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 collections import namedtuple
   6
   7 from django.db import models
   8 from django.db.models import permalink, Q
   9 import django.dispatch
  10 from django.core.cache import cache
  11 from django.core.files.storage import DefaultStorage
  12 from django.utils.translation import ugettext_lazy as _
  13 from django.contrib.auth.models import User
  14 from django.template.loader import render_to_string
  15 from django.utils.datastructures import SortedDict
  16 from django.utils.safestring import mark_safe
  17 from django.utils.translation import get_language
  18 from django.core.urlresolvers import reverse
  19 from django.db.models.signals import post_save, m2m_changed, pre_delete
  20
  21 from django.conf import settings
  22
  23 from newtagging.models import TagBase, tags_updated
  24 from newtagging import managers
  25 from catalogue.fields import JSONField, OverwritingFileField
  26 from catalogue.utils import create_zip
  27 from catalogue.tasks import touch_tag
  28 from shutil import copy
  29 from glob import glob
  30 import re
  31 from os import path
  32
  33
  34 TAG_CATEGORIES = (
  35     ('author', _('author')),
  36     ('epoch', _('epoch')),
  37     ('kind', _('kind')),
  38     ('genre', _('genre')),
  39     ('theme', _('theme')),
  40     ('set', _('set')),
  41     ('book', _('book')),
  42 )
  43
  44 # not quite, but Django wants you to set a timeout
  45 CACHE_FOREVER = 2419200  # 28 days
  46
  47
  48 class TagSubcategoryManager(models.Manager):
  49     def __init__(self, subcategory):
  50         super(TagSubcategoryManager, self).__init__()
  51         self.subcategory = subcategory
  52
  53     def get_query_set(self):
  54         return super(TagSubcategoryManager, self).get_query_set().filter(category=self.subcategory)
  55
  56
  57 class Tag(TagBase):
  58     name = models.CharField(_('name'), max_length=50, db_index=True)
  59     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
  60     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True)
  61     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
  62         db_index=True, choices=TAG_CATEGORIES)
  63     description = models.TextField(_('description'), blank=True)
  64     main_page = models.BooleanField(_('main page'), default=False, db_index=True, help_text=_('Show tag on main page'))
  65
  66     user = models.ForeignKey(User, blank=True, null=True)
  67     book_count = models.IntegerField(_('book count'), blank=True, null=True)
  68     gazeta_link = models.CharField(blank=True, max_length=240)
  69     wiki_link = models.CharField(blank=True, max_length=240)
  70
  71     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
  72     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
  73
  74     class UrlDeprecationWarning(DeprecationWarning):
  75         pass
  76
  77     categories_rev = {
  78         'autor': 'author',
  79         'epoka': 'epoch',
  80         'rodzaj': 'kind',
  81         'gatunek': 'genre',
  82         'motyw': 'theme',
  83         'polka': 'set',
  84     }
  85     categories_dict = dict((item[::-1] for item in categories_rev.iteritems()))
  86
  87     class Meta:
  88         ordering = ('sort_key',)
  89         verbose_name = _('tag')
  90         verbose_name_plural = _('tags')
  91         unique_together = (("slug", "category"),)
  92
  93     def __unicode__(self):
  94         return self.name
  95
  96     def __repr__(self):
  97         return "Tag(slug=%r)" % self.slug
  98
  99     @permalink
 100     def get_absolute_url(self):
 101         return ('catalogue.views.tagged_object_list', [self.url_chunk])
 102
 103     def has_description(self):
 104         return len(self.description) > 0
 105     has_description.short_description = _('description')
 106     has_description.boolean = True
 107
 108     def get_count(self):
 109         """Returns global book count for book tags, fragment count for themes."""
