src/documents/models/book.py

   1 # This file is part of FNP-Redakcja, licensed under GNU Affero GPLv3 or later.
   2 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
   3 #
   4 from django.apps import apps
   5 from django.core.files.base import ContentFile
   6 from django.contrib.sites.models import Site
   7 from django.db import connection, models, transaction
   8 from django.template.loader import render_to_string
   9 from django.urls import reverse
  10 from django.utils.translation import gettext_lazy as _
  11 from django.conf import settings
  12 from slugify import slugify
  13 from librarian.cover import make_cover
  14 from librarian.dcparser import BookInfo
  15
  16 import apiclient
  17 from documents.helpers import cached_in_field, GalleryMerger
  18 from documents.models import BookPublishRecord, ChunkPublishRecord, Project
  19 from documents.signals import post_publish
  20 from documents.xml_tools import compile_text, split_xml
  21 from cover.models import Image
  22 from io import BytesIO
  23 import os
  24 import shutil
  25 import re
  26
  27
  28 class Book(models.Model):
  29     """ A document edited on the wiki """
  30
  31     title = models.CharField(_('title'), max_length=255, db_index=True)
  32     slug = models.SlugField(_('slug'), max_length=128, unique=True, db_index=True)
  33     public = models.BooleanField(_('public'), default=True, db_index=True)
  34     gallery = models.CharField(_('scan gallery name'), max_length=255, blank=True)
  35     project = models.ForeignKey(Project, models.SET_NULL, null=True, blank=True)
  36
  37     parent = models.ForeignKey('self', models.SET_NULL, null=True, blank=True, verbose_name=_('parent'), related_name="children", editable=False)
  38     parent_number = models.IntegerField(_('parent number'), null=True, blank=True, db_index=True, editable=False)
  39
  40     # Cache
  41     _single = models.BooleanField(editable=False, null=True, db_index=True)
  42     _new_publishable = models.BooleanField(editable=False, null=True)
  43     _published = models.BooleanField(editable=False, null=True)
  44     _on_track = models.IntegerField(null=True, blank=True, db_index=True, editable=False)
  45     dc_cover_image = models.ForeignKey(Image, blank=True, null=True,
  46         db_index=True, on_delete=models.SET_NULL, editable=False)
  47     dc = models.JSONField(null=True, editable=False)
  48     cover = models.FileField(blank=True, upload_to='documents/cover')
  49     catalogue_book = models.ForeignKey(
  50         'catalogue.Book',
  51         models.DO_NOTHING,
  52         to_field='slug',
  53         null=True, blank=True,
  54         db_constraint=False,
  55         editable=False, db_index=True,
  56         related_name='document_books',
  57         related_query_name='document_book',
  58     )
  59     legimi_id = models.CharField(max_length=255, blank=True)
  60
  61     class NoTextError(BaseException):
  62         pass
  63
  64     class Meta:
  65         app_label = 'documents'
  66         ordering = ['title', 'slug']
  67         verbose_name = _('book')
  68         verbose_name_plural = _('books')
  69
  70     @classmethod
  71     def get_visible_for(cls, user):
  72         qs = cls.objects.all()
  73         if not user.is_authenticated:
  74             qs = qs.filter(public=True)
  75         return qs
  76
  77     @staticmethod
  78     def q_dc(field, field_plural, value, prefix=''):
  79         if connection.features.supports_json_field_contains:
  80             return models.Q(**{f'{prefix}dc__{field_plural}__contains': value})
  81         else:
  82             return models.Q(**{f'{prefix}dc__{field}': value})
  83
  84
  85     # Representing
  86     # ============
  87
  88     def __iter__(self):
  89         return iter(self.chunk_set.all())
  90
  91     def __getitem__(self, chunk):
  92         return self.chunk_set.all()[chunk]
  93
  94     def __len__(self):
  95         return self.chunk_set.count()
  96
  97     def __bool__(self):
  98         """
