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 datetime import datetime
   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 shutil import copy
  28 from glob import glob
  29 import re
  30 from os import path
  31
  32
  33 TAG_CATEGORIES = (
  34     ('author', _('author')),
  35     ('epoch', _('epoch')),
  36     ('kind', _('kind')),
  37     ('genre', _('genre')),
  38     ('theme', _('theme')),
  39     ('set', _('set')),
  40     ('book', _('book')),
  41 )
  42
  43 MEDIA_FORMATS = (
  44     ('odt', _('ODT file')),
  45     ('mp3', _('MP3 file')),
  46     ('ogg', _('OGG file')),
  47     ('daisy', _('DAISY file')),
  48 )
  49
  50 # not quite, but Django wants you to set a timeout
  51 CACHE_FOREVER = 2419200  # 28 days
  52
  53
  54 class TagSubcategoryManager(models.Manager):
  55     def __init__(self, subcategory):
  56         super(TagSubcategoryManager, self).__init__()
  57         self.subcategory = subcategory
  58
  59     def get_query_set(self):
  60         return super(TagSubcategoryManager, self).get_query_set().filter(category=self.subcategory)
  61
  62
  63 class Tag(TagBase):
  64     name = models.CharField(_('name'), max_length=50, db_index=True)
  65     slug = models.SlugField(_('slug'), max_length=120, db_index=True)
  66     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True)
  67     category = models.CharField(_('category'), max_length=50, blank=False, null=False,
  68         db_index=True, choices=TAG_CATEGORIES)
  69     description = models.TextField(_('description'), blank=True)
  70     main_page = models.BooleanField(_('main page'), default=False, db_index=True, help_text=_('Show tag on main page'))
  71
  72     user = models.ForeignKey(User, blank=True, null=True)
  73     book_count = models.IntegerField(_('book count'), blank=True, null=True)
  74     gazeta_link = models.CharField(blank=True, max_length=240)
  75     wiki_link = models.CharField(blank=True, max_length=240)
  76
  77     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
  78     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
  79
  80     class UrlDeprecationWarning(DeprecationWarning):
  81         pass
  82
  83     categories_rev = {
  84         'autor': 'author',
  85         'epoka': 'epoch',
  86         'rodzaj': 'kind',
  87         'gatunek': 'genre',
  88         'motyw': 'theme',
  89         'polka': 'set',
  90     }
  91     categories_dict = dict((item[::-1] for item in categories_rev.iteritems()))
  92
  93     class Meta:
  94         ordering = ('sort_key',)
  95         verbose_name = _('tag')
  96         verbose_name_plural = _('tags')
  97         unique_together = (("slug", "category"),)
  98
  99     def __unicode__(self):
 100         return self.name
 101
 102     def __repr__(self):
 103         return "Tag(slug=%r)" % self.slug
 104
 105     @permalink
 106     def get_absolute_url(self):
 107         return ('catalogue.views.tagged_object_list', [self.url_chunk])
 108
 109     def has_description(self):
 110         return len(self.description) > 0
 111     has_description.short_description = _('description')
 112     has_description.boolean = True
 113
 114     def get_count(self):
 115         """ returns global book count for book tags, fragment count for themes """
 116
 117         if self.book_count is None:
 118             if self.category == 'book':
 119                 # never used
 120                 objects = Book.objects.none()
 121             elif self.category == 'theme':
 122                 objects = Fragment.tagged.with_all((self,))
 123             else:
 124                 objects = Book.tagged.with_all((self,)).order_by()
 125                 if self.category != 'set':
 126                     # eliminate descendants
 127                     l_tags = Tag.objects.filter(slug__in=[book.book_tag_slug() for book in objects])
 128                     descendants_keys = [book.pk for book in Book.tagged.with_any(l_tags)]
 129                     if descendants_keys:
 130                         objects = objects.exclude(pk__in=descendants_keys)
 131             self.book_count = objects.count()
 132             self.save()
 133         return self.book_count
 134
 135     @staticmethod
 136     def get_tag_list(tags):
 137         if isinstance(tags, basestring):
 138             real_tags = []
 139             ambiguous_slugs = []
 140             category = None
 141             deprecated = False
 142             tags_splitted = tags.split('/')
 143             for name in tags_splitted:
 144                 if category:
 145                     real_tags.append(Tag.objects.get(slug=name, category=category))
 146                     category = None
 147                 elif name in Tag.categories_rev:
 148                     category = Tag.categories_rev[name]
 149                 else:
 150                     try:
 151                         real_tags.append(Tag.objects.exclude(category='book').get(slug=name))
 152                         deprecated = True
 153                     except Tag.MultipleObjectsReturned, e:
 154                         ambiguous_slugs.append(name)
 155
 156             if category:
 157                 # something strange left off
 158                 raise Tag.DoesNotExist()
 159             if ambiguous_slugs:
 160                 # some tags should be qualified
 161                 e = Tag.MultipleObjectsReturned()
 162                 e.tags = real_tags
 163                 e.ambiguous_slugs = ambiguous_slugs
 164                 raise e
 165             if deprecated:
 166                 e = Tag.UrlDeprecationWarning()
 167                 e.tags = real_tags
 168                 raise e
 169             return real_tags
 170         else:
 171             return TagBase.get_tag_list(tags)
 172
 173     @property
 174     def url_chunk(self):
 175         return '/'.join((Tag.categories_dict[self.category], self.slug))
 176
 177
 178 def get_dynamic_path(media, filename, ext=None, maxlen=100):
 179     from slughifi import slughifi
 180
 181     # how to put related book's slug here?
