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, split_tags
  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 import search
  35
  36 TAG_CATEGORIES = (
  37     ('author', _('author')),
  38     ('epoch', _('epoch')),
  39     ('kind', _('kind')),
  40     ('genre', _('genre')),
  41     ('theme', _('theme')),
  42     ('set', _('set')),
  43     ('book', _('book')),
  44 )
  45
  46 # not quite, but Django wants you to set a timeout
  47 CACHE_FOREVER = 2419200  # 28 days
  48
  49
  50 class TagSubcategoryManager(models.Manager):
  51     def __init__(self, subcategory):
  52         super(TagSubcategoryManager, self).__init__()
  53         self.subcategory = subcategory
  54
  55     def get_query_set(self):
  56         return super(TagSubcategoryManager, self).get_query_set().filter(category=self.subcategory)
  57
  58
  59 class Tag(TagBase):
  60     name = models.CharField(_('name'), max_length=50, db_index=True)
  61     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
  62     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True)
  63     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
  64         db_index=True, choices=TAG_CATEGORIES)
  65     description = models.TextField(_('description'), blank=True)
  66
  67     user = models.ForeignKey(User, blank=True, null=True)
  68     book_count = models.IntegerField(_('book count'), blank=True, null=True)
  69     gazeta_link = models.CharField(blank=True, max_length=240)
  70     wiki_link = models.CharField(blank=True, max_length=240)
  71
  72     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
  73     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
  74
  75     class UrlDeprecationWarning(DeprecationWarning):
  76         pass
  77
  78     categories_rev = {
  79         'autor': 'author',
  80         'epoka': 'epoch',
  81         'rodzaj': 'kind',
  82         'gatunek': 'genre',
  83         'motyw': 'theme',
  84         'polka': 'set',
  85     }
  86     categories_dict = dict((item[::-1] for item in categories_rev.iteritems()))
  87
  88     class Meta:
  89         ordering = ('sort_key',)
  90         verbose_name = _('tag')
  91         verbose_name_plural = _('tags')
  92         unique_together = (("slug", "category"),)
  93
  94     def __unicode__(self):
  95         return self.name
  96
  97     def __repr__(self):
  98         return "Tag(slug=%r)" % self.slug
  99
 100     @permalink
 101     def get_absolute_url(self):
 102         return ('catalogue.views.tagged_object_list', [self.url_chunk])
 103
 104     def has_description(self):
 105         return len(self.description) > 0
 106     has_description.short_description = _('description')
 107     has_description.boolean = True
 108
 109     def get_count(self):
 110         """Returns global book count for book tags, fragment count for themes."""
 111
 112         if self.category == 'book':
 113             # never used
 114             objects = Book.objects.none()
 115         elif self.category == 'theme':
 116             objects = Fragment.tagged.with_all((self,))
 117         else:
 118             objects = Book.tagged.with_all((self,)).order_by()
 119             if self.category != 'set':
 120                 # eliminate descendants
 121                 l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects])
 122                 descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
 123                 if descendants_keys:
 124                     objects = objects.exclude(pk__in=descendants_keys)
 125         return objects.count()
 126
 127     @staticmethod
 128     def get_tag_list(tags):
 129         if isinstance(tags, basestring):
 130             real_tags = []
 131             ambiguous_slugs = []
 132             category = None
 133             deprecated = False
 134             tags_splitted = tags.split('/')
 135             for name in tags_splitted:
 136                 if category:
 137                     real_tags.append(Tag.objects.get(slug=name, category=category))
 138                     category = None
 139                 elif name in Tag.categories_rev:
 140                     category = Tag.categories_rev[name]
 141                 else:
 142                     try:
 143                         real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
 144                         deprecated = True
 145                     except Tag.MultipleObjectsReturned, e:
 146                         ambiguous_slugs.append(name)
 147
 148             if category:
 149                 # something strange left off
 150                 raise Tag.DoesNotExist()
 151             if ambiguous_slugs:
 152                 # some tags should be qualified
 153                 e = Tag.MultipleObjectsReturned()
 154                 e.tags = real_tags
 155                 e.ambiguous_slugs = ambiguous_slugs
 156                 raise e
 157             if deprecated:
 158                 e = Tag.UrlDeprecationWarning()
 159                 e.tags = real_tags
 160                 raise e
 161             return real_tags
 162         else:
 163             return TagBase.get_tag_list(tags)
 164
 165     @property
 166     def url_chunk(self):
 167         return '/'.join((Tag.categories_dict[self.category], self.slug))
 168
 169     @staticmethod
 170     def tags_from_info(info):
 171         from slughifi import slughifi
 172         from sortify import sortify
 173         meta_tags = []
 174         categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
 175         for field_name, category in categories:
 176             try:
 177                 tag_names = getattr(info, field_name)
 178             except:
 179                 try:
 180                     tag_names = [getattr(info, category)]
 181                 except:
 182                     # For instance, Pictures do not have 'genre' field.
