35d0e8c4ea522fcd8bb5e35130e81be191d840f9
[wolnelektury.git] / src / catalogue / models / book.py
1 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
2 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
3 #
4 from collections import OrderedDict
5 import json
6 from datetime import date, timedelta
7 from random import randint
8 import os.path
9 import re
10 from urllib.request import urlretrieve
11 from django.apps import apps
12 from django.conf import settings
13 from django.db import connection, models, transaction
14 import django.dispatch
15 from django.contrib.contenttypes.fields import GenericRelation
16 from django.template.loader import render_to_string
17 from django.urls import reverse
18 from django.utils.translation import gettext_lazy as _, get_language
19 from fnpdjango.storage import BofhFileSystemStorage
20 from lxml import html
21 from librarian.cover import WLCover
22 from librarian.html import transform_abstrakt
23 from librarian.builders import builders
24 from newtagging import managers
25 from catalogue import constants
26 from catalogue import fields
27 from catalogue.models import Tag, Fragment, BookMedia
28 from catalogue.utils import create_zip, gallery_url, gallery_path, split_tags, get_random_hash
29 from catalogue.models.tag import prefetched_relations
30 from catalogue import app_settings
31 from wolnelektury.utils import makedirs, cached_render, clear_cached_renders
32
33 bofh_storage = BofhFileSystemStorage()
34
35
36 class Book(models.Model):
37     """Represents a book imported from WL-XML."""
38     title = models.CharField('tytuł', max_length=32767)
39     sort_key = models.CharField('klucz sortowania', max_length=120, db_index=True, editable=False)
40     sort_key_author = models.CharField(
41         'klucz sortowania wg autora', max_length=120, db_index=True, editable=False, default='')
42     slug = models.SlugField('slug', max_length=120, db_index=True, unique=True)
43     common_slug = models.SlugField('wspólny slug', max_length=120, db_index=True)
44     language = models.CharField('kod języka', max_length=3, db_index=True, default=app_settings.DEFAULT_LANGUAGE)
45     description = models.TextField('opis', blank=True)
46     abstract = models.TextField('abstrakt', blank=True)
47     toc = models.TextField('spis treści', blank=True)
48     created_at = models.DateTimeField('data utworzenia', auto_now_add=True, db_index=True)
49     changed_at = models.DateTimeField('data motyfikacji', auto_now=True, db_index=True)
50     parent_number = models.IntegerField('numer w ramach rodzica', default=0)
51     extra_info = models.TextField('dodatkowe informacje', default='{}')
52     gazeta_link = models.CharField(blank=True, max_length=240)
53     wiki_link = models.CharField(blank=True, max_length=240)
54     print_on_demand = models.BooleanField('druk na żądanie', default=False)
55     recommended = models.BooleanField('polecane', default=False)
56     audio_length = models.CharField('długość audio', blank=True, max_length=8)
57     preview = models.BooleanField('prapremiera', default=False)
58     preview_until = models.DateField('prapremiera do', blank=True, null=True)
59     preview_key = models.CharField(max_length=32, blank=True, null=True)
60     findable = models.BooleanField('wyszukiwalna', default=True, db_index=True)
61
62     # files generated during publication
63     xml_file = fields.XmlField(storage=bofh_storage, with_etag=False)
64     html_file = fields.HtmlField(storage=bofh_storage)
65     fb2_file = fields.Fb2Field(storage=bofh_storage)
66     txt_file = fields.TxtField(storage=bofh_storage)
67     epub_file = fields.EpubField(storage=bofh_storage)
68     mobi_file = fields.MobiField(storage=bofh_storage)
69     pdf_file = fields.PdfField(storage=bofh_storage)
70
71     cover = fields.CoverField('okładka', storage=bofh_storage)
72     # Cleaner version of cover for thumbs
73     cover_clean = fields.CoverCleanField('czysta okładka')
74     cover_thumb = fields.CoverThumbField('miniatura okładki')
75     cover_api_thumb = fields.CoverApiThumbField(
76         'mniaturka okładki dla aplikacji')
77     simple_cover = fields.SimpleCoverField('okładka dla aplikacji')
78     cover_ebookpoint = fields.CoverEbookpointField(
79         'okładka dla Ebookpoint')
80
81     ebook_formats = constants.EBOOK_FORMATS