 182     if not ext:
 183         if media.type == 'daisy':
 184             ext = 'daisy.zip'
 185         else:
 186             ext = media.type
 187     if media is None or not media.name:
 188         name = slughifi(filename.split(".")[0])
 189     else:
 190         name = slughifi(media.name)
 191     return 'book/%s/%s.%s' % (ext, name[:maxlen-len('book/%s/.%s' % (ext, ext))-4], ext)
 192
 193
 194 # TODO: why is this hard-coded ?
 195 def book_upload_path(ext=None, maxlen=100):
 196     return lambda *args: get_dynamic_path(*args, ext=ext, maxlen=maxlen)
 197
 198
 199 def get_customized_pdf_path(book, customizations):
 200     """
 201     Returns a MEDIA_ROOT relative path for a customized pdf. The name will contain a hash of customization options.
 202     """
 203     customizations.sort()
 204     h = hash(tuple(customizations))
 205
 206     pdf_name = '%s-custom-%s' % (book.fileid(), h)
 207     pdf_file = get_dynamic_path(None, pdf_name, ext='pdf')
 208
 209     return pdf_file
 210
 211
 212 def get_existing_customized_pdf(book):
 213     """
 214     Returns a list of paths to generated customized pdf of a book
 215     """
 216     pdf_glob = '%s-custom-' % (book.fileid(),)
 217     pdf_glob = get_dynamic_path(None, pdf_glob, ext='pdf')
 218     pdf_glob = re.sub(r"[.]([a-z0-9]+)$", "*.\\1", pdf_glob)
 219     return glob(path.join(settings.MEDIA_ROOT, pdf_glob))
 220
 221
 222 class BookMedia(models.Model):
 223     type        = models.CharField(_('type'), choices=MEDIA_FORMATS, max_length="100")
 224     name        = models.CharField(_('name'), max_length="100")
 225     file        = OverwritingFileField(_('file'), upload_to=book_upload_path())
 226     uploaded_at = models.DateTimeField(_('creation date'), auto_now_add=True, editable=False)
 227     extra_info  = JSONField(_('extra information'), default='{}', editable=False)
 228     book = models.ForeignKey('Book', related_name='media')
 229     source_sha1 = models.CharField(null=True, blank=True, max_length=40, editable=False)
 230
 231     def __unicode__(self):
 232         return "%s (%s)" % (self.name, self.file.name.split("/")[-1])
 233
 234     class Meta:
 235         ordering            = ('type', 'name')
 236         verbose_name        = _('book media')
 237         verbose_name_plural = _('book media')
 238
 239     def save(self, *args, **kwargs):
 240         from slughifi import slughifi
 241         from catalogue.utils import ExistingFile, remove_zip
 242
 243         try:
 244             old = BookMedia.objects.get(pk=self.pk)
 245         except BookMedia.DoesNotExist, e:
 246             pass
 247         else:
 248             # if name changed, change the file name, too
 249             if slughifi(self.name) != slughifi(old.name):
 250                 self.file.save(None, ExistingFile(self.file.path), save=False, leave=True)
 251
 252         super(BookMedia, self).save(*args, **kwargs)
 253
 254         # remove the zip package for book with modified media
 255         remove_zip(self.book.fileid())
 256
 257         extra_info = self.get_extra_info_value()
 258         extra_info.update(self.read_meta())
 259         self.set_extra_info_value(extra_info)
 260         self.source_sha1 = self.read_source_sha1(self.file.path, self.type)
 261         return super(BookMedia, self).save(*args, **kwargs)
 262
 263     def read_meta(self):
 264         """
