Fix child book reimport bugs.
[wolnelektury.git] / apps / catalogue / models / book.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 import re
6 from django.conf import settings
7 from django.core.cache import get_cache
8 from django.db import models
9 from django.db.models import permalink
10 import django.dispatch
11 from django.utils.datastructures import SortedDict
12 from django.utils.translation import ugettext_lazy as _
13 import jsonfield
14 from catalogue.models import Tag, Fragment, BookMedia
15 from catalogue.utils import create_zip, split_tags, truncate_html_words, book_upload_path
16 from catalogue import tasks
17 from newtagging import managers
18
19
20 permanent_cache = get_cache('permanent')
21
22
23 class Book(models.Model):
24     """Represents a book imported from WL-XML."""
25     title         = models.CharField(_('title'), max_length=120)
26     sort_key = models.CharField(_('sort key'), max_length=120, db_index=True, editable=False)
27     slug = models.SlugField(_('slug'), max_length=120, db_index=True,
28             unique=True)
29     common_slug = models.SlugField(_('slug'), max_length=120, db_index=True)
30     language = models.CharField(_('language code'), max_length=3, db_index=True,
31                     default=settings.CATALOGUE_DEFAULT_LANGUAGE)
32     description   = models.TextField(_('description'), blank=True)
33     created_at    = models.DateTimeField(_('creation date'), auto_now_add=True, db_index=True)
34     changed_at    = models.DateTimeField(_('creation date'), auto_now=True, db_index=True)
35     parent_number = models.IntegerField(_('parent number'), default=0)
36     extra_info    = jsonfield.JSONField(_('extra information'), default='{}')
37     gazeta_link   = models.CharField(blank=True, max_length=240)
38     wiki_link     = models.CharField(blank=True, max_length=240)
39     # files generated during publication
40
41     cover = models.FileField(_('cover'), upload_to=book_upload_path('png'),
42                 null=True, blank=True)
43     ebook_formats = ['pdf', 'epub', 'mobi', 'fb2', 'txt']
44     formats = ebook_formats + ['html', 'xml']
45
46     parent        = models.ForeignKey('self', blank=True, null=True, related_name='children')
47
48     _related_info = jsonfield.JSONField(blank=True, null=True, editable=False)
49
50     objects  = models.Manager()
51     tagged   = managers.ModelTaggedItemManager(Tag)
52     tags     = managers.TagDescriptor(Tag)
53
54     html_built = django.dispatch.Signal()
55     published = django.dispatch.Signal()
56
57     class AlreadyExists(Exception):
58         pass
59
60     class Meta:
61         ordering = ('sort_key',)
62         verbose_name = _('book')
63         verbose_name_plural = _('books')
64         app_label = 'catalogue'
65
66     def __unicode__(self):
67         return self.title
68
69     def save(self, force_insert=False, force_update=False, reset_short_html=True, **kwargs):
70         from sortify import sortify
71
72         self.sort_key = sortify(self.title)
73
74         ret = super(Book, self).save(force_insert, force_update)
75
76         if reset_short_html:
77             self.reset_short_html()
78
79         return ret
80
81     @permalink
82     def get_absolute_url(self):
83         return ('catalogue.views.book_detail', [self.slug])
84
85     @property
86     def name(self):
87         return self.title
88
89     def book_tag_slug(self):
90         return ('l-' + self.slug)[:120]
91
92     def book_tag(self):
93         slug = self.book_tag_slug()
94         book_tag, created = Tag.objects.get_or_create(slug=slug, category='book')
95         if created:
96             book_tag.name = self.title[:50]
97             book_tag.sort_key = self.title.lower()
98             book_tag.save()
99         return book_tag
100
101     def has_media(self, type_):
102         if type_ in Book.formats:
103             return bool(getattr(self, "%s_file" % type_))
104         else:
105             return self.media.filter(type=type_).exists()
106
107     def get_media(self, type_):
108         if self.has_media(type_):
109             if type_ in Book.formats:
110                 return getattr(self, "%s_file" % type_)
111             else:                                             
112                 return self.media.filter(type=type_)
113         else:
114             return None
115
116     def get_mp3(self):
117         return self.get_media("mp3")
118     def get_odt(self):
119         return self.get_media("odt")
120     def get_ogg(self):
121         return self.get_media("ogg")
122     def get_daisy(self):
123         return self.get_media("daisy")                       
124
125     def reset_short_html(self):
126         if self.id is None:
127             return
128