82     formats = ebook_formats + ['html', 'xml']
83
84     parent = models.ForeignKey('self', models.CASCADE, blank=True, null=True, related_name='children')
85     ancestor = models.ManyToManyField('self', blank=True, editable=False, related_name='descendant', symmetrical=False)
86
87     cached_author = models.CharField(blank=True, max_length=240, db_index=True)
88     has_audience = models.BooleanField(default=False)
89
90     objects = models.Manager()
91     tagged = managers.ModelTaggedItemManager(Tag)
92     tags = managers.TagDescriptor(Tag)
93     tag_relations = GenericRelation(Tag.intermediary_table_model)
94
95     html_built = django.dispatch.Signal()
96     published = django.dispatch.Signal()
97
98     SORT_KEY_SEP = '$'
99
100     is_book = True
101
102     class AlreadyExists(Exception):
103         pass
104
105     class Meta:
106         ordering = ('sort_key_author', 'sort_key')
107         verbose_name = 'książka'
108         verbose_name_plural = 'książki'
109         app_label = 'catalogue'
110
111     def __str__(self):
112         return self.title
113
114     def get_extra_info_json(self):
115         return json.loads(self.extra_info or '{}')
116
117     def get_initial(self):
118         try:
119             return re.search(r'\w', self.title, re.U).group(0)
120         except AttributeError:
121             return ''
122
123     def authors(self):
124         return self.tags.filter(category='author')
125
126     def epochs(self):
127         return self.tags.filter(category='epoch')
128
129     def genres(self):
130         return self.tags.filter(category='genre')
131
132     def kinds(self):
133         return self.tags.filter(category='kind')
134
135     def tag_unicode(self, category):
136         relations = prefetched_relations(self, category)
137         if relations:
138             return ', '.join(rel.tag.name for rel in relations)
139         else:
140             return ', '.join(self.tags.filter(category=category).values_list('name', flat=True))
141
142     def tags_by_category(self):
143         return split_tags(self.tags.exclude(category__in=('set', 'theme')))
144
145     def author_unicode(self):
146         return self.cached_author
147
148     def kind_unicode(self):
149         return self.tag_unicode('kind')
150
151     def epoch_unicode(self):
152         return self.tag_unicode('epoch')
153
154     def genre_unicode(self):
155         return self.tag_unicode('genre')
156
157     def translators(self):
158         translators = self.get_extra_info_json().get('translators') or []
159         return [
160             '\xa0'.join(reversed(translator.split(', ', 1))) for translator in translators
161         ]
162
163     def translator(self):
164         translators = self.get_extra_info_json().get('translators')
165         if not translators:
166             return None
167         if len(translators) > 3:
168             translators = translators[:2]
169             others = ' i inni'
170         else:
171             others = ''
172         return ', '.join('\xa0'.join(reversed(translator.split(', ', 1))) for translator in translators) + others
173
174     def cover_source(self):
175         return self.get_extra_info_json().get('cover_source', self.parent.cover_source() if self.parent else '')
176
177     @property
178     def isbn_pdf(self):
179         return self.get_extra_info_json().get('isbn_pdf')
180
181     @property
182     def isbn_epub(self):
183         return self.get_extra_info_json().get('isbn_epub')
184
185     @property
186     def isbn_mobi(self):
187         return self.get_extra_info_json().get('isbn_mobi')
188
189     def is_accessible_to(self, user):
190         if not self.preview:
191             return True
192         if not user.is_authenticated:
193             return False
194         Membership = apps.get_model('club', 'Membership')
195         if Membership.is_active_for(user):
196             return True
197         Funding = apps.get_model('funding', 'Funding')
198         if Funding.objects.filter(user=user, offer__book=self):
199             return True
200         return False
201
202     def save(self, force_insert=False, force_update=False, **kwargs):
203         from sortify import sortify
204
205         self.sort_key = sortify(self.title)[:120]
206         self.title = str(self.title)  # ???