 265             Reads some metadata from the audiobook.
 266         """
 267         import mutagen
 268         from mutagen import id3
 269
 270         artist_name = director_name = project = funded_by = ''
 271         if self.type == 'mp3':
 272             try:
 273                 audio = id3.ID3(self.file.path)
 274                 artist_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE1'))
 275                 director_name = ', '.join(', '.join(tag.text) for tag in audio.getall('TPE3'))
 276                 project = ", ".join([t.data for t in audio.getall('PRIV')
 277                         if t.owner=='wolnelektury.pl?project'])
 278                 funded_by = ", ".join([t.data for t in audio.getall('PRIV')
 279                         if t.owner=='wolnelektury.pl?funded_by'])
 280             except:
 281                 pass
 282         elif self.type == 'ogg':
 283             try:
 284                 audio = mutagen.File(self.file.path)
 285                 artist_name = ', '.join(audio.get('artist', []))
 286                 director_name = ', '.join(audio.get('conductor', []))
 287                 project = ", ".join(audio.get('project', []))
 288                 funded_by = ", ".join(audio.get('funded_by', []))
 289             except:
 290                 pass
 291         else:
 292             return {}
 293         return {'artist_name': artist_name, 'director_name': director_name,
 294                 'project': project, 'funded_by': funded_by}
 295
 296     @staticmethod
 297     def read_source_sha1(filepath, filetype):
 298         """
 299             Reads source file SHA1 from audiobok metadata.
 300         """
 301         import mutagen
 302         from mutagen import id3
 303
 304         if filetype == 'mp3':
 305             try:
 306                 audio = id3.ID3(filepath)
 307                 return [t.data for t in audio.getall('PRIV')
 308                         if t.owner=='wolnelektury.pl?flac_sha1'][0]
 309             except:
 310                 return None
 311         elif filetype == 'ogg':
 312             try:
 313                 audio = mutagen.File(filepath)
 314                 return audio.get('flac_sha1', [None])[0]
 315             except:
 316                 return None
 317         else:
 318             return None
 319
 320
 321 class Book(models.Model):
 322     title         = models.CharField(_('title'), max_length=120)
 323     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True, editable=False)
 324     slug          = models.SlugField(_('slug'), max_length=120, db_index=True)
 325     language = models.CharField(_('language code'), max_length=3, db_index=True,
 326                     default=settings.CATALOGUE_DEFAULT_LANGUAGE)
 327     description   = models.TextField(_('description'), blank=True)
 328     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
 329     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
 330     parent_number = models.IntegerField(_('parent number'), default=0)
 331     extra_info    = JSONField(_('extra information'), default='{}')
 332     gazeta_link   = models.CharField(blank=True, max_length=240)
 333     wiki_link     = models.CharField(blank=True, max_length=240)
 334     # files generated during publication
 335
 336     file_types = ['epub', 'html', 'mobi', 'pdf', 'txt', 'xml']
 337
 338     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
 339     objects  = models.Manager()
 340     tagged   = managers.ModelTaggedItemManager(Tag)
 341     tags     = managers.TagDescriptor(Tag)
 342
 343     html_built = django.dispatch.Signal()
 344     published = django.dispatch.Signal()
 345
 346     URLID_RE = r'[a-z0-9-]+(?:/[a-z]{3})?'
 347     FILEID_RE = r'[a-z0-9-]+(?:_[a-z]{3})?'
 348
 349     class AlreadyExists(Exception):
 350         pass
 351
 352     class Meta:
 353         unique_together = [['slug', 'language']]
 354         ordering = ('sort_key',)
 355         verbose_name = _('book')
 356         verbose_name_plural = _('books')
 357
 358     def __unicode__(self):
 359         return self.title
 360
 361     def urlid(self, sep='/'):
 362         stem = self.slug
 363         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 364             stem += sep + self.language
 365         return stem
 366
 367     def fileid(self):
 368         return self.urlid('_')
 369
 370     @staticmethod
 371     def split_urlid(urlid, sep='/', default_lang=settings.CATALOGUE_DEFAULT_LANGUAGE):
 372         """Splits a URL book id into slug and language code.