129         type(self).objects.filter(pk=self.pk).update(_related_info=None)
130         # Fragment.short_html relies on book's tags, so reset it here too
131         for fragm in self.fragments.all().iterator():
132             fragm.reset_short_html()
133
134     def has_description(self):
135         return len(self.description) > 0
136     has_description.short_description = _('description')
137     has_description.boolean = True
138
139     # ugly ugly ugly
140     def has_mp3_file(self):
141         return bool(self.has_media("mp3"))
142     has_mp3_file.short_description = 'MP3'
143     has_mp3_file.boolean = True
144
145     def has_ogg_file(self):
146         return bool(self.has_media("ogg"))
147     has_ogg_file.short_description = 'OGG'
148     has_ogg_file.boolean = True
149
150     def has_daisy_file(self):
151         return bool(self.has_media("daisy"))
152     has_daisy_file.short_description = 'DAISY'
153     has_daisy_file.boolean = True
154
155     def wldocument(self, parse_dublincore=True):
156         from catalogue.import_utils import ORMDocProvider
157         from librarian.parser import WLDocument
158
159         return WLDocument.from_file(self.xml_file.path,
160                 provider=ORMDocProvider(self),
161                 parse_dublincore=parse_dublincore)
162
163     def build_cover(self, book_info=None):
164         """(Re)builds the cover image."""
165         from StringIO import StringIO
166         from django.core.files.base import ContentFile
167         from librarian.cover import WLCover
168
169         if book_info is None:
170             book_info = self.wldocument().book_info
171
172         cover = WLCover(book_info).image()
173         imgstr = StringIO()
174         cover.save(imgstr, 'png')
175         self.cover.save(None, ContentFile(imgstr.getvalue()))
176
177     def build_html(self):
178         from django.core.files.base import ContentFile
179         from slughifi import slughifi
180         from sortify import sortify
181         from librarian import html
182
183         meta_tags = list(self.tags.filter(
184             category__in=('author', 'epoch', 'genre', 'kind')))
185         book_tag = self.book_tag()
186
187         html_output = self.wldocument(parse_dublincore=False).as_html()
188         if html_output:
189             self.html_file.save('%s.html' % self.slug,
190                     ContentFile(html_output.get_string()))
191
192             # get ancestor l-tags for adding to new fragments
193             ancestor_tags = []
194             p = self.parent
195             while p:
196                 ancestor_tags.append(p.book_tag())
197                 p = p.parent
198
199             # Delete old fragments and create them from scratch
200             self.fragments.all().delete()
201             # Extract fragments
202             closed_fragments, open_fragments = html.extract_fragments(self.html_file.path)
203             for fragment in closed_fragments.values():
204                 try:
205                     theme_names = [s.strip() for s in fragment.themes.split(',')]
206                 except AttributeError:
207                     continue
208                 themes = []
209                 for theme_name in theme_names:
210                     if not theme_name:
211                         continue
212                     tag, created = Tag.objects.get_or_create(slug=slughifi(theme_name), category='theme')
213                     if created:
214                         tag.name = theme_name
215                         tag.sort_key = sortify(theme_name.lower())
216                         tag.save()
217                     themes.append(tag)
218                 if not themes:
219                     continue
220
221                 text = fragment.to_string()
222                 short_text = truncate_html_words(text, 15)
223                 if text == short_text:
224                     short_text = ''
225                 new_fragment = Fragment.objects.create(anchor=fragment.id, book=self,
226                     text=text, short_text=short_text)
227
228                 new_fragment.save()
229                 new_fragment.tags = set(meta_tags + themes + [book_tag] + ancestor_tags)
230             self.save()
231             self.html_built.send(sender=self)
232             return True
233         return False
234
235     # Thin wrappers for builder tasks
236     def build_pdf(self, *args, **kwargs):
237         """(Re)builds PDF."""
238         return tasks.build_pdf.delay(self.pk, *args, **kwargs)
239     def build_epub(self, *args, **kwargs):
240         """(Re)builds EPUB."""
241         return tasks.build_epub.delay(self.pk, *args, **kwargs)
242     def build_mobi(self, *args, **kwargs):
243         """(Re)builds MOBI."""