207
208         try:
209             author = self.authors().first().sort_key
210         except AttributeError:
211             author = ''
212         self.sort_key_author = author
213
214         self.cached_author = self.tag_unicode('author')
215         self.has_audience = 'audience' in self.get_extra_info_json()
216
217         if self.preview and not self.preview_key:
218             self.preview_key = get_random_hash(self.slug)[:32]
219
220         ret = super(Book, self).save(force_insert, force_update, **kwargs)
221
222         return ret
223
224     def get_absolute_url(self):
225         return reverse('book_detail', args=[self.slug])
226
227     def gallery_path(self):
228         return gallery_path(self.slug)
229
230     def gallery_url(self):
231         return gallery_url(self.slug)
232
233     def get_first_text(self):
234         if self.html_file:
235             return self
236         child = self.children.all().order_by('parent_number').first()
237         if child is not None:
238             return child.get_first_text()
239
240     def get_last_text(self):
241         if self.html_file:
242             return self
243         child = self.children.all().order_by('parent_number').last()
244         if child is not None:
245             return child.get_last_text()
246
247     def get_prev_text(self):
248         if not self.parent:
249             return None
250         sibling = self.parent.children.filter(parent_number__lt=self.parent_number).order_by('-parent_number').first()
251         if sibling is not None:
252             return sibling.get_last_text()
253
254         if self.parent.html_file:
255             return self.parent
256
257         return self.parent.get_prev_text()
258
259     def get_next_text(self, inside=True):
260         if inside:
261             child = self.children.order_by('parent_number').first()
262             if child is not None:
263                 return child.get_first_text()
264
265         if not self.parent:
266             return None
267         sibling = self.parent.children.filter(parent_number__gt=self.parent_number).order_by('parent_number').first()
268         if sibling is not None:
269             return sibling.get_first_text()
270         return self.parent.get_next_text(inside=False)
271
272     def get_child_audiobook(self):
273         BookMedia = apps.get_model('catalogue', 'BookMedia')
274         if not BookMedia.objects.filter(book__ancestor=self).exists():
275             return None
276         for child in self.children.order_by('parent_number').all():
277             if child.has_mp3_file():
278                 return child
279             child_sub = child.get_child_audiobook()
280             if child_sub is not None:
281                 return child_sub
282
283     def get_siblings(self):
284         if not self.parent:
285             return []
286         return self.parent.children.all().order_by('parent_number')
287
288     def get_children(self):
289         return self.children.all().order_by('parent_number')
290
291     @property
292     def name(self):
293         return self.title
294
295     def language_code(self):
296         return constants.LANGUAGES_3TO2.get(self.language, self.language)
297
298     def language_name(self):
299         return dict(settings.LANGUAGES).get(self.language_code(), "")
300
301     def is_foreign(self):
302         return self.language_code() != settings.LANGUAGE_CODE
303
304     def set_audio_length(self):
305         length = self.get_audio_length()
306         if length > 0:
307             self.audio_length = self.format_audio_length(length)
308             self.save()
309
310     @staticmethod
311     def format_audio_length(seconds):
312         """
313         >>> Book.format_audio_length(1)
314         '0:01'
315         >>> Book.format_audio_length(3661)
316         '1:01:01'
317         """
318         if seconds < 60*60:
319             minutes = seconds // 60
320             seconds = seconds % 60
321             return '%d:%02d' % (minutes, seconds)
322         else:
323             hours = seconds // 3600
324             minutes = seconds % 3600 // 60
325             seconds = seconds % 60
326             return '%d:%02d:%02d' % (hours, minutes, seconds)
327
328     def get_audio_length(self):
329         total = 0
330         for media in self.get_mp3() or ():
331             total += app_settings.GET_MP3_LENGTH(media.file.path)
332         return int(total)
333
334     def has_media(self, type_):
335         if type_ in Book.formats:
336             return bool(getattr(self, "%s_file" % type_))
337         else:
338             return self.media.filter(type=type_).exists()
339
340     def has_audio(self):
341         return self.has_media('mp3')
342
343     def get_media(self, type_):
344         if self.has_media(type_):
345             if type_ in Book.formats:
346                 return getattr(self, "%s_file" % type_)
347             else:
348                 return self.media.filter(type=type_)
349         else:
350             return None
351
352     def get_mp3(self):
353         return self.get_media("mp3")
354
355     def get_odt(self):
356         return self.get_media("odt")
357
358     def get_ogg(self):
359         return self.get_media("ogg")
360
361     def get_daisy(self):
362         return self.get_media("daisy")
363
364     def get_audio_epub(self):
365         return self.get_media("audio.epub")
366
367     def media_url(self, format_):
368         media = self.get_media(format_)
369         if media:
370             if self.preview:
371                 return reverse('embargo_link', kwargs={'key': self.preview_key, 'slug': self.slug, 'format_': format_})
372             else:
373                 return media.url
374         else:
375             return None