 183                     continue
 184             for tag_name in tag_names:
 185                 tag_sort_key = tag_name
 186                 if category == 'author':
 187                     tag_sort_key = tag_name.last_name
 188                     tag_name = tag_name.readable()
 189                 tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
 190                 if created:
 191                     tag.name = tag_name
 192                     tag.sort_key = sortify(tag_sort_key.lower())
 193                     tag.save()
 194                 meta_tags.append(tag)
 195         return meta_tags
 196
 197
 198
 199 def get_dynamic_path(media, filename, ext=None, maxlen=100):
 200     from slughifi import slughifi
 201
 202     # how to put related book's slug here?
 203     if not ext:
 204         # BookMedia case
 205         ext = media.formats[media.type].ext
 206     if media is None or not media.name:
 207         name = slughifi(filename.split(".")[0])
 208     else:
 209         name = slughifi(media.name)
 210     return 'book/%s/%s.%s' % (ext, name[:maxlen-len('book/%s/.%s' % (ext, ext))-4], ext)
 211
 212
 213 # TODO: why is this hard-coded ?
 214 def book_upload_path(ext=None, maxlen=100):
 215     return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen)
 216
 217
 218 def get_customized_pdf_path(book, customizations):
 219     """
 220     Returns a MEDIA_ROOT relative path for a customized pdf. The name will contain a hash of customization options.
 221     """
 222     customizations.sort()
 223     h = hash(tuple(customizations))
 224
 225     pdf_name = '%s-custom-%s' % (book.fileid(), h)
 226     pdf_file = get_dynamic_path(None, pdf_name, ext='pdf')
 227
 228     return pdf_file
 229
 230
 231 def get_existing_customized_pdf(book):
 232     """
 233     Returns a list of paths to generated customized pdf of a book
 234     """
 235     pdf_glob = '%s-custom-' % (book.fileid(),)
 236     pdf_glob = get_dynamic_path(None, pdf_glob, ext='pdf')
 237     pdf_glob = re.sub(r"[.]([a-z0-9]+)$", "*.\\1", pdf_glob)
 238     return glob(path.join(settings.MEDIA_ROOT, pdf_glob))
 239
 240
 241 class BookMedia(models.Model):
 242     FileFormat = namedtuple("FileFormat", "name ext")
 243     formats = SortedDict([
 244         ('mp3', FileFormat(name='MP3', ext='mp3')),
 245         ('ogg', FileFormat(name='Ogg Vorbis', ext='ogg')),
 246         ('daisy', FileFormat(name='DAISY', ext='daisy.zip')),
 247     ])
 248     format_choices = [(k, _('%s file') % t.name)
 249             for k, t in formats.items()]
 250
 251     type        = models.CharField(_('type'), choices=format_choices, max_length="100")
 252     name        = models.CharField(_('name'), max_length="100")
 253     file        = OverwritingFileField(_('file'), upload_to=book_upload_path())
 254     uploaded_at = models.DateTimeField(_('creation date'), auto_now_add=True, editable=False)
 255     extra_info  = JSONField(_('extra information'), default='{}', editable=False)
 256     book = models.ForeignKey('Book', related_name='media')
 257     source_sha1 = models.CharField(null=True, blank=True, max_length=40, editable=False)
 258
 259     def __unicode__(self):
 260         return "%s (%s)" % (self.name, self.file.name.split("/")[-1])
 261
 262     class Meta:
 263         ordering            = ('type', 'name')
 264         verbose_name        = _('book media')
 265         verbose_name_plural = _('book media')
 266
 267     def save(self, *args, **kwargs):
 268         from slughifi import slughifi
 269         from catalogue.utils import ExistingFile, remove_zip
 270
 271         try:
 272             old = BookMedia.objects.get(pk=self.pk)
 273         except BookMedia.DoesNotExist, e:
 274             pass
 275         else:
 276             # if name changed, change the file name, too
 277             if slughifi(self.name) != slughifi(old.name):
 278                 self.file.save(None, ExistingFile(self.file.path), save=False, leave=True)
 279
 280         super(BookMedia, self).save(*args, **kwargs)
 281
 282         # remove the zip package for book with modified media
 283         remove_zip(self.book.fileid())
 284
 285         extra_info = self.get_extra_info_value()
 286         extra_info.update(self.read_meta())
 287         self.set_extra_info_value(extra_info)
 288         self.source_sha1 = self.read_source_sha1(self.file.path, self.type)
 289         return super(BookMedia, self).save(*args, **kwargs)
 290
 291     def read_meta(self):
 292         """