  99             Necessary so that __len__ isn't used for bool evaluation.
 100         """
 101         return True
 102
 103     def __str__(self):
 104         return self.title
 105
 106     def get_absolute_url(self):
 107         return reverse("documents_book", args=[self.slug])
 108
 109     def correct_about(self):
 110         return "http://%s%s" % (
 111             Site.objects.get_current().domain,
 112             self.get_absolute_url()
 113         )
 114
 115     def gallery_path(self):
 116         return os.path.join(settings.MEDIA_ROOT, settings.IMAGE_DIR, self.gallery)
 117
 118     def gallery_url(self):
 119         return '%s%s%s/' % (settings.MEDIA_URL, settings.IMAGE_DIR, self.gallery)
 120
 121     # Creating & manipulating
 122     # =======================
 123
 124     def accessible(self, request):
 125         return self.public or request.user.is_authenticated
 126
 127     @classmethod
 128     @transaction.atomic
 129     def create(cls, creator, text, *args, **kwargs):
 130         b = cls.objects.create(*args, **kwargs)
 131         b.chunk_set.all().update(creator=creator)
 132         b[0].commit(text, author=creator)
 133         return b
 134
 135     def add(self, *args, **kwargs):
 136         """Add a new chunk at the end."""
 137         return self.chunk_set.reverse()[0].split(*args, **kwargs)
 138
 139     @classmethod
 140     @transaction.atomic
 141     def import_xml_text(cls, text=u'', previous_book=None,
 142                 commit_args=None, **kwargs):
 143         """Imports a book from XML, splitting it into chunks as necessary."""
 144         texts = split_xml(text)
 145         if previous_book:
 146             instance = previous_book
 147         else:
 148             instance = cls(**kwargs)
 149             instance.save()
 150
 151         # if there are more parts, set the rest to empty strings
 152         book_len = len(instance)
 153         for i in range(book_len - len(texts)):
 154             texts.append((u'pusta część %d' % (i + 1), u''))
 155
 156         i = 0
 157         for i, (title, text) in enumerate(texts):
 158             if not title:
 159                 title = u'część %d' % (i + 1)
 160
 161             slug = slugify(title)
 162
 163             if i < book_len:
 164                 chunk = instance[i]
 165                 chunk.slug = slug[:50]
 166                 chunk.title = title[:255]
 167                 chunk.save()
 168             else:
 169                 chunk = instance.add(slug, title)
 170
 171             chunk.commit(text, **commit_args)
 172
 173         return instance
 174
 175     def make_chunk_slug(self, proposed):
 176         """
 177             Finds a chunk slug not yet used in the book.
 178         """
 179         slugs = set(c.slug for c in self)
 180         i = 1
 181         new_slug = proposed[:50]
 182         while new_slug in slugs:
 183             new_slug = "%s_%d" % (proposed[:45], i)
 184             i += 1
 185         return new_slug
 186
 187     @transaction.atomic
 188     def append(self, other, slugs=None, titles=None):
 189         """Add all chunks of another book to self."""
 190         assert self != other
 191
 192         number = self[len(self) - 1].number + 1
 193         len_other = len(other)
 194         single = len_other == 1
 195
 196         if slugs is not None:
 197             assert len(slugs) == len_other
 198         if titles is not None:
 199             assert len(titles) == len_other
 200             if slugs is None:
 201                 slugs = [slugify(t) for t in titles]
 202
 203         for i, chunk in enumerate(other):
 204             # move chunk to new book
 205             chunk.book = self
 206             chunk.number = number
 207
 208             if titles is None:
 209                 # try some title guessing
 210                 if other.title.startswith(self.title):
 211                     other_title_part = other.title[len(self.title):].lstrip(' /')
 212                 else:
 213                     other_title_part = other.title
 214
 215                 if single:
 216                     # special treatment for appending one-parters:
 217                     # just use the guessed title and original book slug
 218                     chunk.title = other_title_part
 219                     if other.slug.startswith(self.slug):
 220                         chunk.slug = other.slug[len(self.slug):].lstrip('-_')
 221                     else:
 222                         chunk.slug = other.slug
 223                 else:
 224                     chunk.title = ("%s, %s" % (other_title_part, chunk.title))[:255]
 225             else:
 226                 chunk.slug = slugs[i]
 227                 chunk.title = titles[i]
 228
 229             chunk.slug = self.make_chunk_slug(chunk.slug)
 230             chunk.save()
 231             number += 1
 232         assert not other.chunk_set.exists()
 233
 234         gm = GalleryMerger(self.gallery, other.gallery)
 235         self.gallery = gm.merge()
 236
 237         # and move the gallery starts
 238         if gm.was_merged:
 239                 for chunk in self[len(self) - len_other:]:
 240                         old_start = chunk.gallery_start or 1
 241                         chunk.gallery_start = old_start + gm.dest_size - gm.num_deleted
 242                         chunk.save()
 243
 244         other.delete()
 245
 246
 247     @transaction.atomic
 248     def prepend_history(self, other):
 249         """Prepend history from all the other book's chunks to own."""