244         return tasks.build_mobi.delay(self.pk, *args, **kwargs)
245     def build_fb2(self, *args, **kwargs):
246         """(Re)build FB2"""
247         return tasks.build_fb2.delay(self.pk, *args, **kwargs)
248     def build_txt(self, *args, **kwargs):
249         """(Re)builds TXT."""
250         return tasks.build_txt.delay(self.pk, *args, **kwargs)
251
252     @staticmethod
253     def zip_format(format_):
254         def pretty_file_name(book):
255             return "%s/%s.%s" % (
256                 b.extra_info['author'],
257                 b.slug,
258                 format_)
259
260         field_name = "%s_file" % format_
261         books = Book.objects.filter(parent=None).exclude(**{field_name: ""})
262         paths = [(pretty_file_name(b), getattr(b, field_name).path)
263                     for b in books.iterator()]
264         return create_zip(paths,
265                     getattr(settings, "ALL_%s_ZIP" % format_.upper()))
266
267     def zip_audiobooks(self, format_):
268         bm = BookMedia.objects.filter(book=self, type=format_)
269         paths = map(lambda bm: (None, bm.file.path), bm)
270         return create_zip(paths, "%s_%s" % (self.slug, format_))
271
272     def search_index(self, book_info=None, reuse_index=False, index_tags=True):
273         import search
274         if reuse_index:
275             idx = search.ReusableIndex()
276         else:
277             idx = search.Index()
278             
279         idx.open()
280         try:
281             idx.index_book(self, book_info)
282             if index_tags:
283                 idx.index_tags()
284         finally:
285             idx.close()
286
287     @classmethod
288     def from_xml_file(cls, xml_file, **kwargs):
289         from django.core.files import File
290         from librarian import dcparser
291
292         # use librarian to parse meta-data
293         book_info = dcparser.parse(xml_file)
294
295         if not isinstance(xml_file, File):
296             xml_file = File(open(xml_file))
297
298         try:
299             return cls.from_text_and_meta(xml_file, book_info, **kwargs)
300         finally:
301             xml_file.close()
302
303     @classmethod
304     def from_text_and_meta(cls, raw_file, book_info, overwrite=False,
305             build_epub=True, build_txt=True, build_pdf=True, build_mobi=True, build_fb2=True,
306             search_index=True, search_index_tags=True, search_index_reuse=False):
307
308         # check for parts before we do anything
309         children = []
310         if hasattr(book_info, 'parts'):
311             for part_url in book_info.parts:
312                 try:
313                     children.append(Book.objects.get(slug=part_url.slug))
314                 except Book.DoesNotExist:
315                     raise Book.DoesNotExist(_('Book "%s" does not exist.') %
316                             part_url.slug)
317
318
319         # Read book metadata
320         book_slug = book_info.url.slug
321         if re.search(r'[^a-z0-9-]', book_slug):
322             raise ValueError('Invalid characters in slug')
323         book, created = Book.objects.get_or_create(slug=book_slug)
324
325         if created:
326             book_shelves = []
327         else:
328             if not overwrite:
329                 raise Book.AlreadyExists(_('Book %s already exists') % (
330                         book_slug))
331             # Save shelves for this book
332             book_shelves = list(book.tags.filter(category='set'))
333
334         book.language = book_info.language
335         book.title = book_info.title
336         if book_info.variant_of:
337             book.common_slug = book_info.variant_of.slug
338         else:
339             book.common_slug = book.slug
340         book.extra_info = book_info.to_dict()
341         book.save()
342
343         meta_tags = Tag.tags_from_info(book_info)
344
345         book.tags = set(meta_tags + book_shelves)
346
347         obsolete_children = set(b for b in book.children.all() if b not in children)
348         for n, child_book in enumerate(children):
349             child_book.parent = book
350             child_book.parent_number = n
351             child_book.save()
352         # Disown unfaithful children and let them cope on their own.
353         for child in obsolete_children:
354             child.parent = None
355             child.parent_number = 0
356             child.save()
357             tasks.fix_tree_tags.delay(child)
358
359         # Save XML and HTML files
360         book.xml_file.save('%s.xml' % book.slug, raw_file, save=False)
361         book.build_cover(book_info)
362
363         # delete old fragments when overwriting
364         book.fragments.all().delete()
365
366         if book.build_html():
367             # No direct saves behind this point.