376
377     def html_url(self):
378         return self.media_url('html')
379
380     def pdf_url(self):
381         return self.media_url('pdf')
382
383     def epub_url(self):
384         return self.media_url('epub')
385
386     def mobi_url(self):
387         return self.media_url('mobi')
388
389     def txt_url(self):
390         return self.media_url('txt')
391
392     def fb2_url(self):
393         return self.media_url('fb2')
394
395     def xml_url(self):
396         return self.media_url('xml')
397
398     def has_description(self):
399         return len(self.description) > 0
400     has_description.short_description = 'opis'
401     has_description.boolean = True
402
403     def has_mp3_file(self):
404         return self.has_media("mp3")
405     has_mp3_file.short_description = 'MP3'
406     has_mp3_file.boolean = True
407
408     def has_ogg_file(self):
409         return self.has_media("ogg")
410     has_ogg_file.short_description = 'OGG'
411     has_ogg_file.boolean = True
412
413     def has_daisy_file(self):
414         return self.has_media("daisy")
415     has_daisy_file.short_description = 'DAISY'
416     has_daisy_file.boolean = True
417
418     def has_sync_file(self):
419         return self.has_media("sync")
420
421     def get_sync(self):
422         with self.get_media('sync').first().file.open('r') as f:
423             sync = f.read().split('\n')
424         offset = float(sync[0])
425         items = []
426         for line in sync[1:]:
427             if not line:
428                 continue
429             start, end, elid = line.split()
430             items.append([elid, float(start) + offset])
431         return json.dumps(items)
432     
433     def has_audio_epub_file(self):
434         return self.has_media("audio.epub")
435
436     @property
437     def media_daisy(self):
438         return self.get_media('daisy')
439
440     @property
441     def media_audio_epub(self):
442         return self.get_media('audio.epub')
443
444     def get_audiobooks(self):
445         ogg_files = {}
446         for m in self.media.filter(type='ogg').order_by().iterator():
447             ogg_files[m.name] = m
448
449         audiobooks = []
450         projects = set()
451         total_duration = 0
452         for mp3 in self.media.filter(type='mp3').iterator():
453             # ogg files are always from the same project
454             meta = mp3.get_extra_info_json()
455             project = meta.get('project')
456             if not project:
457                 # temporary fallback
458                 project = 'CzytamySłuchając'
459
460             projects.add((project, meta.get('funded_by', '')))
461             total_duration += mp3.duration or 0
462
463             media = {'mp3': mp3}
464
465             ogg = ogg_files.get(mp3.name)
466             if ogg:
467                 media['ogg'] = ogg
468             audiobooks.append(media)
469
470         projects = sorted(projects)
471         total_duration = '%d:%02d' % (
472             total_duration // 60,
473             total_duration % 60
474         )
475         return audiobooks, projects, total_duration
476
477     def wldocument(self, parse_dublincore=True, inherit=True):
478         from catalogue.import_utils import ORMDocProvider
479         from librarian.parser import WLDocument
480
481         if inherit and self.parent:
482             meta_fallbacks = self.parent.cover_info()
483         else:
484             meta_fallbacks = None
485
486         return WLDocument.from_file(
487             self.xml_file.path,
488             provider=ORMDocProvider(self),
489             parse_dublincore=parse_dublincore,
490             meta_fallbacks=meta_fallbacks)
491
492     def wldocument2(self):
493         from catalogue.import_utils import ORMDocProvider
494         from librarian.document import WLDocument
495         doc = WLDocument(
496             self.xml_file.path,
497             provider=ORMDocProvider(self)
498         )
499         doc.meta.update(self.cover_info())
500         return doc
501
502
503     @staticmethod
504     def zip_format(format_):
505         def pretty_file_name(book):
506             return "%s/%s.%s" % (
507                 book.get_extra_info_json()['author'],
508                 book.slug,
509                 format_)
510
511         field_name = "%s_file" % format_
512         field = getattr(Book, field_name)
513         books = Book.objects.filter(parent=None).exclude(**{field_name: ""}).exclude(preview=True).exclude(findable=False)
514         paths = [(pretty_file_name(b), getattr(b, field_name).path) for b in books.iterator()]
515         return create_zip(paths, field.ZIP)
516
517     def zip_audiobooks(self, format_):
518         bm = BookMedia.objects.filter(book=self, type=format_)
519         paths = map(lambda bm: (bm.get_nice_filename(), bm.file.path), bm)
520         licenses = set()
521         for m in bm:
522             license = constants.LICENSES.get(
523                 m.get_extra_info_json().get('license'), {}
524             ).get('locative')
525             if license:
526                 licenses.add(license)
527         readme = render_to_string('catalogue/audiobook_zip_readme.txt', {
528             'licenses': licenses,
529             'meta': self.wldocument2().meta,
530         })
531         return create_zip(paths, "%s_%s" % (self.slug, format_), {'informacje.txt': readme})
532
533     def search_index(self, index=None):
534         if not self.findable:
535             return
536         from search.index import Index
537         Index.index_book(self)
538
539     # will make problems in conjunction with paid previews
540     def download_pictures(self, remote_gallery_url):
541         # This is only needed for legacy relative image paths.