 373
 374         Returns a dictionary usable i.e. for object lookup, or None.
 375
 376         >>> Book.split_urlid("a-slug/pol", default_lang="eng")
 377         {'slug': 'a-slug', 'language': 'pol'}
 378         >>> Book.split_urlid("a-slug", default_lang="eng")
 379         {'slug': 'a-slug', 'language': 'eng'}
 380         >>> Book.split_urlid("a-slug_pol", "_", default_lang="eng")
 381         {'slug': 'a-slug', 'language': 'pol'}
 382         >>> Book.split_urlid("a-slug/eng", default_lang="eng")
 383
 384         """
 385         parts = urlid.rsplit(sep, 1)
 386         if len(parts) == 2:
 387             if parts[1] == default_lang:
 388                 return None
 389             return {'slug': parts[0], 'language': parts[1]}
 390         else:
 391             return {'slug': urlid, 'language': default_lang}
 392
 393     @classmethod
 394     def split_fileid(cls, fileid):
 395         return cls.split_urlid(fileid, '_')
 396
 397     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
 398         from sortify import sortify
 399
 400         self.sort_key = sortify(self.title)
 401
 402         ret = super(Book, self).save(force_insert, force_update)
 403
 404         if reset_short_html:
 405             self.reset_short_html()
 406
 407         return ret
 408
 409     @permalink
 410     def get_absolute_url(self):
 411         return ('catalogue.views.book_detail', [self.urlid()])
 412
 413     @property
 414     def name(self):
 415         return self.title
 416
 417     def book_tag_slug(self):
 418         stem = 'l-' + self.slug
 419         if self.language != settings.CATALOGUE_DEFAULT_LANGUAGE:
 420             return stem[:116] + ' ' + self.language
 421         else:
 422             return stem[:120]
 423
 424     def book_tag(self):
 425         slug = self.book_tag_slug()
 426         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
 427         if created:
 428             book_tag.name = self.title[:50]
 429             book_tag.sort_key = self.title.lower()
 430             book_tag.save()
 431         return book_tag
 432
 433     def has_media(self, type):
 434         if type in Book.file_types:
 435             return bool(getattr(self, "%s_file" % type))
 436         else:
 437             return self.media.filter(type=type).exists()
 438
 439     def get_media(self, type):
 440         if self.has_media(type):
 441             if type in Book.file_types:
 442                 return getattr(self, "%s_file" % type)
 443             else:
 444                 return self.media.filter(type=type)
 445         else:
 446             return None
 447
 448     def get_mp3(self):
 449         return self.get_media("mp3")
 450     def get_odt(self):
 451         return self.get_media("odt")
 452     def get_ogg(self):
 453         return self.get_media("ogg")
 454     def get_daisy(self):
 455         return self.get_media("daisy")
 456
 457     def reset_short_html(self):
 458         if self.id is None:
 459             return
 460
 461         cache_key = "Book.short_html/%d/%s"
 462         for lang, langname in settings.LANGUAGES:
 463             cache.delete(cache_key % (self.id, lang))
 464         # Fragment.short_html relies on book's tags, so reset it here too
 465         for fragm in self.fragments.all():
 466             fragm.reset_short_html()
 467
 468     def short_html(self):
 469         if self.id:
 470             cache_key = "Book.short_html/%d/%s" % (self.id, get_language())
 471             short_html = cache.get(cache_key)
 472         else:
 473             short_html = None
 474
 475         if short_html is not None:
 476             return mark_safe(short_html)
 477         else:
 478             tags = self.tags.filter(~Q(category__in=('set', 'theme', 'book')))
 479             tags = [mark_safe(u'<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name)) for tag in tags]
 480
 481             formats = []
 482             # files generated during publication
 483             if self.has_media("html"):
 484                 formats.append(u'<a href="%s">%s</a>' % (reverse('book_text', [self.fileid()]), _('Read online')))
 485             if self.has_media("pdf"):
 486                 formats.append(u'<a href="%s">PDF</a>' % self.get_media('pdf').url)
 487             if self.has_media("mobi"):
 488                 formats.append(u'<a href="%s">MOBI</a>' % self.get_media('mobi').url)
 489             if self.root_ancestor.has_media("epub"):
 490                 formats.append(u'<a href="%s">EPUB</a>' % self.root_ancestor.get_media('epub').url)
 491             if self.has_media("txt"):
 492                 formats.append(u'<a href="%s">TXT</a>' % self.get_media('txt').url)
 493             # other files
 494             for m in self.media.order_by('type'):
 495                 formats.append(u'<a href="%s">%s</a>' % (m.file.url, m.type.upper()))
 496
 497             formats = [mark_safe(format) for format in formats]
 498
 499             short_html = unicode(render_to_string('catalogue/book_short.html',
 500                 {'book': self, 'tags': tags, 'formats': formats}))
 501
 502             if self.id:
 503                 cache.set(cache_key, short_html, CACHE_FOREVER)
 504             return mark_safe(short_html)
 505
 506     @property
 507     def root_ancestor(self):
 508         """ returns the oldest ancestor """
 509
 510         if not hasattr(self, '_root_ancestor'):
 511             book = self
 512             while book.parent:
 513                 book = book.parent
 514             self._root_ancestor = book
 515         return self._root_ancestor
 516
 517
 518     def has_description(self):
 519         return len(self.description) > 0
 520     has_description.short_description = _('description')
 521     has_description.boolean = True
 522
 523     # ugly ugly ugly
 524     def has_odt_file(self):
 525         return bool(self.has_media("odt"))
 526     has_odt_file.short_description = 'ODT'
 527     has_odt_file.boolean = True
 528
 529     def has_mp3_file(self):
 530         return bool(self.has_media("mp3"))
 531     has_mp3_file.short_description = 'MP3'
 532     has_mp3_file.boolean = True
 533
 534     def has_ogg_file(self):
 535         return bool(self.has_media("ogg"))
 536     has_ogg_file.short_description = 'OGG'
 537     has_ogg_file.boolean = True
 538
 539     def has_daisy_file(self):
 540         return bool(self.has_media("daisy"))
 541     has_daisy_file.short_description = 'DAISY'
 542     has_daisy_file.boolean = True
 543
 544     def wldocument(self, parse_dublincore=True):
 545         from catalogue.utils import ORMDocProvider
 546         from librarian.parser import WLDocument
 547
 548         return WLDocument.from_file(self.xml_file.path,
 549                 provider=ORMDocProvider(self),
 550                 parse_dublincore=parse_dublincore)
 551
 552     def build_pdf(self, customizations=None, file_name=None):
 553         """ (Re)builds the pdf file.