 110
 111         if self.category == 'book':
 112             # never used
 113             objects = Book.objects.none()
 114         elif self.category == 'theme':
 115             objects = Fragment.tagged.with_all((self,))
 116         else:
 117             objects = Book.tagged.with_all((self,)).order_by()
 118             if self.category != 'set':
 119                 # eliminate descendants
 120                 l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects])
 121                 descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
 122                 if descendants_keys:
 123                     objects = objects.exclude(pk__in=descendants_keys)
 124         return objects.count()
 125
 126     @staticmethod
 127     def get_tag_list(tags):
 128         if isinstance(tags, basestring):
 129             real_tags = []
 130             ambiguous_slugs = []
 131             category = None
 132             deprecated = False
 133             tags_splitted = tags.split('/')
 134             for name in tags_splitted:
 135                 if category:
 136                     real_tags.append(Tag.objects.get(slug=name, category=category))
 137                     category = None
 138                 elif name in Tag.categories_rev:
 139                     category = Tag.categories_rev[name]
 140                 else:
 141                     try:
 142                         real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
 143                         deprecated = True
 144                     except Tag.MultipleObjectsReturned, e:
 145                         ambiguous_slugs.append(name)
 146
 147             if category:
 148                 # something strange left off
 149                 raise Tag.DoesNotExist()
 150             if ambiguous_slugs:
 151                 # some tags should be qualified
 152                 e = Tag.MultipleObjectsReturned()
 153                 e.tags = real_tags
 154                 e.ambiguous_slugs = ambiguous_slugs
 155                 raise e
 156             if deprecated:
 157                 e = Tag.UrlDeprecationWarning()
 158                 e.tags = real_tags
 159                 raise e
 160             return real_tags
 161         else:
 162             return TagBase.get_tag_list(tags)
 163
 164     @property
 165     def url_chunk(self):
 166         return '/'.join((Tag.categories_dict[self.category], self.slug))
 167
 168     @staticmethod
 169     def tags_from_info(info):
 170         from slughifi import slughifi
 171         from sortify import sortify
 172         meta_tags = []
 173         categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
 174         for field_name, category in categories:
 175             try:
 176                 tag_names = getattr(info, field_name)
 177             except:
 178                 tag_names = [getattr(info, category)]
 179             for tag_name in tag_names:
 180                 tag_sort_key = tag_name
 181                 if category == 'author':
 182                     tag_sort_key = tag_name.last_name
 183                     tag_name = ' '.join(tag_name.first_names) + ' ' + tag_name.last_name
 184                 tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
 185                 if created:
 186                     tag.name = tag_name
 187                     tag.sort_key = sortify(tag_sort_key.lower())
 188                     tag.save()
 189                 meta_tags.append(tag)
 190         return meta_tags
 191
 192
 193
 194 def get_dynamic_path(media, filename, ext=None, maxlen=100):
 195     from slughifi import slughifi
 196
 197     # how to put related book's slug here?
 198     if not ext:
 199         # BookMedia case
 200         ext = media.formats[media.type].ext
 201     if media is None or not media.name:
 202         name = slughifi(filename.split(".")[0])
 203     else:
 204         name = slughifi(media.name)
 205     return 'book/%s/%s.%s' % (ext, name[:maxlen-len('book/%s/.%s' % (ext, ext))-4], ext)
 206
 207
 208 # TODO: why is this hard-coded ?
 209 def book_upload_path(ext=None, maxlen=100):
 210     return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen)
 211
 212
 213 def get_customized_pdf_path(book, customizations):
 214     """
 215     Returns a MEDIA_ROOT relative path for a customized pdf. The name will contain a hash of customization options.
 216     """
 217     customizations.sort()
 218     h = hash(tuple(customizations))
 219
 220     pdf_name = '%s-custom-%s' % (book.fileid(), h)
 221     pdf_file = get_dynamic_path(None, pdf_name, ext='pdf')
 222
 223     return pdf_file
 224
 225
 226 def get_existing_customized_pdf(book):
 227     """
 228     Returns a list of paths to generated customized pdf of a book
 229     """
 230     pdf_glob = '%s-custom-' % (book.fileid(),)
 231     pdf_glob = get_dynamic_path(None, pdf_glob, ext='pdf')
 232     pdf_glob = re.sub(r"[.]([a-z0-9]+)$", "*.\\1", pdf_glob)
 233     return glob(path.join(settings.MEDIA_ROOT, pdf_glob))
 234
 235
 236 class BookMedia(models.Model):
 237     FileFormat = namedtuple("FileFormat", "name ext")
 238     formats = SortedDict([
 239         ('mp3', FileFormat(name='MP3', ext='mp3')),
 240         ('ogg', FileFormat(name='Ogg Vorbis', ext='ogg')),
 241         ('daisy', FileFormat(name='DAISY', ext='daisy.zip')),
 242     ])
 243     format_choices = [(k, _('%s file') % t.name)
 244             for k, t in formats.items()]
 245
 246     type        = models.CharField(_('type'), choices=format_choices, max_length="100")
 247     name        = models.CharField(_('name'), max_length="100")
 248     file        = OverwritingFileField(_('file'), upload_to=book_upload_path())
 249     uploaded_at = models.DateTimeField(_('creation date'), auto_now_add=True, editable=False)
 250     extra_info  = JSONField(_('extra information'), default='{}', editable=False)
 251     book = models.ForeignKey('Book', related_name='media')
 252     source_sha1 = models.CharField(null=True, blank=True, max_length=40, editable=False)
 253
 254     def __unicode__(self):
 255         return "%s (%s)" % (self.name, self.file.name.split("/")[-1])
 256
 257     class Meta:
 258         ordering            = ('type', 'name')
 259         verbose_name        = _('book media')
 260         verbose_name_plural = _('book media')
 261
 262     def save(self, *args, **kwargs):
 263         from slughifi import slughifi
 264         from catalogue.utils import ExistingFile, remove_zip
 265
 266         try:
 267             old = BookMedia.objects.get(pk=self.pk)
 268         except BookMedia.DoesNotExist, e:
 269             pass
 270         else:
 271             # if name changed, change the file name, too
 272             if slughifi(self.name) != slughifi(old.name):
 273                 self.file.save(None, ExistingFile(self.file.path), save=False, leave=True)
 274
 275         super(BookMedia, self).save(*args, **kwargs)
 276
 277         # remove the zip package for book with modified media
 278         remove_zip(self.book.fileid())
 279
 280         extra_info = self.get_extra_info_value()
 281         extra_info.update(self.read_meta())
 282         self.set_extra_info_value(extra_info)
 283         self.source_sha1 = self.read_source_sha1(self.file.path, self.type)
 284         return super(BookMedia, self).save(*args, **kwargs)
 285
 286     def read_meta(self):
 287         """