 250         assert self != other
 251
 252         for i in range(len(self), len(other)):
 253             title = u"pusta część %d" % i
 254             chunk = self.add(slugify(title), title)
 255             chunk.commit('')
 256
 257         for i in range(len(other)):
 258             self[i].prepend_history(other[0])
 259
 260         assert not other.chunk_set.exists()
 261         other.delete()
 262
 263     def split(self):
 264         """Splits all the chunks into separate books."""
 265         self.title
 266         for chunk in self:
 267             book = Book.objects.create(title=chunk.title, slug=chunk.slug,
 268                     public=self.public, gallery=self.gallery)
 269             book[0].delete()
 270             chunk.book = book
 271             chunk.number = 1
 272             chunk.save()
 273         assert not self.chunk_set.exists()
 274         self.delete()
 275
 276     # State & cache
 277     # =============
 278
 279     def last_published(self):
 280         try:
 281             return self.publish_log.all()[0].timestamp
 282         except IndexError:
 283             return None
 284
 285     def last_legimi_publish(self):
 286         return self.legimibookpublish_set.order_by('-created_at').first()
 287
 288     def assert_publishable(self):
 289         assert self.chunk_set.exists(), _('No chunks in the book.')
 290         try:
 291             changes = self.get_current_changes(publishable=True)
 292         except self.NoTextError:
 293             raise AssertionError(_('Not all chunks have publishable revisions.'))
 294
 295         from librarian import NoDublinCore, ParseError, ValidationError
 296
 297         try:
 298             bi = self.wldocument(changes=changes, strict=True).book_info
 299         except ParseError as e:
 300             raise AssertionError(_('Invalid XML') + ': ' + str(e))
 301         except NoDublinCore:
 302             raise AssertionError(_('No Dublin Core found.'))
 303         except ValidationError as e:
 304             raise AssertionError(_('Invalid Dublin Core') + ': ' + str(e))
 305
 306         valid_about = self.correct_about()
 307         assert bi.about == valid_about, _("rdf:about is not") + " " + valid_about
 308
 309     def publishable_error(self):
 310         try:
 311             return self.assert_publishable()
 312         except AssertionError as e:
 313             return e
 314         else:
 315             return None
 316
 317     def hidden(self):
 318         return self.slug.startswith('.')
 319
 320     def is_new_publishable(self):
 321         """Checks if book is ready for publishing.
 322
 323         Returns True if there is a publishable version newer than the one
 324         already published.
 325
 326         """
 327         new_publishable = False
 328         if not self.chunk_set.exists():
 329             return False
 330         for chunk in self:
 331             change = chunk.publishable()
 332             if not change:
 333                 return False
 334             if not new_publishable and not change.publish_log.exists():
 335                 new_publishable = True
 336         return new_publishable
 337     new_publishable = cached_in_field('_new_publishable')(is_new_publishable)
 338
 339     def is_published(self):
 340         return self.publish_log.exists()
 341     published = cached_in_field('_published')(is_published)
 342
 343     def get_on_track(self):
 344         if self.published:
 345             return -1
 346         stages = [ch.stage.ordering if ch.stage is not None else 0
 347                     for ch in self]
 348         if not len(stages):
 349             return 0
 350         return min(stages)
 351     on_track = cached_in_field('_on_track')(get_on_track)
 352
 353     def is_single(self):
 354         return len(self) == 1
 355     single = cached_in_field('_single')(is_single)
 356
 357     def book_info(self, publishable=True):
 358         try:
 359             book_xml = self.materialize(publishable=publishable)
 360         except self.NoTextError:
 361             pass
 362         else:
 363             from librarian.dcparser import BookInfo
 364             from librarian import NoDublinCore, ParseError, ValidationError
 365             try:
 366                 return BookInfo.from_bytes(book_xml.encode('utf-8'))
 367             except (self.NoTextError, ParseError, NoDublinCore, ValidationError):
 368                 return None
 369
 370     def refresh_dc_cache(self):
 371         update = {
 372             'catalogue_book_id': None,
 373             'dc_cover_image': None,
 374         }
 375
 376         info = self.book_info()
 377         if info is not None:
 378             update['catalogue_book_id'] = info.url.slug
 379             if info.cover_source:
 380                 try:
 381                     image = Image.objects.get(pk=int(info.cover_source.rstrip('/').rsplit('/', 1)[-1]))
 382                 except:
 383                     pass
 384                 else:
 385                     if info.cover_source == image.get_full_url():
 386                         update['dc_cover_image'] = image
 387             update['dc'] = info.to_dict()
 388         Book.objects.filter(pk=self.pk).update(**update)
 389
 390     def touch(self):
 391         update = {
 392             "_new_publishable": self.is_new_publishable(),
 393             "_published": self.is_published(),
 394             "_single": self.is_single(),
 395             "_on_track": self.get_on_track(),
 396         }
 397         Book.objects.filter(pk=self.pk).update(**update)
 398         self.refresh_dc_cache()
 399         self.build_cover()
 400
 401     def build_cover(self):
 402         width, height = 212, 300
 403         try:
 404             xml = self.materialize(publishable=True).encode('utf-8')
 405             info = BookInfo.from_bytes(xml)
 406             cover = make_cover(info, width=width, height=height)
 407             out = BytesIO()
 408             ext = cover.ext()
 409             cover.save(out)
 410             self.cover.save(f'{self.slug}.{ext}', out, save=False)
 411             type(self).objects.filter(pk=self.pk).update(cover=self.cover)
 412         except:
 413             type(self).objects.filter(pk=self.pk).update(cover='')
 414
 415     # Materializing & publishing
 416     # ==========================
 417
 418     def get_current_changes(self, publishable=True):
 419         """
 420             Returns a list containing one Change for every Chunk in the Book.
 421             Takes the most recent revision (publishable, if set).
 422             Throws an error, if a proper revision is unavailable for a Chunk.
 423         """
 424         if publishable:
 425             changes = [chunk.publishable() for chunk in self]
 426         else:
 427             changes = [chunk.head for chunk in self if chunk.head is not None]
 428         if None in changes:
 429             raise self.NoTextError('Some chunks have no available text.')
 430         return changes
 431
 432     def materialize(self, publishable=False, changes=None):
 433         """
 434             Get full text of the document compiled from chunks.
 435             Takes the current versions of all texts
 436             or versions most recently tagged for publishing,
 437             or a specified iterable changes.
 438         """
 439         if changes is None:
 440             changes = self.get_current_changes(publishable)
 441         return compile_text(change.materialize() for change in changes)
 442
 443     def wldocument(self, publishable=True, changes=None,
 444                    parse_dublincore=True, strict=False, librarian2=False):
 445         from documents.ebook_utils import RedakcjaDocProvider
 446         from librarian.parser import WLDocument
 447         from librarian.document import WLDocument as WLDocument2
 448
 449         provider = RedakcjaDocProvider(publishable=publishable)
 450         xml = self.materialize(publishable=publishable, changes=changes).encode('utf-8')
 451
 452         if librarian2:
 453             return WLDocument2(
 454                 BytesIO(xml),
 455                 provider=provider)
 456         return WLDocument.from_bytes(
 457                 xml,
 458                 provider=provider,
 459                 parse_dublincore=parse_dublincore,
 460                 strict=strict)
 461
 462     def publish(self, user, fake=False, host=None, days=0, beta=False, hidden=False):
 463         """
 464             Publishes a book on behalf of a (local) user.
 465         """
 466         self.assert_publishable()
 467         changes = self.get_current_changes(publishable=True)
 468         if not fake:
 469             book_xml = self.materialize(changes=changes)
 470             data = {"book_xml": book_xml, "days": days, "hidden": hidden}
 471             if host:
 472                 data['gallery_url'] = host + self.gallery_url()
 473             apiclient.api_call(user, "books/", data, beta=beta)
 474         if not beta:
 475             # record the publish
 476             br = BookPublishRecord.objects.create(book=self, user=user)
 477             for c in changes:
 478                 ChunkPublishRecord.objects.create(book_record=br, change=c)
 479             if not self.public and days == 0:
 480                 self.public = True
 481                 self.save()
 482             if self.public and days > 0:
 483                 self.public = False
 484                 self.save()
 485             post_publish.send(sender=br)
 486
 487     def latex_dir(self):
 488         doc = self.wldocument()
 489         return doc.latex_dir(cover=True, ilustr_path=self.gallery_path())