 293             Reads some metadata from the audiobook.
 294         """
 295         import mutagen
 296         from mutagen import id3
 297
 298         artist_name = director_name = project = funded_by = ''
 299         if self.type == 'mp3':
 300             try:
 301                 audio = id3.ID3(self.file.path)
 302                 artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
 303                 director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
 304                 project = ", ".join([t.data for t in audio.getall('PRIV')
 305                         if t.owner=='wolnelektury.pl?project'])
 306                 funded_by = ", ".join([t.data for t in audio.getall('PRIV')
 307                         if t.owner=='wolnelektury.pl?funded_by'])
 308             except:
 309                 pass
 310         elif self.type == 'ogg':
 311             try:
 312                 audio = mutagen.File(self.file.path)
 313                 artist_name = ', '.join(audio.get('artist', []))
 314                 director_name = ', '.join(audio.get('conductor', []))
 315                 project = ", ".join(audio.get('project', []))
 316                 funded_by = ", ".join(audio.get('funded_by', []))
 317             except:
 318                 pass
 319         else:
 320             return {}
 321         return {'artist_name': artist_name, 'director_name': director_name,
 322                 'project': project, 'funded_by': funded_by}
 323
 324     @staticmethod
 325     def read_source_sha1(filepath, filetype):
 326         """
 327             Reads source file SHA1 from audiobok metadata.
 328         """
 329         import mutagen
 330         from mutagen import id3
 331
 332         if filetype == 'mp3':
 333             try:
 334                 audio = id3.ID3(filepath)
 335                 return [t.data for t in audio.getall('PRIV')
 336                         if t.owner=='wolnelektury.pl?flac_sha1'][0]
 337             except:
 338                 return None
 339         elif filetype == 'ogg':
 340             try:
 341                 audio = mutagen.File(filepath)
 342                 return audio.get('flac_sha1', [None])[0]
 343             except:
 344                 return None
 345         else:
 346             return None
 347
 348
 349 class Book(models.Model):
 350     title         = models.CharField(_('title'), max_length=120)
 351     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True, editable=False)
 352     slug          = models.SlugField(_('slug'), max_length=120, db_index=True)
 353     language = models.CharField(_('language code'), max_length=3, db_index=True,
 354                     default=settings.CATALOGUE_DEFAULT_LANGUAGE)
 355     description   = models.TextField(_('description'), blank=True)
 356     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
 357     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
 358     parent_number = models.IntegerField(_('parent number'), default=0)
 359     extra_info    = JSONField(_('extra information'), default='{}')
 360     gazeta_link   = models.CharField(blank=True, max_length=240)
 361     wiki_link     = models.CharField(blank=True, max_length=240)
 362     # files generated during publication
 363
 364     cover = models.FileField(_('cover'), upload_to=book_upload_path('png'),
 365                 null=True, blank=True)
 366     ebook_formats = ['pdf', 'epub', 'mobi', 'txt']
 367     formats = ebook_formats + ['html', 'xml']
 368
 369     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
 370     objects  = models.Manager()
 371     tagged   = managers.ModelTaggedItemManager(Tag)
 372     tags     = managers.TagDescriptor(Tag)
 373
 374     html_built = django.dispatch.Signal()
 375     published = django.dispatch.Signal()
 376
 377     URLID_RE = r'[a-z0-9-]+(?:/[a-z]{3})?'
 378     FILEID_RE = r'[a-z0-9-]+(?:_[a-z]{3})?'
 379
 380     class AlreadyExists(Exception):
 381         pass
 382
 383     class Meta:
 384         unique_together = [['slug', 'language']]
 385         ordering = ('sort_key',)
 386         verbose_name = _('book')
 387         verbose_name_plural = _('books')
 388
 389     def __unicode__(self):
 390         return self.title
 391
 392     def urlid(self, sep='/'):
 393         stem = self.slug
 394         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 395             stem += sep + self.language
 396         return stem
 397
 398     def fileid(self):
 399         return self.urlid('_')
 400
 401     @staticmethod
 402     def split_urlid(urlid, sep='/', default_lang=settings.CATALOGUE_DEFAULT_LANGUAGE):
 403         """Splits a URL book id into slug and language code.
