_tag_counter()
 663         return dict((int(k), v) for k, v in self.get__tag_counter_value().iteritems())
 664
 665     def refresh_theme_counter(self):
 666         tags = {}
 667         for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
 668             for tag in fragment.tags.filter(category='theme').order_by():
 669                 tags[tag.pk] = tags.get(tag.pk, 0) + 1
 670         self.set__theme_counter_value(tags)
 671         self.save(reset_short_html=False)
 672         return tags
 673
 674     def reset_theme_counter(self):
 675         self._theme_counter = None
 676         self.save(reset_short_html=False)
 677         if self.parent:
 678             self.parent.reset_theme_counter()
 679
 680     @property
 681     def theme_counter(self):
 682         if self._theme_counter is None:
 683             return self.refresh_theme_counter()
 684         return dict((int(k), v) for k, v in self.get__theme_counter_value().iteritems())
 685
 686     def pretty_title(self, html_links=False):
 687         book = self
 688         names = list(book.tags.filter(category='author'))
 689
 690         books = []
 691         while book:
 692             books.append(book)
 693             book = book.parent
 694         names.extend(reversed(books))
 695
 696         if html_links:
 697             names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
 698         else:
 699             names = [tag.name for tag in names]
 700
 701         return ', '.join(names)
 702
 703
 704 class Fragment(models.Model):
 705     text = models.TextField()
 706     short_text = models.TextField(editable=False)
 707     _short_html = models.TextField(editable=False)
 708     anchor = models.CharField(max_length=120)
 709     book = models.ForeignKey(Book, related_name='fragments')
 710
 711     objects = models.Manager()
 712     tagged = managers.ModelTaggedItemManager(Tag)
 713     tags = managers.TagDescriptor(Tag)
 714
 715     class Meta:
 716         ordering = ('book', 'anchor',)
 717         verbose_name = _('fragment')
 718         verbose_name_plural = _('fragments')
 719
 720     def get_absolute_url(self):
 721         return '%s#m%s' % (reverse('book_text', kwargs={'slug': self.book.slug}), self.anchor)
 722
 723     def short_html(self):
 724         key = '_short_html_%s' % get_language()
 725         short_html = getattr(self, key)
 726         if short_html and len(short_html):
 727             return mark_safe(short_html)
 728         else:
 729             setattr(self, key, unicode(render_to_string('catalogue/fragment_short.html',
 730                 {'fragment': self})))
 731             self.save()
 732             return mark_safe(getattr(self, key))
 733
 734
 735 class FileRecord(models.Model):
 736     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
 737     type = models.CharField(_('type'), max_length=20, db_index=True)
 738     sha1 = models.CharField(_('sha-1 hash'), max_length=40)
 739     time = models.DateTimeField(_('time'), auto_now_add=True)
 740
 741     class Meta:
 742         ordering = ('-time','-slug', '-type')
 743         verbose_name = _('file record')
 744         verbose_name_plural = _('file records')
 745
 746     def __unicode__(self):
 747         return "%s %s.%s" % (self.sha1,  self.slug, self.type)
 748
 749 ###########
 750 #
 751 # SIGNALS
 752 #
 753 ###########
 754
 755
 756 def _tags_updated_handler(sender, affected_tags, **kwargs):
 757     # reset tag global counter
 758     Tag.objects.filter(pk__in=[tag.pk for tag in affected_tags]).update(book_count=None)
 759
 760     # if book tags changed, reset book tag counter
 761     if isinstance(sender, Book) and \
 762                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
 763                     exclude(category__in=('book', 'theme', 'set')).count():
 764         sender.reset_tag_counter()
 765     # if fragment theme changed, reset book theme counter
 766     elif isinstance(sender, Fragment) and \
 767                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
 768                     filter(category='theme').count():
 769         sender.book.reset_theme_counter()
 770 tags_updated.connect(_tags_updated_handler)
 771
 772
 773 def _m2m_changed_handler(sender, instance, action, reverse, pk_set, **kwargs):
 774     """ refresh all the short_html stuff on BookMedia delete """
 775     if sender == Book.medias.through and reverse and action == 'pre_clear':
 776         for book in instance.book_set.all():
 777             book.save()
 778 m2m_changed.connect(_m2m_changed_handler)
 779
 780 def _pre_delete_handler(sender, instance, **kwargs):
 781     """ explicitly clear m2m, so that Books can be refreshed """
 782     if sender == BookMedia:
 783         instance.book_set.clear()
 784 pre_delete.connect(_pre_delete_handler)
 785
 786 def _post_save_handler(sender, instance, **kwargs):
 787     """ refresh all the short_html stuff on BookMedia update """
 788     if sender == BookMedia:
 789         for book in instance.book_set.all():
 790             book.save()
 791 post_save.connect(_post_save_handler)
 792