542         gallery_path = self.gallery_path()
543         # delete previous files, so we don't include old files in ebooks
544         if os.path.isdir(gallery_path):
545             for filename in os.listdir(gallery_path):
546                 file_path = os.path.join(gallery_path, filename)
547                 os.unlink(file_path)
548         ilustr_elements = list(self.wldocument().edoc.findall('//ilustr'))
549         if ilustr_elements:
550             makedirs(gallery_path)
551             for ilustr in ilustr_elements:
552                 ilustr_src = ilustr.get('src')
553                 if '/' in ilustr_src:
554                     continue
555                 ilustr_path = os.path.join(gallery_path, ilustr_src)
556                 urlretrieve('%s/%s' % (remote_gallery_url, ilustr_src), ilustr_path)
557
558     def load_abstract(self):
559         abstract = self.wldocument(parse_dublincore=False).edoc.getroot().find('.//abstrakt')
560         if abstract is not None:
561             self.abstract = transform_abstrakt(abstract)
562         else:
563             self.abstract = ''
564
565     def load_toc(self):
566         self.toc = ''
567         if self.html_file:
568             parser = html.HTMLParser(encoding='utf-8')
569             tree = html.parse(self.html_file.path, parser=parser)
570             toc = tree.find('//div[@id="toc"]/ol')
571             if toc is None or not len(toc):
572                 return
573             html_link = reverse('book_text', args=[self.slug])
574             for a in toc.findall('.//a'):
575                 a.attrib['href'] = html_link + a.attrib['href']
576             self.toc = html.tostring(toc, encoding='unicode')
577             # div#toc
578
579     @classmethod
580     def from_xml_file(cls, xml_file, **kwargs):
581         from django.core.files import File
582         from librarian import dcparser
583
584         # use librarian to parse meta-data
585         book_info = dcparser.parse(xml_file)
586
587         if not isinstance(xml_file, File):
588             xml_file = File(open(xml_file))
589
590         try:
591             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
592         finally:
593             xml_file.close()
594
595     @classmethod
596     def from_text_and_meta(cls, raw_file, book_info, overwrite=False, dont_build=None, search_index=True,
597                            remote_gallery_url=None, days=0, findable=True):
598         from catalogue import tasks
599
600         if dont_build is None:
601             dont_build = set()
602         dont_build = set.union(set(dont_build), set(app_settings.DONT_BUILD))
603
604         # check for parts before we do anything
605         children = []
606         if hasattr(book_info, 'parts'):
607             for part_url in book_info.parts:
608                 try:
609                     children.append(Book.objects.get(slug=part_url.slug))
610                 except Book.DoesNotExist:
611                     raise Book.DoesNotExist('Książka "%s" nie istnieje.' % part_url.slug)
612
613         # Read book metadata
614         book_slug = book_info.url.slug
615         if re.search(r'[^a-z0-9-]', book_slug):
616             raise ValueError('Invalid characters in slug')
617         book, created = Book.objects.get_or_create(slug=book_slug)
618
619         if created:
620             book_shelves = []
621             old_cover = None
622             book.preview = bool(days)
623             if book.preview:
624                 book.preview_until = date.today() + timedelta(days)
625         else:
626             if not overwrite:
627                 raise Book.AlreadyExists('Książka %s już istnieje' % book_slug)
628             # Save shelves for this book
629             book_shelves = list(book.tags.filter(category='set'))
630             old_cover = book.cover_info()
631
632         # Save XML file
633         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
634         if book.preview:
635             book.xml_file.set_readable(False)
636
637         book.findable = findable
638         book.language = book_info.language
639         book.title = book_info.title
640         if book_info.variant_of:
641             book.common_slug = book_info.variant_of.slug
642         else:
643             book.common_slug = book.slug
644         book.extra_info = json.dumps(book_info.to_dict())
645         book.load_abstract()
646         book.load_toc()
647         book.save()
648
649         meta_tags = Tag.tags_from_info(book_info)
650
651         for tag in meta_tags:
652             if not tag.for_books:
653                 tag.for_books = True
654                 tag.save()
655
656         book.tags = set(meta_tags + book_shelves)
657         book.save()  # update sort_key_author
658
659         cover_changed = old_cover != book.cover_info()
660         obsolete_children = set(b for b in book.children.all()
661                                 if b not in children)
662         notify_cover_changed = []
663         for n, child_book in enumerate(children):
664             new_child = child_book.parent != book
665             child_book.parent = book
666             child_book.parent_number = n
667             child_book.save()
668             if new_child or cover_changed:
669                 notify_cover_changed.append(child_book)
670         # Disown unfaithful children and let them cope on their own.