 404
 405         Returns a dictionary usable i.e. for object lookup, or None.
 406
 407         >>> Book.split_urlid("a-slug/pol", default_lang="eng")
 408         {'slug': 'a-slug', 'language': 'pol'}
 409         >>> Book.split_urlid("a-slug", default_lang="eng")
 410         {'slug': 'a-slug', 'language': 'eng'}
 411         >>> Book.split_urlid("a-slug_pol", "_", default_lang="eng")
 412         {'slug': 'a-slug', 'language': 'pol'}
 413         >>> Book.split_urlid("a-slug/eng", default_lang="eng")
 414
 415         """
 416         parts = urlid.rsplit(sep, 1)
 417         if len(parts) == 2:
 418             if parts[1] == default_lang:
 419                 return None
 420             return {'slug': parts[0], 'language': parts[1]}
 421         else:
 422             return {'slug': urlid, 'language': default_lang}
 423
 424     @classmethod
 425     def split_fileid(cls, fileid):
 426         return cls.split_urlid(fileid, '_')
 427
 428     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
 429         from sortify import sortify
 430
 431         self.sort_key = sortify(self.title)
 432
 433         ret = super(Book, self).save(force_insert, force_update)
 434
 435         if reset_short_html:
 436             self.reset_short_html()
 437
 438         return ret
 439
 440     @permalink
 441     def get_absolute_url(self):
 442         return ('catalogue.views.book_detail', [self.urlid()])
 443
 444     @property
 445     def name(self):
 446         return self.title
 447
 448     def book_tag_slug(self):
 449         stem = 'l-' + self.slug
 450         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 451             return stem[:116] + ' ' + self.language
 452         else:
 453             return stem[:120]
 454
 455     def book_tag(self):
 456         slug = self.book_tag_slug()
 457         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
 458         if created:
 459             book_tag.name = self.title[:50]
 460             book_tag.sort_key = self.title.lower()
 461             book_tag.save()
 462         return book_tag
 463
 464     def has_media(self, type):
 465         if type in Book.formats:
 466             return bool(getattr(self, "%s_file" % type))
 467         else:
 468             return self.media.filter(type=type).exists()
 469
 470     def get_media(self, type):
 471         if self.has_media(type):
 472             if type in Book.formats:
 473                 return getattr(self, "%s_file" % type)
 474             else:
 475                 return self.media.filter(type=type)
 476         else:
 477             return None
 478
 479     def get_mp3(self):
 480         return self.get_media("mp3")
 481     def get_odt(self):
 482         return self.get_media("odt")
 483     def get_ogg(self):
 484         return self.get_media("ogg")
 485     def get_daisy(self):
 486         return self.get_media("daisy")
 487
 488     def reset_short_html(self):
 489         if self.id is None:
 490             return
 491
 492         cache_key = "Book.short_html/%d/%s"
 493         for lang, langname in settings.LANGUAGES:
 494             cache.delete(cache_key % (self.id, lang))
 495         cache.delete("Book.mini_box/%d" % (self.id, ))
 496         # Fragment.short_html relies on book's tags, so reset it here too
 497         for fragm in self.fragments.all():
 498             fragm.reset_short_html()
 499
 500     def short_html(self):
 501         if self.id:
 502             cache_key = "Book.short_html/%d/%s" % (self.id, get_language())
 503             short_html = cache.get(cache_key)
 504         else:
 505             short_html = None
 506
 507         if short_html is not None:
 508             return mark_safe(short_html)
 509         else:
 510             tags = self.tags.filter(category__in=('author', 'kind', 'genre', 'epoch'))
 511             tags = split_tags(tags)
 512
 513             formats = []
 514             # files generated during publication
 515             for ebook_format in self.ebook_formats:
 516                 if self.has_media(ebook_format):
 517                     formats.append(u'<a href="%s">%s</a>' % (
 518                         "", #self.get_media(ebook_format).url,
 519                         ebook_format.upper()
 520                     ))
 521
 522             formats = [mark_safe(format) for format in formats]
 523
 524             short_html = unicode(render_to_string('catalogue/book_short.html',
 525                 {'book': self, 'tags': tags, 'formats': formats}))
 526
 527             if self.id:
 528                 cache.set(cache_key, short_html, CACHE_FOREVER)
 529             return mark_safe(short_html)
 530
 531     def mini_box(self):
 532         if self.id:
 533             cache_key = "Book.mini_box/%d" % (self.id, )
 534             short_html = cache.get(cache_key)
 535         else:
 536             short_html = None
 537
 538         if short_html is None:
 539             authors = self.tags.filter(category='author')
 540
 541             short_html = unicode(render_to_string('catalogue/book_mini_box.html',
 542                 {'book': self, 'authors': authors, 'STATIC_URL': settings.STATIC_URL}))
 543
 544             if self.id:
 545                 cache.set(cache_key, short_html, CACHE_FOREVER)
 546         return mark_safe(short_html)
 547
 548     def has_description(self):
 549         return len(self.description) > 0
 550     has_description.short_description = _('description')
 551     has_description.boolean = True
 552
 553     # ugly ugly ugly
 554     def has_mp3_file(self):
 555         return bool(self.has_media("mp3"))
 556     has_mp3_file.short_description = 'MP3'
 557     has_mp3_file.boolean = True
 558
 559     def has_ogg_file(self):
 560         return bool(self.has_media("ogg"))
 561     has_ogg_file.short_description = 'OGG'
 562     has_ogg_file.boolean = True
 563
 564     def has_daisy_file(self):
 565         return bool(self.has_media("daisy"))
 566     has_daisy_file.short_description = 'DAISY'
 567     has_daisy_file.boolean = True
 568
 569     def wldocument(self, parse_dublincore=True):
 570         from catalogue.utils import ORMDocProvider
 571         from librarian.parser import WLDocument
 572
 573         return WLDocument.from_file(self.xml_file.path,
 574                 provider=ORMDocProvider(self),
 575                 parse_dublincore=parse_dublincore)
 576
 577     def build_cover(self, book_info=None):
 578         """(Re)builds the cover image."""