 554         customizations - customizations which are passed to LaTeX class file.
 555         file_name - save the pdf file under a different name and DO NOT save it in db.
 556         """
 557         from os import unlink
 558         from django.core.files import File
 559         from catalogue.utils import remove_zip
 560
 561         pdf = self.wldocument().as_pdf(customizations=customizations)
 562
 563         if file_name is None:
 564             # we'd like to be sure not to overwrite changes happening while
 565             # (timely) pdf generation is taking place (async celery scenario)
 566             current_self = Book.objects.get(id=self.id)
 567             current_self.pdf_file.save('%s.pdf' % self.fileid(),
 568                     File(open(pdf.get_filename())))
 569             self.pdf_file = current_self.pdf_file
 570
 571             # remove cached downloadables
 572             remove_zip(settings.ALL_PDF_ZIP)
 573
 574             for customized_pdf in get_existing_customized_pdf(self):
 575                 unlink(customized_pdf)
 576         else:
 577             print "saving %s" % file_name
 578             print "to: %s" % DefaultStorage().path(file_name)
 579             DefaultStorage().save(file_name, File(open(pdf.get_filename())))
 580
 581     def build_mobi(self):
 582         """ (Re)builds the MOBI file.
 583
 584         """
 585         from django.core.files import File
 586         from catalogue.utils import remove_zip
 587
 588         mobi = self.wldocument().as_mobi()
 589
 590         self.mobi_file.save('%s.mobi' % self.fileid(), File(open(mobi.get_filename())))
 591
 592         # remove zip with all mobi files
 593         remove_zip(settings.ALL_MOBI_ZIP)
 594
 595     def build_epub(self, remove_descendants=True):
 596         """ (Re)builds the epub file.
 597             If book has a parent, does nothing.
 598             Unless remove_descendants is False, descendants' epubs are removed.
 599         """
 600         from django.core.files import File
 601         from catalogue.utils import remove_zip
 602
 603         if self.parent:
 604             # don't need an epub
 605             return
 606
 607         epub = self.wldocument().as_epub()
 608
 609         try:
 610             epub.transform(ORMDocProvider(self), self.fileid(), output_file=epub_file)
 611             self.epub_file.save('%s.epub' % self.fileid(), File(open(epub.get_filename())))
 612         except NoDublinCore:
 613             pass
 614
 615         book_descendants = list(self.children.all())
 616         while len(book_descendants) > 0:
 617             child_book = book_descendants.pop(0)
 618             if remove_descendants and child_book.has_epub_file():
 619                 child_book.epub_file.delete()
 620             # save anyway, to refresh short_html
 621             child_book.save()
 622             book_descendants += list(child_book.children.all())
 623
 624         # remove zip package with all epub files
 625         remove_zip(settings.ALL_EPUB_ZIP)
 626
 627     def build_txt(self):
 628         from django.core.files.base import ContentFile
 629
 630         text = self.wldocument().as_text()
 631         self.txt_file.save('%s.txt' % self.fileid(), ContentFile(text.get_string()))
 632
 633
 634     def build_html(self):
 635         from markupstring import MarkupString
 636         from django.core.files.base import ContentFile
 637         from slughifi import slughifi
 638         from librarian import html
 639
 640         meta_tags = list(self.tags.filter(
 641             category__in=('author', 'epoch', 'genre', 'kind')))
 642         book_tag = self.book_tag()
 643
 644         html_output = self.wldocument(parse_dublincore=False).as_html()
 645         if html_output:
 646             self.html_file.save('%s.html' % self.fileid(),
 647                     ContentFile(html_output.get_string()))
 648
 649             # get ancestor l-tags for adding to new fragments
 650             ancestor_tags = []
 651             p = self.parent
 652             while p:
 653                 ancestor_tags.append(p.book_tag())
 654                 p = p.parent
 655
 656             # Delete old fragments and create them from scratch
 657             self.fragments.all().delete()
 658             # Extract fragments
 659             closed_fragments, open_fragments = html.extract_fragments(self.html_file.path)
 660             for fragment in closed_fragments.values():
 661                 try:
 662                     theme_names = [s.strip() for s in fragment.themes.split(',')]
 663                 except AttributeError:
 664                     continue
 665                 themes = []
 666                 for theme_name in theme_names:
 667                     if not theme_name:
 668                         continue
 669                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
 670                     if created:
 671                         tag.name = theme_name