 288             Reads some metadata from the audiobook.
 289         """
 290         import mutagen
 291         from mutagen import id3
 292
 293         artist_name = director_name = project = funded_by = ''
 294         if self.type == 'mp3':
 295             try:
 296                 audio = id3.ID3(self.file.path)
 297                 artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
 298                 director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
 299                 project = ", ".join([t.data for t in audio.getall('PRIV')
 300                         if t.owner=='wolnelektury.pl?project'])
 301                 funded_by = ", ".join([t.data for t in audio.getall('PRIV')
 302                         if t.owner=='wolnelektury.pl?funded_by'])
 303             except:
 304                 pass
 305         elif self.type == 'ogg':
 306             try:
 307                 audio = mutagen.File(self.file.path)
 308                 artist_name = ', '.join(audio.get('artist', []))
 309                 director_name = ', '.join(audio.get('conductor', []))
 310                 project = ", ".join(audio.get('project', []))
 311                 funded_by = ", ".join(audio.get('funded_by', []))
 312             except:
 313                 pass
 314         else:
 315             return {}
 316         return {'artist_name': artist_name, 'director_name': director_name,
 317                 'project': project, 'funded_by': funded_by}
 318
 319     @staticmethod
 320     def read_source_sha1(filepath, filetype):
 321         """
 322             Reads source file SHA1 from audiobok metadata.
 323         """
 324         import mutagen
 325         from mutagen import id3
 326
 327         if filetype == 'mp3':
 328             try:
 329                 audio = id3.ID3(filepath)
 330                 return [t.data for t in audio.getall('PRIV')
 331                         if t.owner=='wolnelektury.pl?flac_sha1'][0]
 332             except:
 333                 return None
 334         elif filetype == 'ogg':
 335             try:
 336                 audio = mutagen.File(filepath)
 337                 return audio.get('flac_sha1', [None])[0]
 338             except:
 339                 return None
 340         else:
 341             return None
 342
 343
 344 class Book(models.Model):
 345     title         = models.CharField(_('title'), max_length=120)
 346     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True, editable=False)
 347     slug          = models.SlugField(_('slug'), max_length=120, db_index=True)
 348     language = models.CharField(_('language code'), max_length=3, db_index=True,
 349                     default=settings.CATALOGUE_DEFAULT_LANGUAGE)
 350     description   = models.TextField(_('description'), blank=True)
 351     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
 352     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
 353     parent_number = models.IntegerField(_('parent number'), default=0)
 354     extra_info    = JSONField(_('extra information'), default='{}')
 355     gazeta_link   = models.CharField(blank=True, max_length=240)
 356     wiki_link     = models.CharField(blank=True, max_length=240)
 357     # files generated during publication
 358
 359     ebook_formats = ['pdf', 'epub', 'mobi', 'txt']
 360     formats = ebook_formats + ['html', 'xml']
 361
 362     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
 363     objects  = models.Manager()
 364     tagged   = managers.ModelTaggedItemManager(Tag)
 365     tags     = managers.TagDescriptor(Tag)
 366
 367     html_built = django.dispatch.Signal()
 368     published = django.dispatch.Signal()
 369
 370     URLID_RE = r'[a-z0-9-]+(?:/[a-z]{3})?'
 371     FILEID_RE = r'[a-z0-9-]+(?:_[a-z]{3})?'
 372
 373     class AlreadyExists(Exception):
 374         pass
 375
 376     class Meta:
 377         unique_together = [['slug', 'language']]
 378         ordering = ('sort_key',)
 379         verbose_name = _('book')
 380         verbose_name_plural = _('books')
 381
 382     def __unicode__(self):
 383         return self.title
 384
 385     def urlid(self, sep='/'):
 386         stem = self.slug
 387         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 388             stem += sep + self.language
 389         return stem
 390
 391     def fileid(self):
 392         return self.urlid('_')
 393
 394     @staticmethod
 395     def split_urlid(urlid, sep='/', default_lang=settings.CATALOGUE_DEFAULT_LANGUAGE):
 396         """Splits a URL book id into slug and language code.