 579         from StringIO import StringIO
 580         from django.core.files.base import ContentFile
 581         from librarian.cover import WLCover
 582
 583         if book_info is None:
 584             book_info = self.wldocument().book_info
 585
 586         cover = WLCover(book_info).image()
 587         imgstr = StringIO()
 588         cover.save(imgstr, 'png')
 589         self.cover.save(None, ContentFile(imgstr.getvalue()))
 590
 591     def build_pdf(self, customizations=None, file_name=None):
 592         """ (Re)builds the pdf file.
 593         customizations - customizations which are passed to LaTeX class file.
 594         file_name - save the pdf file under a different name and DO NOT save it in db.
 595         """
 596         from os import unlink
 597         from django.core.files import File
 598         from catalogue.utils import remove_zip
 599
 600         pdf = self.wldocument().as_pdf(customizations=customizations)
 601
 602         if file_name is None:
 603             # we'd like to be sure not to overwrite changes happening while
 604             # (timely) pdf generation is taking place (async celery scenario)
 605             current_self = Book.objects.get(id=self.id)
 606             current_self.pdf_file.save('%s.pdf' % self.fileid(),
 607                     File(open(pdf.get_filename())))
 608             self.pdf_file = current_self.pdf_file
 609
 610             # remove cached downloadables
 611             remove_zip(settings.ALL_PDF_ZIP)
 612
 613             for customized_pdf in get_existing_customized_pdf(self):
 614                 unlink(customized_pdf)
 615         else:
 616             print "saving %s" % file_name
 617             print "to: %s" % DefaultStorage().path(file_name)
 618             DefaultStorage().save(file_name, File(open(pdf.get_filename())))
 619
 620     def build_mobi(self):
 621         """ (Re)builds the MOBI file.
 622
 623         """
 624         from django.core.files import File
 625         from catalogue.utils import remove_zip
 626
 627         mobi = self.wldocument().as_mobi()
 628
 629         self.mobi_file.save('%s.mobi' % self.fileid(), File(open(mobi.get_filename())))
 630
 631         # remove zip with all mobi files
 632         remove_zip(settings.ALL_MOBI_ZIP)
 633
 634     def build_epub(self):
 635         """(Re)builds the epub file."""
 636         from django.core.files import File
 637         from catalogue.utils import remove_zip
 638
 639         epub = self.wldocument().as_epub()
 640
 641         self.epub_file.save('%s.epub' % self.fileid(),
 642                 File(open(epub.get_filename())))
 643
 644         # remove zip package with all epub files
 645         remove_zip(settings.ALL_EPUB_ZIP)
 646
 647     def build_txt(self):
 648         from django.core.files.base import ContentFile
 649
 650         text = self.wldocument().as_text()
 651         self.txt_file.save('%s.txt' % self.fileid(), ContentFile(text.get_string()))
 652
 653
 654     def build_html(self):
 655         from markupstring import MarkupString
 656         from django.core.files.base import ContentFile
 657         from slughifi import slughifi
 658         from librarian import html
 659
 660         meta_tags = list(self.tags.filter(
 661             category__in=('author', 'epoch', 'genre', 'kind')))
 662         book_tag = self.book_tag()
 663
 664         html_output = self.wldocument(parse_dublincore=False).as_html()
 665         if html_output:
 666             self.html_file.save('%s.html' % self.fileid(),
 667                     ContentFile(html_output.get_string()))
 668
 669             # get ancestor l-tags for adding to new fragments
 670             ancestor_tags = []
 671             p = self.parent
 672             while p:
 673                 ancestor_tags.append(p.book_tag())
 674                 p = p.parent
 675
 676             # Delete old fragments and create them from scratch
 677             self.fragments.all().delete()
 678             # Extract fragments
 679             closed_fragments, open_fragments = html.extract_fragments(self.html_file.path)
 680             for fragment in closed_fragments.values():
 681                 try:
 682                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 683                 except AttributeError:
 684                     continue
 685                 themes = []
 686                 for theme_name in theme_names:
 687                     if not theme_name:
 688                         continue
 689                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
 690                     if created:
 691                         tag.name = theme_name
 692                         tag.sort_key = theme_name.lower()
 693                         tag.save()
 694                     themes.append(tag)