671         for child in obsolete_children:
672             child.parent = None
673             child.parent_number = 0
674             child.save()
675             if old_cover:
676                 notify_cover_changed.append(child)
677
678         cls.repopulate_ancestors()
679         tasks.update_counters.delay()
680
681         if remote_gallery_url:
682             book.download_pictures(remote_gallery_url)
683
684         # No saves beyond this point.
685
686         # Build cover.
687         if 'cover' not in dont_build:
688             book.cover.build_delay()
689             book.cover_clean.build_delay()
690             book.cover_thumb.build_delay()
691             book.cover_api_thumb.build_delay()
692             book.simple_cover.build_delay()
693             book.cover_ebookpoint.build_delay()
694
695         # Build HTML and ebooks.
696         book.html_file.build_delay()
697         if not children:
698             for format_ in constants.EBOOK_FORMATS_WITHOUT_CHILDREN:
699                 if format_ not in dont_build:
700                     getattr(book, '%s_file' % format_).build_delay()
701         for format_ in constants.EBOOK_FORMATS_WITH_CHILDREN:
702             if format_ not in dont_build:
703                 getattr(book, '%s_file' % format_).build_delay()
704
705         if not settings.NO_SEARCH_INDEX and search_index and findable:
706             tasks.index_book.delay(book.id)
707
708         for child in notify_cover_changed:
709             child.parent_cover_changed()
710
711         book.update_popularity()
712         tasks.update_references.delay(book.id)
713
714         cls.published.send(sender=cls, instance=book)
715         return book
716
717     def update_references(self):
718         Entity = apps.get_model('references', 'Entity')
719         doc = self.wldocument2()
720         doc._compat_assign_section_ids()
721         doc._compat_assign_ordered_ids()
722         refs = {}
723         for ref_elem in doc.references():
724             uri = ref_elem.attrib.get('href', '')
725             if not uri:
726                 continue
727             if uri in refs:
728                 ref = refs[uri]
729             else:
730                 entity, entity_created = Entity.objects.get_or_create(uri=uri)
731                 if entity_created:
732                     entity.populate()
733                     entity.save()
734                 ref, ref_created = entity.reference_set.get_or_create(book=self)
735                 refs[uri] = ref
736                 if not ref_created:
737                     ref.occurence_set.all().delete()
738             sec = ref_elem.get_link()
739             m = re.match(r'sec(\d+)', sec)
740             assert m is not None
741             sec = int(m.group(1))
742             snippet = ref_elem.get_snippet()
743             b = builders['html-snippet']()
744             for s in snippet:
745                 s.html_build(b)
746             html = b.output().get_bytes().decode('utf-8')
747
748             ref.occurence_set.create(
749                 section=sec,
750                 html=html
751             )
752         self.reference_set.exclude(entity__uri__in=refs).delete()
753
754     @property
755     def references(self):
756         return self.reference_set.all().select_related('entity')
757
758     @classmethod
759     @transaction.atomic
760     def repopulate_ancestors(cls):
761         """Fixes the ancestry cache."""