 397
 398         Returns a dictionary usable i.e. for object lookup, or None.
 399
 400         >>> Book.split_urlid("a-slug/pol", default_lang="eng")
 401         {'slug': 'a-slug', 'language': 'pol'}
 402         >>> Book.split_urlid("a-slug", default_lang="eng")
 403         {'slug': 'a-slug', 'language': 'eng'}
 404         >>> Book.split_urlid("a-slug_pol", "_", default_lang="eng")
 405         {'slug': 'a-slug', 'language': 'pol'}
 406         >>> Book.split_urlid("a-slug/eng", default_lang="eng")
 407
 408         """
 409         parts = urlid.rsplit(sep, 1)
 410         if len(parts) == 2:
 411             if parts[1] == default_lang:
 412                 return None
 413             return {'slug': parts[0], 'language': parts[1]}
 414         else:
 415             return {'slug': urlid, 'language': default_lang}
 416
 417     @classmethod
 418     def split_fileid(cls, fileid):
 419         return cls.split_urlid(fileid, '_')
 420
 421     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
 422         from sortify import sortify
 423
 424         self.sort_key = sortify(self.title)
 425
 426         ret = super(Book, self).save(force_insert, force_update)
 427
 428         if reset_short_html:
 429             self.reset_short_html()
 430
 431         return ret
 432
 433     @permalink
 434     def get_absolute_url(self):
 435         return ('catalogue.views.book_detail', [self.urlid()])
 436
 437     @property
 438     def name(self):
 439         return self.title
 440
 441     def book_tag_slug(self):
 442         stem = 'l-' + self.slug
 443         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 444             return stem[:116] + ' ' + self.language
 445         else:
 446             return stem[:120]
 447
 448     def book_tag(self):
 449         slug = self.book_tag_slug()
 450         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
 451         if created:
 452             book_tag.name = self.title[:50]
 453             book_tag.sort_key = self.title.lower()
 454             book_tag.save()
 455         return book_tag
 456
 457     def has_media(self, type):
 458         if type in Book.formats:
 459             return bool(getattr(self, "%s_file" % type))
 460         else:
 461             return self.media.filter(type=type).exists()
 462
 463     def get_media(self, type):
 464         if self.has_media(type):
 465             if type in Book.formats:
 466                 return getattr(self, "%s_file" % type)
 467             else:
 468                 return self.media.filter(type=type)
 469         else:
 470             return None
 471
 472     def get_mp3(self):
 473         return self.get_media("mp3")
 474     def get_odt(self):
 475         return self.get_media("odt")
 476     def get_ogg(self):
 477         return self.get_media("ogg")
 478     def get_daisy(self):
 479         return self.get_media("daisy")
 480
 481     def reset_short_html(self):
 482         if self.id is None:
 483             return
 484
 485         cache_key = "Book.short_html/%d/%s"
 486         for lang, langname in settings.LANGUAGES:
 487             cache.delete(cache_key % (self.id, lang))
 488         # Fragment.short_html relies on book's tags, so reset it here too
 489         for fragm in self.fragments.all():
 490             fragm.reset_short_html()
 491
 492     def short_html(self):
 493         if self.id:
 494             cache_key = "Book.short_html/%d/%s" % (self.id, get_language())
 495             short_html = cache.get(cache_key)
 496         else:
 497             short_html = None
 498
 499         if short_html is not None:
 500             return mark_safe(short_html)
 501         else:
 502             tags = self.tags.filter(~Q(category__in=('set', 'theme', 'book')))
 503             tags = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name)) for tag in tags]
 504
 505             formats = []
 506             # files generated during publication
 507             if self.has_media("html"):
 508                 formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', args=[self.fileid()]), _('Read online')))
 509             for ebook_format in self.ebook_formats:
 510                 if self.has_media(ebook_format):
 511                     formats.append(u'<a href="%s">%s</a>' % (