368             if not settings.NO_BUILD_TXT and build_txt:
369                 book.build_txt()
370
371         if not settings.NO_BUILD_EPUB and build_epub:
372             book.build_epub()
373
374         if not settings.NO_BUILD_PDF and build_pdf:
375             book.build_pdf()
376
377         if not settings.NO_BUILD_MOBI and build_mobi:
378             book.build_mobi()
379
380         if not settings.NO_BUILD_FB2 and build_fb2:
381             book.build_fb2()
382
383         if not settings.NO_SEARCH_INDEX and search_index:
384             book.search_index(index_tags=search_index_tags, reuse_index=search_index_reuse)
385             #index_book.delay(book.id, book_info)
386
387         tasks.fix_tree_tags.delay(book)
388         cls.published.send(sender=book)
389         return book
390
391     def fix_tree_tags(self):
392         """Fixes the l-tags on the book's subtree.
393
394         Makes sure that:
395         * the book has its parents book-tags,
396         * its fragments have the book's and its parents book-tags,
397         * runs those for every child book too,
398         * touches all relevant tags,
399         * resets tag and theme counter on the book and its ancestry.
400         """
401         def fix_subtree(book, parent_tags):
402             affected_tags = set(book.tags)
403             book.tags = list(book.tags.exclude(category='book')) + parent_tags
404             sub_parent_tags = parent_tags + [book.book_tag()]
405             for frag in book.fragments.all():
406                 affected_tags.update(frag.tags)
407                 frag.tags = list(frag.tags.exclude(category='book')) + sub_parent_tags
408             for child in book.children.all():
409                 affected_tags.update(fix_subtree(child, sub_parent_tags))
410             return affected_tags
411
412         parent_tags = []
413         parent = self.parent
414         while parent is not None:
415             parent_tags.append(parent.book_tag())
416             parent = parent.parent
417
418         affected_tags = fix_subtree(self, parent_tags)
419         for tag in affected_tags:
420             tasks.touch_tag(tag)
421
422         book = self
423         while book is not None:
424             book.reset_tag_counter()
425             book.reset_theme_counter()
426             book = book.parent
427
428     def related_info(self):
429         """Keeps info about related objects (tags, media) in cache field."""
430         if self._related_info is not None:
431             return self._related_info
432         else:
433             rel = {'tags': {}, 'media': {}}
434
435             tags = self.tags.filter(category__in=(
436                     'author', 'kind', 'genre', 'epoch'))
437             tags = split_tags(tags)
438             for category in tags:
439                 rel['tags'][category] = [
440                         (t.name, t.slug) for t in tags[category]]
441
442             for media_format in BookMedia.formats:
443                 rel['media'][media_format] = self.has_media(media_format)
444
445             book = self
446             parents = []
447             while book.parent:
448                 parents.append((book.parent.title, book.parent.slug))
449                 book = book.parent
450             parents = parents[::-1]
451             if parents:
452                 rel['parents'] = parents
453
454             if self.pk:
455                 type(self).objects.filter(pk=self.pk).update(_related_info=rel)
456             return rel
457
458     def related_themes(self):
459         theme_counter = self.theme_counter
460         book_themes = list(Tag.objects.filter(pk__in=theme_counter.keys()))
461         for tag in book_themes:
462             tag.count = theme_counter[tag.pk]
463         return book_themes
464
465     def reset_tag_counter(self):
466         if self.id is None:
467             return
468
469         cache_key = "Book.tag_counter/%d" % self.id
470         permanent_cache.delete(cache_key)
471         if self.parent:
472             self.parent.reset_tag_counter()
473
474     @property
475     def tag_counter(self):
476         if self.id:
477             cache_key = "Book.tag_counter/%d" % self.id
478             tags = permanent_cache.get(cache_key)
479         else:
480             tags = None
481
482         if tags is None:
483             tags = {}
484             for child in self.children.all().order_by().iterator():
485                 for tag_pk, value in child.tag_counter.iteritems():
486                     tags[tag_pk] = tags.get(tag_pk, 0) + value
487             for tag in self.tags.exclude(category__in=('book', 'theme', 'set')).order_by().iterator():
488                 tags[tag.pk] = 1
489
490             if self.id:
491                 permanent_cache.set(cache_key, tags)
492         return tags
493
494     def reset_theme_counter(self):
495         if self.id is None:
496             return
497
498         cache_key = "Book.theme_counter/%d" % self.id
499         permanent_cache.delete(cache_key)
500         if self.parent:
501             self.parent.reset_theme_counter()
502
503     @property
504     def theme_counter(self):
505         if self.id:
506             cache_key = "Book.theme_counter/%d" % self.id
507             tags = permanent_cache.get(cache_key)
508         else:
509             tags = None
510
511         if tags is None:
512             tags = {}
513             for fragment in Fragment.tagged.with_any([self.book_tag()]).order_by().iterator():
514                 for tag in fragment.tags.filter(category='theme').order_by().iterator():
515                     tags[tag.pk] = tags.get(tag.pk, 0) + 1
516
517             if self.id:
518                 permanent_cache.set(cache_key, tags)
519         return tags
520
521     def pretty_title(self, html_links=False):
522         book = self
523         names = list(book.tags.filter(category='author'))
524
525         books = []
526         while book:
527             books.append(book)
528             book = book.parent
529         names.extend(reversed(books))
530
531         if html_links:
532             names = ['<a href="%s">%s</a>' % (tag.get_absolute_url(), tag.name) for tag in names]
533         else:
534             names = [tag.name for tag in names]
535
536         return ', '.join(names)
537
538     @classmethod
539     def tagged_top_level(cls, tags):
540         """ Returns top-level books tagged with `tags`.