 695                 if not themes:
 696                     continue
 697
 698                 text = fragment.to_string()
 699                 short_text = ''
 700                 if (len(MarkupString(text)) > 240):
 701                     short_text = unicode(MarkupString(text)[:160])
 702                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=self,
 703                     text=text, short_text=short_text)
 704
 705                 new_fragment.save()
 706                 new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
 707             self.save()
 708             self.html_built.send(sender=self)
 709             return True
 710         return False
 711
 712     @staticmethod
 713     def zip_format(format_):
 714         def pretty_file_name(book):
 715             return "%s/%s.%s" % (
 716                 b.get_extra_info_value()['author'],
 717                 b.fileid(),
 718                 format_)
 719
 720         field_name = "%s_file" % format_
 721         books = Book.objects.filter(parent=None).exclude(**{field_name: ""})
 722         paths = [(pretty_file_name(b), getattr(b, field_name).path)
 723                     for b in books]
 724         result = create_zip.delay(paths,
 725                     getattr(settings, "ALL_%s_ZIP" % format_.upper()))
 726         return result.wait()
 727
 728     def zip_audiobooks(self):
 729         bm = BookMedia.objects.filter(book=self, type='mp3')
 730         paths = map(lambda bm: (None, bm.file.path), bm)
 731         result = create_zip.delay(paths, self.fileid())
 732         return result.wait()
 733
 734     def search_index(self, book_info=None):
 735         if settings.SEARCH_INDEX_PARALLEL:
 736             if instance(settings.SEARCH_INDEX_PARALLEL, int):
 737                 idx = search.ReusableIndex(threads=4)
 738             else:
 739                 idx = search.ReusableIndex()
 740         else:
 741             idx = search.Index()
 742
 743         idx.open()
 744         try:
 745             idx.index_book(self, book_info)
 746         finally:
 747             idx.close()
 748
 749     @classmethod
 750     def from_xml_file(cls, xml_file, **kwargs):
 751         from django.core.files import File
 752         from librarian import dcparser
 753
 754         # use librarian to parse meta-data
 755         book_info = dcparser.parse(xml_file)
 756
 757         if not isinstance(xml_file, File):
 758             xml_file = File(open(xml_file))
 759
 760         try:
 761             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
 762         finally:
 763             xml_file.close()
 764
 765     @classmethod
 766     def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
 767             build_epub=True, build_txt=True, build_pdf=True, build_mobi=True,
 768             search_index=True):
 769         import re
 770         from sortify import sortify
 771
 772         # check for parts before we do anything
 773         children = []
 774         if hasattr(book_info, 'parts'):
 775             for part_url in book_info.parts:
 776                 try:
 777                     children.append(Book.objects.get(
 778                         slug=part_url.slug, language=part_url.language))
 779                 except Book.DoesNotExist, e:
 780                     raise Book.DoesNotExist(_('Book "%s/%s" does not exist.') %
 781                             (part_url.slug, part_url.language))
 782
 783
 784         # Read book metadata
 785         book_slug = book_info.url.slug
 786         language = book_info.language
 787         if re.search(r'[^a-zA-Z0-9-]', book_slug):
 788             raise ValueError('Invalid characters in slug')
 789         book, created = Book.objects.get_or_create(slug=book_slug, language=language)
 790
 791         if created:
 792             book_shelves = []
 793         else:
 794             if not overwrite:
 795                 raise Book.AlreadyExists(_('Book %s/%s already exists') % (
 796                         book_slug, language))
 797             # Save shelves for this book
 798             book_shelves = list(book.tags.filter(category='set'))
 799
 800         book.title = book_info.title
 801         book.set_extra_info_value(book_info.to_dict())
 802         book.save()
 803
 804         meta_tags = Tag.tags_from_info(book_info)
 805
 806         book.tags = set(meta_tags + book_shelves)
 807
 808         book_tag = book.book_tag()
 809
 810         for n, child_book in enumerate(children):
 811             child_book.parent = book
 812             child_book.parent_number = n
 813             child_book.save()
 814
 815         # Save XML and HTML files
 816         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
 817
 818         # delete old fragments when overwriting
 819         book.fragments.all().delete()