 512                         self.get_media(ebook_format).url,
 513                         ebook_format.upper()
 514                     ))
 515             # other files
 516             for m in self.media.order_by('type'):
 517                 formats.append(u'<a href="%s">%s</a>' % (m.file.url, m.type.upper()))
 518
 519             formats = [mark_safe(format) for format in formats]
 520
 521             short_html = unicode(render_to_string('catalogue/book_short.html',
 522                 {'book': self, 'tags': tags, 'formats': formats}))
 523
 524             if self.id:
 525                 cache.set(cache_key, short_html, CACHE_FOREVER)
 526             return mark_safe(short_html)
 527
 528     def mini_box(self):
 529         if self.id:
 530             cache_key = "Book.mini_boxs/%d" % (self.id, )
 531             short_html = cache.get(cache_key)
 532         else:
 533             short_html = None
 534
 535         if short_html is None:
 536             authors = self.tags.filter(category='author')
 537
 538             short_html = unicode(render_to_string('catalogue/book_mini_box.html',
 539                 {'book': self, 'authors': authors, 'STATIC_URL': settings.STATIC_URL}))
 540
 541             if self.id:
 542                 cache.set(cache_key, short_html, CACHE_FOREVER)
 543         return mark_safe(short_html)
 544
 545     def has_description(self):
 546         return len(self.description) > 0
 547     has_description.short_description = _('description')
 548     has_description.boolean = True
 549
 550     # ugly ugly ugly
 551     def has_mp3_file(self):
 552         return bool(self.has_media("mp3"))
 553     has_mp3_file.short_description = 'MP3'
 554     has_mp3_file.boolean = True
 555
 556     def has_ogg_file(self):
 557         return bool(self.has_media("ogg"))
 558     has_ogg_file.short_description = 'OGG'
 559     has_ogg_file.boolean = True
 560
 561     def has_daisy_file(self):
 562         return bool(self.has_media("daisy"))
 563     has_daisy_file.short_description = 'DAISY'
 564     has_daisy_file.boolean = True
 565
 566     def wldocument(self, parse_dublincore=True):
 567         from catalogue.utils import ORMDocProvider
 568         from librarian.parser import WLDocument
 569
 570         return WLDocument.from_file(self.xml_file.path,
 571                 provider=ORMDocProvider(self),
 572                 parse_dublincore=parse_dublincore)
 573
 574     def build_pdf(self, customizations=None, file_name=None):
 575         """ (Re)builds the pdf file.
 576         customizations - customizations which are passed to LaTeX class file.
 577         file_name - save the pdf file under a different name and DO NOT save it in db.
 578         """
 579         from os import unlink
 580         from django.core.files import File
 581         from catalogue.utils import remove_zip
 582
 583         pdf = self.wldocument().as_pdf(customizations=customizations)
 584
 585         if file_name is None:
 586             # we'd like to be sure not to overwrite changes happening while
 587             # (timely) pdf generation is taking place (async celery scenario)
 588             current_self = Book.objects.get(id=self.id)
 589             current_self.pdf_file.save('%s.pdf' % self.fileid(),
 590                     File(open(pdf.get_filename())))
 591             self.pdf_file = current_self.pdf_file
 592
 593             # remove cached downloadables
 594             remove_zip(settings.ALL_PDF_ZIP)
 595
 596             for customized_pdf in get_existing_customized_pdf(self):
 597                 unlink(customized_pdf)
 598         else:
 599             print "saving %s" % file_name
 600             print "to: %s" % DefaultStorage().path(file_name)
 601             DefaultStorage().save(file_name, File(open(pdf.get_filename())))
 602
 603     def build_mobi(self):
 604         """ (Re)builds the MOBI file.
 605
 606         """
 607         from django.core.files import File
 608         from catalogue.utils import remove_zip
 609
 610         mobi = self.wldocument().as_mobi()
 611
 612         self.mobi_file.save('%s.mobi' % self.fileid(), File(open(mobi.get_filename())))
 613
 614         # remove zip with all mobi files
 615         remove_zip(settings.ALL_MOBI_ZIP)
 616
 617     def build_epub(self):
 618         """(Re)builds the epub file."""