 672                         tag.sort_key = theme_name.lower()
 673                         tag.save()
 674                     themes.append(tag)
 675                 if not themes:
 676                     continue
 677
 678                 text = fragment.to_string()
 679                 short_text = ''
 680                 if (len(MarkupString(text)) > 240):
 681                     short_text = unicode(MarkupString(text)[:160])
 682                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=self,
 683                     text=text, short_text=short_text)
 684
 685                 new_fragment.save()
 686                 new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
 687             self.save()
 688             self.html_built.send(sender=self)
 689             return True
 690         return False
 691
 692     @staticmethod
 693     def zip_format(format_):
 694         def pretty_file_name(book):
 695             return "%s/%s.%s" % (
 696                 b.get_extra_info_value()['author'],
 697                 b.fileid(),
 698                 format_)
 699
 700         field_name = "%s_file" % format_
 701         books = Book.objects.filter(parent=None).exclude(**{field_name: ""})
 702         paths = [(pretty_file_name(b), getattr(b, field_name).path)
 703                     for b in books]
 704         result = create_zip.delay(paths,
 705                     getattr(settings, "ALL_%s_ZIP" % format_.upper()))
 706         return result.wait()
 707
 708     def zip_audiobooks(self):
 709         bm = BookMedia.objects.filter(book=self, type='mp3')
 710         paths = map(lambda bm: (None, bm.file.path), bm)
 711         result = create_zip.delay(paths, self.fileid())
 712         return result.wait()
 713
 714     @classmethod
 715     def from_xml_file(cls, xml_file, **kwargs):
 716         from django.core.files import File
 717         from librarian import dcparser
 718
 719         # use librarian to parse meta-data
 720         book_info = dcparser.parse(xml_file)
 721
 722         if not isinstance(xml_file, File):
 723             xml_file = File(open(xml_file))
 724
 725         try:
 726             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
 727         finally:
 728             xml_file.close()
 729
 730     @classmethod
 731     def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
 732             build_epub=True, build_txt=True, build_pdf=True, build_mobi=True):
 733         import re
 734         from slughifi import slughifi
 735         from sortify import sortify
 736
 737         # check for parts before we do anything
 738         children = []
 739         if hasattr(book_info, 'parts'):
 740             for part_url in book_info.parts:
 741                 try:
 742                     children.append(Book.objects.get(
 743                         slug=part_url.slug, language=part_url.language))
 744                 except Book.DoesNotExist, e:
 745                     raise Book.DoesNotExist(_('Book "%s/%s" does not exist.') %
 746                             (part_url.slug, part_url.language))
 747
 748
 749         # Read book metadata
 750         book_slug = book_info.url.slug
 751         language = book_info.language
 752         if re.search(r'[^a-zA-Z0-9-]', book_slug):
 753             raise ValueError('Invalid characters in slug')
 754         book, created = Book.objects.get_or_create(slug=book_slug, language=language)
 755
 756         if created:
 757             book_shelves = []
 758         else:
 759             if not overwrite:
 760                 raise Book.AlreadyExists(_('Book %s/%s already exists') % (
 761                         book_slug, language))
 762             # Save shelves for this book
 763             book_shelves = list(book.tags.filter(category='set'))
 764
 765         book.title = book_info.title
 766         book.set_extra_info_value(book_info.to_dict())
 767         book.save()
 768
 769         meta_tags = []
 770         categories = (('kinds', 'kind'), ('genres', 'genre'), ('authors', 'author'), ('epochs', 'epoch'))
 771         for field_name, category in categories:
 772             try:
 773                 tag_names = getattr(book_info, field_name)
 774             except:
 775                 tag_names = [getattr(book_info, category)]
 776             for tag_name in tag_names:
 777                 tag_sort_key = tag_name
 778                 if category == 'author':
 779                     tag_sort_key = tag_name.last_name
 780                     tag_name = ' '.join(tag_name.first_names) + ' ' + tag_name.last_name
 781                 tag, created = Tag.objects.get_or_create(slug=slughifi(tag_name), category=category)
 782                 if created:
 783                     tag.name = tag_name
 784                     tag.sort_key = sortify(tag_sort_key.lower())
 785                     tag.save()
 786                 meta_tags.append(tag)
 787
 788         book.tags = set(meta_tags + book_shelves)
 789