 820
 821         if book.build_html():
 822             if not settings.NO_BUILD_TXT and build_txt:
 823                 book.build_txt()
 824
 825         book.build_cover(book_info)
 826
 827         if not settings.NO_BUILD_EPUB and build_epub:
 828             book.build_epub()
 829
 830         if not settings.NO_BUILD_PDF and build_pdf:
 831             book.build_pdf()
 832
 833         if not settings.NO_BUILD_MOBI and build_mobi:
 834             book.build_mobi()
 835
 836         if not settings.NO_SEARCH_INDEX and search_index:
 837             book.search_index(book_info)
 838
 839         book_descendants = list(book.children.all())
 840         descendants_tags = set()
 841         # add l-tag to descendants and their fragments
 842         while len(book_descendants) > 0:
 843             child_book = book_descendants.pop(0)
 844             descendants_tags.update(child_book.tags)
 845             child_book.tags = list(child_book.tags) + [book_tag]
 846             child_book.save()
 847             for fragment in child_book.fragments.all():
 848                 fragment.tags = set(list(fragment.tags) + [book_tag])
 849             book_descendants += list(child_book.children.all())
 850
 851         for tag in descendants_tags:
 852             touch_tag.delay(tag)
 853
 854         book.save()
 855
 856         # refresh cache
 857         book.reset_tag_counter()
 858         book.reset_theme_counter()
 859
 860         cls.published.send(sender=book)
 861         return book
 862
 863     def reset_tag_counter(self):
 864         if self.id is None:
 865             return
 866
 867         cache_key = "Book.tag_counter/%d" % self.id
 868         cache.delete(cache_key)
 869         if self.parent:
 870             self.parent.reset_tag_counter()
 871
 872     @property
 873     def tag_counter(self):
 874         if self.id:
 875             cache_key = "Book.tag_counter/%d" % self.id
 876             tags = cache.get(cache_key)
 877         else:
 878             tags = None
 879
 880         if tags is None:
 881             tags = {}
 882             for child in self.children.all().order_by():
 883                 for tag_pk, value in child.tag_counter.iteritems():
 884                     tags[tag_pk] = tags.get(tag_pk, 0) + value
 885             for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
 886                 tags[tag.pk] = 1
 887
 888             if self.id:
 889                 cache.set(cache_key, tags, CACHE_FOREVER)
 890         return tags
 891
 892     def reset_theme_counter(self):
 893         if self.id is None:
 894             return
 895
 896         cache_key = "Book.theme_counter/%d" % self.id
 897         cache.delete(cache_key)
 898         if self.parent:
 899             self.parent.reset_theme_counter()
 900
 901     @property
 902     def theme_counter(self):
 903         if self.id:
 904             cache_key = "Book.theme_counter/%d" % self.id
 905             tags = cache.get(cache_key)
 906         else:
 907             tags = None
 908
 909         if tags is None:
 910             tags = {}
 911             for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
 912                 for tag in fragment.tags.filter(category='theme').order_by():
 913                     tags[tag.pk] = tags.get(tag.pk, 0) + 1
 914
 915             if self.id:
 916                 cache.set(cache_key, tags, CACHE_FOREVER)
 917         return tags
 918
 919     def pretty_title(self, html_links=False):
 920         book = self
 921         names = list(book.tags.filter(category='author'))
 922
 923         books = []
 924         while book:
 925             books.append(book)
 926             book = book.parent
 927         names.extend(reversed(books))
 928
 929         if html_links:
 930             names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
 931         else:
 932             names = [tag.name for tag in names]
 933
 934         return ', '.join(names)
 935
 936     @classmethod
 937     def tagged_top_level(cls, tags):
 938         """ Returns top-level books tagged with `tags'.
 939
 940         It only returns those books which don't have ancestors which are
 941         also tagged with those tags.
 942
 943         """
 944         # get relevant books and their tags
 945         objects = cls.tagged.with_all(tags)
 946         # eliminate descendants
 947         l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in objects])
 948         descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags)]
 949         if descendants_keys:
 950             objects = objects.exclude(pk__in=descendants_keys)
 951
 952         return objects
 953
 954     @classmethod
 955     def book_list(cls, filter=None):
 956         """Generates a hierarchical listing of all books.