 619         from django.core.files import File
 620         from catalogue.utils import remove_zip
 621
 622         epub = self.wldocument().as_epub()
 623
 624         self.epub_file.save('%s.epub' % self.fileid(),
 625                 File(open(epub.get_filename())))
 626
 627         # remove zip package with all epub files
 628         remove_zip(settings.ALL_EPUB_ZIP)
 629
 630     def build_txt(self):
 631         from django.core.files.base import ContentFile
 632
 633         text = self.wldocument().as_text()
 634         self.txt_file.save('%s.txt' % self.fileid(), ContentFile(text.get_string()))
 635
 636
 637     def build_html(self):
 638         from markupstring import MarkupString
 639         from django.core.files.base import ContentFile
 640         from slughifi import slughifi
 641         from librarian import html
 642
 643         meta_tags = list(self.tags.filter(
 644             category__in=('author', 'epoch', 'genre', 'kind')))
 645         book_tag = self.book_tag()
 646
 647         html_output = self.wldocument(parse_dublincore=False).as_html()
 648         if html_output:
 649             self.html_file.save('%s.html' % self.fileid(),
 650                     ContentFile(html_output.get_string()))
 651
 652             # get ancestor l-tags for adding to new fragments
 653             ancestor_tags = []
 654             p = self.parent
 655             while p:
 656                 ancestor_tags.append(p.book_tag())
 657                 p = p.parent
 658
 659             # Delete old fragments and create them from scratch
 660             self.fragments.all().delete()
 661             # Extract fragments
 662             closed_fragments, open_fragments = html.extract_fragments(self.html_file.path)
 663             for fragment in closed_fragments.values():
 664                 try:
 665                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 666                 except AttributeError:
 667                     continue
 668                 themes = []
 669                 for theme_name in theme_names:
 670                     if not theme_name:
 671                         continue
 672                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
 673                     if created:
 674                         tag.name = theme_name
 675                         tag.sort_key = theme_name.lower()
 676                         tag.save()
 677                     themes.append(tag)
 678                 if not themes:
 679                     continue
 680
 681                 text = fragment.to_string()
 682                 short_text = ''
 683                 if (len(MarkupString(text)) > 240):
 684                     short_text = unicode(MarkupString(text)[:160])
 685                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=self,
 686                     text=text, short_text=short_text)
 687
 688                 new_fragment.save()
 689                 new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
 690             self.save()
 691             self.html_built.send(sender=self)
 692             return True
 693         return False
 694
 695     @staticmethod
 696     def zip_format(format_):
 697         def pretty_file_name(book):
 698             return "%s/%s.%s" % (
 699                 b.get_extra_info_value()['author'],
 700                 b.fileid(),
 701                 format_)
 702
 703         field_name = "%s_file" % format_
 704         books = Book.objects.filter(parent=None).exclude(**{field_name: ""})
 705         paths = [(pretty_file_name(b), getattr(b, field_name).path)
 706                     for b in books]
 707         result = create_zip.delay(paths,
 708                     getattr(settings, "ALL_%s_ZIP" % format_.upper()))
 709         return result.wait()
 710
 711     def zip_audiobooks(self):
 712         bm = BookMedia.objects.filter(book=self, type='mp3')
 713         paths = map(lambda bm: (None, bm.file.path), bm)
 714         result = create_zip.delay(paths, self.fileid())
 715         return result.wait()
 716
 717     @classmethod
 718     def from_xml_file(cls, xml_file, **kwargs):
 719         from django.core.files import File
 720         from librarian import dcparser
 721
 722         # use librarian to parse meta-data
 723         book_info = dcparser.parse(xml_file)
 724
 725         if not isinstance(xml_file, File):
 726             xml_file = File(open(xml_file))
 727
 728         try:
 729             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
 730         finally:
 731             xml_file.close()
 732
 733     @classmethod
 734     def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
 735             build_epub=True, build_txt=True, build_pdf=True, build_mobi=True):
 736         import re
 737         from sortify import sortify
 738
 739         # check for parts before we do anything
 740         children = []
 741         if hasattr(book_info, 'parts'):
 742             for part_url in book_info.parts:
 743                 try:
 744                     children.append(Book.objects.get(
 745                         slug=part_url.slug, language=part_url.language))
 746                 except Book.DoesNotExist, e:
 747                     raise Book.DoesNotExist(_('Book "%s/%s" does not exist.') %
 748                             (part_url.slug, part_url.language))
 749
 750
 751         # Read book metadata
 752         book_slug = book_info.url.slug
 753         language = book_info.language
 754         if re.search(r'[^a-zA-Z0-9-]', book_slug):
 755             raise ValueError('Invalid characters in slug')
 756         book, created = Book.objects.get_or_create(slug=book_slug, language=language)