541
542         It only returns those books which don't have ancestors which are
543         also tagged with those tags.
544
545         """
546         # get relevant books and their tags
547         objects = cls.tagged.with_all(tags)
548         # eliminate descendants
549         l_tags = Tag.objects.filter(category='book',
550             slug__in=[book.book_tag_slug() for book in objects.iterator()])
551         descendants_keys = [book.pk for book in cls.tagged.with_any(l_tags).iterator()]
552         if descendants_keys:
553             objects = objects.exclude(pk__in=descendants_keys)
554
555         return objects
556
557     @classmethod
558     def book_list(cls, filter=None):
559         """Generates a hierarchical listing of all books.
560
561         Books are optionally filtered with a test function.
562
563         """
564
565         books_by_parent = {}
566         books = cls.objects.all().order_by('parent_number', 'sort_key').only(
567                 'title', 'parent', 'slug')
568         if filter:
569             books = books.filter(filter).distinct()
570             
571             book_ids = set(b['pk'] for b in books.values("pk").iterator())
572             for book in books.iterator():
573                 parent = book.parent_id
574                 if parent not in book_ids:
575                     parent = None
576                 books_by_parent.setdefault(parent, []).append(book)
577         else:
578             for book in books.iterator():
579                 books_by_parent.setdefault(book.parent_id, []).append(book)
580
581         orphans = []
582         books_by_author = SortedDict()
583         for tag in Tag.objects.filter(category='author').iterator():
584             books_by_author[tag] = []
585
586         for book in books_by_parent.get(None,()):
587             authors = list(book.tags.filter(category='author'))
588             if authors:
589                 for author in authors:
590                     books_by_author[author].append(book)
591             else:
592                 orphans.append(book)
593
594         return books_by_author, orphans, books_by_parent
595
596     _audiences_pl = {
597         "SP1": (1, u"szkoła podstawowa"),
598         "SP2": (1, u"szkoła podstawowa"),
599         "P": (1, u"szkoła podstawowa"),
600         "G": (2, u"gimnazjum"),
601         "L": (3, u"liceum"),
602         "LP": (3, u"liceum"),
603     }
604     def audiences_pl(self):
605         audiences = self.extra_info.get('audiences', [])
606         audiences = sorted(set([self._audiences_pl[a] for a in audiences]))
607         return [a[1] for a in audiences]
608
609     def choose_fragment(self):
610         tag = self.book_tag()
611         fragments = Fragment.tagged.with_any([tag])
612         if fragments.exists():
613             return fragments.order_by('?')[0]
614         elif self.parent:
615             return self.parent.choose_fragment()
616         else:
617             return None
618
619
620 def _has_factory(ftype):
621     has = lambda self: bool(getattr(self, "%s_file" % ftype))
622     has.short_description = ftype.upper()
623     has.__doc__ = None
624     has.boolean = True
625     has.__name__ = "has_%s_file" % ftype
626     return has
627
628     
629 # add the file fields
630 for t in Book.formats:
631     field_name = "%s_file" % t
632     models.FileField(_("%s file" % t.upper()),
633             upload_to=book_upload_path(t),
634             blank=True).contribute_to_class(Book, field_name)
635
636     setattr(Book, "has_%s_file" % t, _has_factory(t))