 957
 958         Books are optionally filtered with a test function.
 959
 960         """
 961
 962         books_by_parent = {}
 963         books = cls.objects.all().order_by('parent_number', 'sort_key').only(
 964                 'title', 'parent', 'slug', 'language')
 965         if filter:
 966             books = books.filter(filter).distinct()
 967             book_ids = set((book.pk for book in books))
 968             for book in books:
 969                 parent = book.parent_id
 970                 if parent not in book_ids:
 971                     parent = None
 972                 books_by_parent.setdefault(parent, []).append(book)
 973         else:
 974             for book in books:
 975                 books_by_parent.setdefault(book.parent_id, []).append(book)
 976
 977         orphans = []
 978         books_by_author = SortedDict()
 979         for tag in Tag.objects.filter(category='author'):
 980             books_by_author[tag] = []
 981
 982         for book in books_by_parent.get(None,()):
 983             authors = list(book.tags.filter(category='author'))
 984             if authors:
 985                 for author in authors:
 986                     books_by_author[author].append(book)
 987             else:
 988                 orphans.append(book)
 989
 990         return books_by_author, orphans, books_by_parent
 991
 992     _audiences_pl = {
 993         "SP1": (1, u"szkoła podstawowa"),
 994         "SP2": (1, u"szkoła podstawowa"),
 995         "P": (1, u"szkoła podstawowa"),
 996         "G": (2, u"gimnazjum"),
 997         "L": (3, u"liceum"),
 998         "LP": (3, u"liceum"),
 999     }
1000     def audiences_pl(self):
1001         audiences = self.get_extra_info_value().get('audiences', [])
1002         audiences = sorted(set([self._audiences_pl[a] for a in audiences]))
1003         return [a[1] for a in audiences]
1004
1005
1006 def _has_factory(ftype):
1007     has = lambda self: bool(getattr(self, "%s_file" % ftype))
1008     has.short_description = t.upper()
1009     has.boolean = True
1010     has.__name__ = "has_%s_file" % ftype
1011     return has
1012
1013
1014 # add the file fields
1015 for t in Book.formats:
1016     field_name = "%s_file" % t
1017     models.FileField(_("%s file" % t.upper()),
1018             upload_to=book_upload_path(t),
1019             blank=True).contribute_to_class(Book, field_name)
1020
1021     setattr(Book, "has_%s_file" % t, _has_factory(t))
1022
1023
1024 class Fragment(models.Model):
1025     text = models.TextField()
1026     short_text = models.TextField(editable=False)
1027     anchor = models.CharField(max_length=120)
1028     book = models.ForeignKey(Book, related_name='fragments')
1029
1030     objects = models.Manager()
1031     tagged = managers.ModelTaggedItemManager(Tag)
1032     tags = managers.TagDescriptor(Tag)
1033
1034     class Meta:
1035         ordering = ('book', 'anchor',)
1036         verbose_name = _('fragment')
1037         verbose_name_plural = _('fragments')
1038
1039     def get_absolute_url(self):
1040         return '%s#m%s' % (self.book.get_html_url(), self.anchor)
1041
1042     def reset_short_html(self):
1043         if self.id is None:
1044             return
1045
1046         cache_key = "Fragment.short_html/%d/%s"
1047         for lang, langname in settings.LANGUAGES:
1048             cache.delete(cache_key % (self.id, lang))
1049
1050     def short_html(self):
1051         if self.id:
1052             cache_key = "Fragment.short_html/%d/%s" % (self.id, get_language())
1053             short_html = cache.get(cache_key)
1054         else:
1055             short_html = None
1056
1057         if short_html is not None:
1058             return mark_safe(short_html)
1059         else:
1060             short_html = unicode(render_to_string('catalogue/fragment_short.html',
1061                 {'fragment': self}))
1062             if self.id:
1063                 cache.set(cache_key, short_html, CACHE_FOREVER)
1064             return mark_safe(short_html)
1065
1066
1067 ###########
1068 #
1069 # SIGNALS
1070 #
1071 ###########
1072
1073
1074 def _tags_updated_handler(sender, affected_tags, **kwargs):
1075     # reset tag global counter
1076     # we want Tag.changed_at updated for API to know the tag was touched
1077     for tag in affected_tags:
1078         touch_tag.delay(tag)
1079
1080     # if book tags changed, reset book tag counter
1081     if isinstance(sender, Book) and \
1082                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1083                     exclude(category__in=('book', 'theme', 'set')).count():
1084         sender.reset_tag_counter()
1085     # if fragment theme changed, reset book theme counter
1086     elif isinstance(sender, Fragment) and \
1087                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1088                     filter(category='theme').count():
1089         sender.book.reset_theme_counter()
1090 tags_updated.connect(_tags_updated_handler)
1091
1092
1093 def _pre_delete_handler(sender, instance, **kwargs):
1094     """ refresh Book on BookMedia delete """
1095     if sender == BookMedia:
1096         instance.book.save()
1097 pre_delete.connect(_pre_delete_handler)
1098
1099 def _post_save_handler(sender, instance, **kwargs):
1100     """ refresh all the short_html stuff on BookMedia update """
1101     if sender == BookMedia:
1102         instance.book.save()
1103 post_save.connect(_post_save_handler)