 757
 758         if created:
 759             book_shelves = []
 760         else:
 761             if not overwrite:
 762                 raise Book.AlreadyExists(_('Book %s/%s already exists') % (
 763                         book_slug, language))
 764             # Save shelves for this book
 765             book_shelves = list(book.tags.filter(category='set'))
 766
 767         book.title = book_info.title
 768         book.set_extra_info_value(book_info.to_dict())
 769         book.save()
 770
 771         meta_tags = Tag.tags_from_info(book_info)
 772
 773         book.tags = set(meta_tags + book_shelves)
 774
 775         book_tag = book.book_tag()
 776
 777         for n, child_book in enumerate(children):
 778             child_book.parent = book
 779             child_book.parent_number = n
 780             child_book.save()
 781
 782         # Save XML and HTML files
 783         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
 784
 785         # delete old fragments when overwriting
 786         book.fragments.all().delete()
 787
 788         if book.build_html():
 789             if not settings.NO_BUILD_TXT and build_txt:
 790                 book.build_txt()
 791
 792         if not settings.NO_BUILD_EPUB and build_epub:
 793             book.build_epub()
 794
 795         if not settings.NO_BUILD_PDF and build_pdf:
 796             book.build_pdf()
 797
 798         if not settings.NO_BUILD_MOBI and build_mobi:
 799             book.build_mobi()
 800
 801         book_descendants = list(book.children.all())
 802         descendants_tags = set()
 803         # add l-tag to descendants and their fragments
 804         while len(book_descendants) > 0:
 805             child_book = book_descendants.pop(0)
 806             descendants_tags.update(child_book.tags)
 807             child_book.tags = list(child_book.tags) + [book_tag]
 808             child_book.save()
 809             for fragment in child_book.fragments.all():
 810                 fragment.tags = set(list(fragment.tags) + [book_tag])
 811             book_descendants += list(child_book.children.all())
 812
 813         for tag in descendants_tags:
 814             touch_tag.delay(tag)
 815
 816         book.save()
 817
 818         # refresh cache
 819         book.reset_tag_counter()
 820         book.reset_theme_counter()
 821
 822         cls.published.send(sender=book)
 823         return book
 824
 825     def reset_tag_counter(self):
 826         if self.id is None:
 827             return
 828
 829         cache_key = "Book.tag_counter/%d" % self.id
 830         cache.delete(cache_key)
 831         if self.parent:
 832             self.parent.reset_tag_counter()
 833
 834     @property
 835     def tag_counter(self):
 836         if self.id:
 837             cache_key = "Book.tag_counter/%d" % self.id
 838             tags = cache.get(cache_key)
 839         else:
 840             tags = None
 841
 842         if tags is None:
 843             tags = {}
 844             for child in self.children.all().order_by():
 845                 for tag_pk, value in child.tag_counter.iteritems():
 846                     tags[tag_pk] = tags.get(tag_pk, 0) + value
 847             for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
 848                 tags[tag.pk] = 1
 849
 850             if self.id:
 851                 cache.set(cache_key, tags, CACHE_FOREVER)
 852         return tags
 853
 854     def reset_theme_counter(self):
 855         if self.id is None:
 856             return
 857
 858         cache_key = "Book.theme_counter/%d" % self.id
 859         cache.delete(cache_key)
 860         if self.parent:
 861             self.parent.reset_theme_counter()
 862
 863     @property
 864     def theme_counter(self):
 865         if self.id:
 866             cache_key = "Book.theme_counter/%d" % self.id
 867             tags = cache.get(cache_key)
 868         else:
 869             tags = None
 870
 871         if tags is None:
 872             tags = {}
 873             for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
 874                 for tag in fragment.tags.filter(category='theme').order_by():
 875                     tags[tag.pk] = tags.get(tag.pk, 0) + 1
 876
 877             if self.id:
 878                 cache.set(cache_key, tags, CACHE_FOREVER)
 879         return tags
 880
 881     def pretty_title(self, html_links=False):
 882         book = self
 883         names = list(book.tags.filter(category='author'))
 884
 885         books = []
 886         while book:
 887             books.append(book)
 888             book = book.parent
 889         names.extend(reversed(books))
 890
 891         if html_links:
 892             names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
 893         else:
 894             names = [tag.name for tag in names]
 895
 896         return ', '.join(names)
 897
 898     @classmethod
 899     def tagged_top_level(cls, tags):
 900         """ Returns top-level books tagged with `tags'.
 901
 902         It only returns those books which don't have ancestors which are
 903         also tagged with those tags.
 904
 905         """
 906         # get relevant books and their tags
 907         objects = cls.tagged.with_all(tags)
 908         # eliminate descendants
 909         l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in objects])
 910         descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags)]
 911         if descendants_keys:
 912             objects = objects.exclude(pk__in=descendants_keys)
 913
 914         return objects
 915
 916     @classmethod
 917     def book_list(cls, filter=None):
 918         """Generates a hierarchical listing of all books.