 790         book_tag = book.book_tag()
 791
 792         for n, child_book in enumerate(children):
 793             child_book.parent = book
 794             child_book.parent_number = n
 795             child_book.save()
 796
 797         # Save XML and HTML files
 798         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
 799
 800         # delete old fragments when overwriting
 801         book.fragments.all().delete()
 802
 803         if book.build_html():
 804             if not settings.NO_BUILD_TXT and build_txt:
 805                 book.build_txt()
 806
 807         if not settings.NO_BUILD_EPUB and build_epub:
 808             book.root_ancestor.build_epub()
 809
 810         if not settings.NO_BUILD_PDF and build_pdf:
 811             book.root_ancestor.build_pdf()
 812
 813         if not settings.NO_BUILD_MOBI and build_mobi:
 814             book.build_mobi()
 815
 816         book_descendants = list(book.children.all())
 817         # add l-tag to descendants and their fragments
 818         # delete unnecessary EPUB files
 819         while len(book_descendants) > 0:
 820             child_book = book_descendants.pop(0)
 821             child_book.tags = list(child_book.tags) + [book_tag]
 822             child_book.save()
 823             for fragment in child_book.fragments.all():
 824                 fragment.tags = set(list(fragment.tags) + [book_tag])
 825             book_descendants += list(child_book.children.all())
 826
 827         book.save()
 828
 829         # refresh cache
 830         book.reset_tag_counter()
 831         book.reset_theme_counter()
 832
 833         cls.published.send(sender=book)
 834         return book
 835
 836     def reset_tag_counter(self):
 837         if self.id is None:
 838             return
 839
 840         cache_key = "Book.tag_counter/%d" % self.id
 841         cache.delete(cache_key)
 842         if self.parent:
 843             self.parent.reset_tag_counter()
 844
 845     @property
 846     def tag_counter(self):
 847         if self.id:
 848             cache_key = "Book.tag_counter/%d" % self.id
 849             tags = cache.get(cache_key)
 850         else:
 851             tags = None
 852
 853         if tags is None:
 854             tags = {}
 855             for child in self.children.all().order_by():
 856                 for tag_pk, value in child.tag_counter.iteritems():
 857                     tags[tag_pk] = tags.get(tag_pk, 0) + value
 858             for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by():
 859                 tags[tag.pk] = 1
 860
 861             if self.id:
 862                 cache.set(cache_key, tags, CACHE_FOREVER)
 863         return tags
 864
 865     def reset_theme_counter(self):
 866         if self.id is None:
 867             return
 868
 869         cache_key = "Book.theme_counter/%d" % self.id
 870         cache.delete(cache_key)
 871         if self.parent:
 872             self.parent.reset_theme_counter()
 873
 874     @property
 875     def theme_counter(self):
 876         if self.id:
 877             cache_key = "Book.theme_counter/%d" % self.id
 878             tags = cache.get(cache_key)
 879         else:
 880             tags = None
 881
 882         if tags is None:
 883             tags = {}
 884             for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by():
 885                 for tag in fragment.tags.filter(category='theme').order_by():
 886                     tags[tag.pk] = tags.get(tag.pk, 0) + 1
 887
 888             if self.id:
 889                 cache.set(cache_key, tags, CACHE_FOREVER)
 890         return tags
 891
 892     def pretty_title(self, html_links=False):
 893         book = self
 894         names = list(book.tags.filter(category='author'))
 895
 896         books = []
 897         while book:
 898             books.append(book)
 899             book = book.parent
 900         names.extend(reversed(books))
 901
 902         if html_links:
 903             names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
 904         else:
 905             names = [tag.name for tag in names]
 906
 907         return ', '.join(names)
 908
 909     @classmethod
 910     def tagged_top_level(cls, tags):
 911         """ Returns top-level books tagged with `tags'.
 912
 913         It only returns those books which don't have ancestors which are
 914         also tagged with those tags.
 915
 916         """
 917         # get relevant books and their tags
 918         objects = cls.tagged.with_all(tags)
 919         # eliminate descendants
 920         l_tags = Tag.objects.filter(category='book', slug__in=[book.book_tag_slug() for book in objects])
 921         descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags)]
 922         if descendants_keys:
 923             objects = objects.exclude(pk__in=descendants_keys)
 924
 925         return objects
 926
 927     @classmethod
 928     def book_list(cls, filter=None):
 929         """Generates a hierarchical listing of all books.