762         # TODO: table names
763         cursor = connection.cursor()
764         if connection.vendor == 'postgres':
765             cursor.execute("TRUNCATE catalogue_book_ancestor")
766             cursor.execute("""
767                 WITH RECURSIVE ancestry AS (
768                     SELECT book.id, book.parent_id
769                     FROM catalogue_book AS book
770                     WHERE book.parent_id IS NOT NULL
771                     UNION
772                     SELECT ancestor.id, book.parent_id
773                     FROM ancestry AS ancestor, catalogue_book AS book
774                     WHERE ancestor.parent_id = book.id
775                         AND book.parent_id IS NOT NULL
776                     )
777                 INSERT INTO catalogue_book_ancestor
778                     (from_book_id, to_book_id)
779                     SELECT id, parent_id
780                     FROM ancestry
781                     ORDER BY id;
782                 """)
783         else:
784             cursor.execute("DELETE FROM catalogue_book_ancestor")
785             for b in cls.objects.exclude(parent=None):
786                 parent = b.parent
787                 while parent is not None:
788                     b.ancestor.add(parent)
789                     parent = parent.parent
790
791     @property
792     def ancestors(self):
793         if self.parent:
794             for anc in self.parent.ancestors:
795                 yield anc
796             yield self.parent
797         else:
798             return []
799
800     def clear_cache(self):
801         clear_cached_renders(self.mini_box)
802         clear_cached_renders(self.mini_box_nolink)
803
804     def cover_info(self, inherit=True):
805         """Returns a dictionary to serve as fallback for BookInfo.
806
807         For now, the only thing inherited is the cover image.
808         """
809         need = False
810         info = {}
811         for field in ('cover_url', 'cover_by', 'cover_source'):
812             val = self.get_extra_info_json().get(field)
813             if val:
814                 info[field] = val
815             else:
816                 need = True
817         if inherit and need and self.parent is not None:
818             parent_info = self.parent.cover_info()
819             parent_info.update(info)
820             info = parent_info
821         return info
822
823     def related_themes(self):
824         return Tag.objects.usage_for_queryset(
825             Fragment.objects.filter(models.Q(book=self) | models.Q(book__ancestor=self)),
826             counts=True).filter(category='theme').order_by('-count')
827
828     def parent_cover_changed(self):
829         """Called when parent book's cover image is changed."""
830         if not self.cover_info(inherit=False):
831             if 'cover' not in app_settings.DONT_BUILD:
832                 self.cover.build_delay()
833                 self.cover_clean.build_delay()
834                 self.cover_thumb.build_delay()
835                 self.cover_api_thumb.build_delay()
836                 self.simple_cover.build_delay()
837                 self.cover_ebookpoint.build_delay()
838             for format_ in constants.EBOOK_FORMATS_WITH_COVERS:
839                 if format_ not in app_settings.DONT_BUILD:
840                     getattr(self, '%s_file' % format_).build_delay()
841             for child in self.children.all():
842                 child.parent_cover_changed()
843
844     def other_versions(self):
845         """Find other versions (i.e. in other languages) of the book."""
846         return type(self).objects.filter(common_slug=self.common_slug, findable=True).exclude(pk=self.pk)
847
848     def parents(self):
849         books = []
850         parent = self.parent
851         while parent is not None:
852             books.insert(0, parent)
853             parent = parent.parent
854         return books
855
856     def pretty_title(self, html_links=False):
857         names = [(tag.name, tag.get_absolute_url()) for tag in self.authors().only('name', 'category', 'slug')]
858         books = self.parents() + [self]
859         names.extend([(b.title, b.get_absolute_url()) for b in books])
860
861         if html_links:
862             names = ['<a href="%s">%s</a>' % (tag[1], tag[0]) for tag in names]
863         else:
864             names = [tag[0] for tag in names]
865         return ', '.join(names)
866
867     def publisher(self):
868         publisher = self.get_extra_info_json()['publisher']
869         if isinstance(publisher, str):
870             return publisher
871         elif isinstance(publisher, list):
872             return ', '.join(publisher)
873
874     @classmethod
875     def tagged_top_level(cls, tags):
876         """ Returns top-level books tagged with `tags`.
877
878         It only returns those books which don't have ancestors which are
879         also tagged with those tags.
880
881         """
882         objects = cls.tagged.with_all(tags)
883         return objects.filter(findable=True).exclude(ancestor__in=objects)
884
885     @classmethod
886     def book_list(cls, book_filter=None):
887         """Generates a hierarchical listing of all books.