 919
 920         Books are optionally filtered with a test function.
 921
 922         """
 923
 924         books_by_parent = {}
 925         books = cls.objects.all().order_by('parent_number', 'sort_key').only(
 926                 'title', 'parent', 'slug', 'language')
 927         if filter:
 928             books = books.filter(filter).distinct()
 929             book_ids = set((book.pk for book in books))
 930             for book in books:
 931                 parent = book.parent_id
 932                 if parent not in book_ids:
 933                     parent = None
 934                 books_by_parent.setdefault(parent, []).append(book)
 935         else:
 936             for book in books:
 937                 books_by_parent.setdefault(book.parent_id, []).append(book)
 938
 939         orphans = []
 940         books_by_author = SortedDict()
 941         for tag in Tag.objects.filter(category='author'):
 942             books_by_author[tag] = []
 943
 944         for book in books_by_parent.get(None,()):
 945             authors = list(book.tags.filter(category='author'))
 946             if authors:
 947                 for author in authors:
 948                     books_by_author[author].append(book)
 949             else:
 950                 orphans.append(book)
 951
 952         return books_by_author, orphans, books_by_parent
 953
 954     _audiences_pl = {
 955         "SP1": (1, u"szkoła podstawowa"),
 956         "SP2": (1, u"szkoła podstawowa"),
 957         "P": (1, u"szkoła podstawowa"),
 958         "G": (2, u"gimnazjum"),
 959         "L": (3, u"liceum"),
 960         "LP": (3, u"liceum"),
 961     }
 962     def audiences_pl(self):
 963         audiences = self.get_extra_info_value().get('audiences', [])
 964         audiences = sorted(set([self._audiences_pl[a] for a in audiences]))
 965         return [a[1] for a in audiences]
 966
 967
 968 def _has_factory(ftype):
 969     has = lambda self: bool(getattr(self, "%s_file" % ftype))
 970     has.short_description = t.upper()
 971     has.boolean = True
 972     has.__name__ = "has_%s_file" % ftype
 973     return has
 974
 975
 976 # add the file fields
 977 for t in Book.formats:
 978     field_name = "%s_file" % t
 979     models.FileField(_("%s file" % t.upper()),
 980             upload_to=book_upload_path(t),
 981             blank=True).contribute_to_class(Book, field_name)
 982
 983     setattr(Book, "has_%s_file" % t, _has_factory(t))
 984
 985
 986 class Fragment(models.Model):
 987     text = models.TextField()
 988     short_text = models.TextField(editable=False)
 989     anchor = models.CharField(max_length=120)
 990     book = models.ForeignKey(Book, related_name='fragments')
 991
 992     objects = models.Manager()
 993     tagged = managers.ModelTaggedItemManager(Tag)
 994     tags = managers.TagDescriptor(Tag)
 995
 996     class Meta:
 997         ordering = ('book', 'anchor',)
 998         verbose_name = _('fragment')
 999         verbose_name_plural = _('fragments')
1000
1001     def get_absolute_url(self):
1002         return '%s#m%s' % (self.book.get_html_url(), self.anchor)
1003
1004     def reset_short_html(self):
1005         if self.id is None:
1006             return
1007
1008         cache_key = "Fragment.short_html/%d/%s"
1009         for lang, langname in settings.LANGUAGES:
1010             cache.delete(cache_key % (self.id, lang))
1011
1012     def short_html(self):
1013         if self.id:
1014             cache_key = "Fragment.short_html/%d/%s" % (self.id, get_language())
1015             short_html = cache.get(cache_key)
1016         else:
1017             short_html = None
1018
1019         if short_html is not None:
1020             return mark_safe(short_html)
1021         else:
1022             short_html = unicode(render_to_string('catalogue/fragment_short.html',
1023                 {'fragment': self}))
1024             if self.id:
1025                 cache.set(cache_key, short_html, CACHE_FOREVER)
1026             return mark_safe(short_html)
1027
1028
1029 ###########
1030 #
1031 # SIGNALS
1032 #
1033 ###########
1034
1035
1036 def _tags_updated_handler(sender, affected_tags, **kwargs):
1037     # reset tag global counter
1038     # we want Tag.changed_at updated for API to know the tag was touched
1039     for tag in affected_tags:
1040         touch_tag.delay(tag)
1041
1042     # if book tags changed, reset book tag counter
1043     if isinstance(sender, Book) and \
1044                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1045                     exclude(category__in=('book', 'theme', 'set')).count():
1046         sender.reset_tag_counter()
1047     # if fragment theme changed, reset book theme counter
1048     elif isinstance(sender, Fragment) and \
1049                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1050                     filter(category='theme').count():
1051         sender.book.reset_theme_counter()
1052 tags_updated.connect(_tags_updated_handler)
1053
1054
1055 def _pre_delete_handler(sender, instance, **kwargs):
1056     """ refresh Book on BookMedia delete """
1057     if sender == BookMedia:
1058         instance.book.save()
1059 pre_delete.connect(_pre_delete_handler)
1060
1061 def _post_save_handler(sender, instance, **kwargs):
1062     """ refresh all the short_html stuff on BookMedia update """
1063     if sender == BookMedia:
1064         instance.book.save()
1065 post_save.connect(_post_save_handler)