 930
 931         Books are optionally filtered with a test function.
 932
 933         """
 934
 935         books_by_parent = {}
 936         books = cls.objects.all().order_by('parent_number', 'sort_key').only(
 937                 'title', 'parent', 'slug', 'language')
 938         if filter:
 939             books = books.filter(filter).distinct()
 940             book_ids = set((book.pk for book in books))
 941             for book in books:
 942                 parent = book.parent_id
 943                 if parent not in book_ids:
 944                     parent = None
 945                 books_by_parent.setdefault(parent, []).append(book)
 946         else:
 947             for book in books:
 948                 books_by_parent.setdefault(book.parent_id, []).append(book)
 949
 950         orphans = []
 951         books_by_author = SortedDict()
 952         for tag in Tag.objects.filter(category='author'):
 953             books_by_author[tag] = []
 954
 955         for book in books_by_parent.get(None,()):
 956             authors = list(book.tags.filter(category='author'))
 957             if authors:
 958                 for author in authors:
 959                     books_by_author[author].append(book)
 960             else:
 961                 orphans.append(book)
 962
 963         return books_by_author, orphans, books_by_parent
 964
 965     _audiences_pl = {
 966         "SP1": (1, u"szkoła podstawowa"),
 967         "SP2": (1, u"szkoła podstawowa"),
 968         "P": (1, u"szkoła podstawowa"),
 969         "G": (2, u"gimnazjum"),
 970         "L": (3, u"liceum"),
 971         "LP": (3, u"liceum"),
 972     }
 973     def audiences_pl(self):
 974         audiences = self.get_extra_info_value().get('audiences', [])
 975         audiences = sorted(set([self._audiences_pl[a] for a in audiences]))
 976         return [a[1] for a in audiences]
 977
 978
 979 def _has_factory(ftype):
 980     has = lambda self: bool(getattr(self, "%s_file" % ftype))
 981     has.short_description = t.upper()
 982     has.boolean = True
 983     has.__name__ = "has_%s_file" % ftype
 984     return has
 985
 986
 987 # add the file fields
 988 for t in Book.file_types:
 989     field_name = "%s_file" % t
 990     models.FileField(_("%s file" % t.upper()),
 991             upload_to=book_upload_path(t),
 992             blank=True).contribute_to_class(Book, field_name)
 993
 994     setattr(Book, "has_%s_file" % t, _has_factory(t))
 995
 996
 997 class Fragment(models.Model):
 998     text = models.TextField()
 999     short_text = models.TextField(editable=False)
1000     anchor = models.CharField(max_length=120)
1001     book = models.ForeignKey(Book, related_name='fragments')
1002
1003     objects = models.Manager()
1004     tagged = managers.ModelTaggedItemManager(Tag)
1005     tags = managers.TagDescriptor(Tag)
1006
1007     class Meta:
1008         ordering = ('book', 'anchor',)
1009         verbose_name = _('fragment')
1010         verbose_name_plural = _('fragments')
1011
1012     def get_absolute_url(self):
1013         return '%s#m%s' % (self.book.get_html_url(), self.anchor)
1014
1015     def reset_short_html(self):
1016         if self.id is None:
1017             return
1018
1019         cache_key = "Fragment.short_html/%d/%s"
1020         for lang, langname in settings.LANGUAGES:
1021             cache.delete(cache_key % (self.id, lang))
1022
1023     def short_html(self):
1024         if self.id:
1025             cache_key = "Fragment.short_html/%d/%s" % (self.id, get_language())
1026             short_html = cache.get(cache_key)
1027         else:
1028             short_html = None
1029
1030         if short_html is not None:
1031             return mark_safe(short_html)
1032         else:
1033             short_html = unicode(render_to_string('catalogue/fragment_short.html',
1034                 {'fragment': self}))
1035             if self.id:
1036                 cache.set(cache_key, short_html, CACHE_FOREVER)
1037             return mark_safe(short_html)
1038
1039
1040 ###########
1041 #
1042 # SIGNALS
1043 #
1044 ###########
1045
1046
1047 def _tags_updated_handler(sender, affected_tags, **kwargs):
1048     # reset tag global counter
1049     # we want Tag.changed_at updated for API to know the tag was touched
1050     Tag.objects.filter(pk__in=[tag.pk for tag in affected_tags]).update(book_count=None, changed_at=datetime.now())
1051
1052     # if book tags changed, reset book tag counter
1053     if isinstance(sender, Book) and \
1054                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1055                     exclude(category__in=('book', 'theme', 'set')).count():
1056         sender.reset_tag_counter()
1057     # if fragment theme changed, reset book theme counter
1058     elif isinstance(sender, Fragment) and \
1059                 Tag.objects.filter(pk__in=(tag.pk for tag in affected_tags)).\
1060                     filter(category='theme').count():
1061         sender.book.reset_theme_counter()
1062 tags_updated.connect(_tags_updated_handler)
1063
1064
1065 def _pre_delete_handler(sender, instance, **kwargs):
1066     """ refresh Book on BookMedia delete """
1067     if sender == BookMedia:
1068         instance.book.save()
1069 pre_delete.connect(_pre_delete_handler)
1070
1071 def _post_save_handler(sender, instance, **kwargs):
1072     """ refresh all the short_html stuff on BookMedia update """
1073     if sender == BookMedia:
1074         instance.book.save()
1075 post_save.connect(_post_save_handler)