888
889         Books are optionally filtered with a test function.
890
891         """
892
893         books_by_parent = {}
894         books = cls.objects.filter(findable=True).order_by('parent_number', 'sort_key').only('title', 'parent', 'slug', 'extra_info')
895         if book_filter:
896             books = books.filter(book_filter).distinct()
897
898             book_ids = set(b['pk'] for b in books.values("pk").iterator())
899             for book in books.iterator():
900                 parent = book.parent_id
901                 if parent not in book_ids:
902                     parent = None
903                 books_by_parent.setdefault(parent, []).append(book)
904         else:
905             for book in books.iterator():
906                 books_by_parent.setdefault(book.parent_id, []).append(book)
907
908         orphans = []
909         books_by_author = OrderedDict()
910         for tag in Tag.objects.filter(category='author').iterator():
911             books_by_author[tag] = []
912
913         for book in books_by_parent.get(None, ()):
914             authors = list(book.authors().only('pk'))
915             if authors:
916                 for author in authors:
917                     books_by_author[author].append(book)
918             else:
919                 orphans.append(book)
920
921         return books_by_author, orphans, books_by_parent
922
923     _audiences_pl = {
924         "SP": (1, "szkoła podstawowa"),
925         "SP1": (1, "szkoła podstawowa"),
926         "SP2": (1, "szkoła podstawowa"),
927         "SP3": (1, "szkoła podstawowa"),
928         "P": (1, "szkoła podstawowa"),
929         "G": (2, "gimnazjum"),
930         "L": (3, "liceum"),
931         "LP": (3, "liceum"),
932     }
933
934     def audiences_pl(self):
935         audiences = self.get_extra_info_json().get('audiences', [])
936         audiences = sorted(set([self._audiences_pl.get(a, (99, a)) for a in audiences]))
937         return [a[1] for a in audiences]
938
939     def stage_note(self):
940         stage = self.get_extra_info_json().get('stage')
941         if stage and stage < '0.4':
942             return (_('Ten utwór wymaga uwspółcześnienia'),
943                     reverse('infopage', args=['wymagajace-uwspolczesnienia']))
944         else:
945             return None, None
946
947     def choose_fragments(self, number):
948         fragments = self.fragments.order_by()
949         fragments_count = fragments.count()
950         if not fragments_count and self.children.exists():
951             fragments = Fragment.objects.filter(book__ancestor=self).order_by()
952             fragments_count = fragments.count()
953         if fragments_count:
954             if fragments_count > number:
955                 offset = randint(0, fragments_count - number)
956             else:
957                 offset = 0
958             return fragments[offset : offset + number]
959         elif self.parent:
960             return self.parent.choose_fragments(number)
961         else:
962             return []
963
964     def choose_fragment(self):
965         fragments = self.choose_fragments(1)
966         if fragments:
967             return fragments[0]
968         else:
969             return None
970
971     def fragment_data(self):
972         fragment = self.choose_fragment()
973         if fragment:
974             return {
975                 'title': fragment.book.pretty_title(),
976                 'html': re.sub('</?blockquote[^>]*>', '', fragment.get_short_text()),
977             }
978         else:
979             return None
980
981     def update_popularity(self):
982         count = self.tags.filter(category='set').values('user').order_by('user').distinct().count()
983         try:
984             pop = self.popularity
985             pop.count = count
986             pop.save()
987         except BookPopularity.DoesNotExist:
988             BookPopularity.objects.create(book=self, count=count)
989
990     def ridero_link(self):
991         return 'https://ridero.eu/%s/books/wl_%s/' % (get_language(), self.slug.replace('-', '_'))
992
993     def like(self, user):
994         from social.utils import likes, get_set, set_sets
995         if not likes(user, self):
996             tag = get_set(user, '')
997             set_sets(user, self, [tag])
998
999     def unlike(self, user):
1000         from social.utils import likes, set_sets
1001         if likes(user, self):
1002             set_sets(user, self, [])
1003
1004     def full_sort_key(self):
1005         return self.SORT_KEY_SEP.join((self.sort_key_author, self.sort_key, str(self.id)))
1006
1007     def cover_color(self):
1008         return WLCover.epoch_colors.get(self.get_extra_info_json().get('epoch'), '#000000')
1009
1010     @cached_render('catalogue/book_mini_box.html')
1011     def mini_box(self):
1012         return {
1013             'book': self
1014         }
1015
1016     @cached_render('catalogue/book_mini_box.html')
1017     def mini_box_nolink(self):
1018         return {
1019             'book': self,
1020             'no_link': True,
1021         }
1022
1023
1024 class BookPopularity(models.Model):
1025     book = models.OneToOneField(Book, models.CASCADE, related_name='popularity')
1026     count = models.IntegerField(default=0, db_index=True)