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 from urllib.parse import urljoin
  27
  28
  29 class Book(models.Model):
  30     """ A document edited on the wiki """
  31
  32     title = models.CharField(_('title'), max_length=255, db_index=True)
  33     slug = models.SlugField(_('slug'), max_length=128, unique=True, db_index=True)
  34     public = models.BooleanField(_('public'), default=True, db_index=True)
  35     gallery = models.CharField(_('scan gallery name'), max_length=255, blank=True)
  36     project = models.ForeignKey(Project, models.SET_NULL, null=True, blank=True)
  37
  38     parent = models.ForeignKey('self', models.SET_NULL, null=True, blank=True, verbose_name=_('parent'), related_name="children", editable=False)
  39     parent_number = models.IntegerField(_('parent number'), null=True, blank=True, db_index=True, editable=False)
  40
  41     # Cache
  42     _single = models.BooleanField(editable=False, null=True, db_index=True)
  43     _new_publishable = models.BooleanField(editable=False, null=True)
  44     _published = models.BooleanField(editable=False, null=True)
  45     _on_track = models.IntegerField(null=True, blank=True, db_index=True, editable=False)
  46     dc_cover_image = models.ForeignKey(Image, blank=True, null=True,
  47         db_index=True, on_delete=models.SET_NULL, editable=False)
  48     dc = models.JSONField(null=True, editable=False)
  49     cover = models.FileField(blank=True, upload_to='documents/cover')
  50
  51     dc_slug = models.CharField(
  52         max_length=2048,
  53         null=True, blank=True,
  54         editable=False,
  55     )
  56     catalogue_book = models.ForeignKey(
  57         'catalogue.Book',
  58         models.PROTECT,
  59         null=True, blank=True,
  60         editable=False,
  61         related_name='document_books',
  62         related_query_name='document_book',
  63     )
  64
  65     class NoTextError(BaseException):
  66         pass
  67
  68     class Meta:
  69         app_label = 'documents'
  70         ordering = ['title', 'slug']
  71         verbose_name = _('book')
  72         verbose_name_plural = _('books')
  73
  74     @classmethod
  75     def get_visible_for(cls, user):
  76         qs = cls.objects.all()
  77         if not user.is_authenticated:
  78             qs = qs.filter(public=True)
  79         return qs
  80
  81     @staticmethod
  82     def q_dc(field, field_plural, value, prefix=''):
  83         if connection.features.supports_json_field_contains:
  84             return models.Q(**{f'{prefix}dc__{field_plural}__contains': value})
  85         else:
  86             return models.Q(**{f'{prefix}dc__{field}': value})
  87
  88
  89     # Representing
  90     # ============
  91
  92     def __iter__(self):
  93         return iter(self.chunk_set.all())
  94
  95     def __getitem__(self, chunk):
  96         return self.chunk_set.all()[chunk]
  97
  98     def __len__(self):
  99         return self.chunk_set.count()
 100
 101     def __bool__(self):
 102         """
 103             Necessary so that __len__ isn't used for bool evaluation.
 104         """
 105         return True
 106
 107     def __str__(self):
 108         return self.title
 109
 110     def get_absolute_url(self):
 111         return reverse("documents_book", args=[self.slug])
 112
 113     def correct_about(self):
 114         return "http://%s%s" % (
 115             Site.objects.get_current().domain,
 116             self.get_absolute_url()
 117         )
 118
 119     def gallery_path(self):
 120         return os.path.join(settings.MEDIA_ROOT, settings.IMAGE_DIR, self.gallery)
 121
 122     def gallery_url(self):
 123         return '%s%s%s/' % (settings.MEDIA_URL, settings.IMAGE_DIR, self.gallery)
 124
 125     # Creating & manipulating
 126     # =======================
 127
 128     def accessible(self, request):
 129         return self.public or request.user.is_authenticated
 130
 131     @classmethod
 132     @transaction.atomic
 133     def create(cls, creator, text, *args, **kwargs):
 134         b = cls.objects.create(*args, **kwargs)
 135         b.chunk_set.all().update(creator=creator)
 136         b[0].commit(text, author=creator)
 137         return b
 138
 139     def add(self, *args, **kwargs):
 140         """Add a new chunk at the end."""
 141         return self.chunk_set.reverse()[0].split(*args, **kwargs)
 142
 143     @classmethod
 144     @transaction.atomic
 145     def import_xml_text(cls, text=u'', previous_book=None,
 146                 commit_args=None, **kwargs):
 147         """Imports a book from XML, splitting it into chunks as necessary."""
 148         texts = split_xml(text)
 149         if previous_book:
 150             instance = previous_book
 151         else:
 152             instance = cls(**kwargs)
 153             instance.save()
 154
 155         # if there are more parts, set the rest to empty strings
 156         book_len = len(instance)
 157         for i in range(book_len - len(texts)):
 158             texts.append((u'pusta część %d' % (i + 1), u''))
 159
 160         i = 0
 161         for i, (title, text) in enumerate(texts):
 162             if not title:
 163                 title = u'część %d' % (i + 1)
 164
 165             slug = slugify(title)
 166
 167             if i < book_len:
 168                 chunk = instance[i]
 169                 chunk.slug = slug[:50]
 170                 chunk.title = title[:255]
 171                 chunk.save()
 172             else:
 173                 chunk = instance.add(slug, title)
 174
 175             chunk.commit(text, **commit_args)
 176
 177         return instance
 178
 179     def make_chunk_slug(self, proposed):
 180         """
 181             Finds a chunk slug not yet used in the book.
 182         """
 183         slugs = set(c.slug for c in self)
 184         i = 1
 185         new_slug = proposed[:50]
 186         while new_slug in slugs:
 187             new_slug = "%s_%d" % (proposed[:45], i)
 188             i += 1
 189         return new_slug
 190
 191     @transaction.atomic
 192     def append(self, other, slugs=None, titles=None):
 193         """Add all chunks of another book to self."""
 194         assert self != other
 195
 196         number = self[len(self) - 1].number + 1
 197         len_other = len(other)
 198         single = len_other == 1
 199
 200         if slugs is not None:
 201             assert len(slugs) == len_other
 202         if titles is not None:
 203             assert len(titles) == len_other
 204             if slugs is None:
 205                 slugs = [slugify(t) for t in titles]
 206
 207         for i, chunk in enumerate(other):
 208             # move chunk to new book
 209             chunk.book = self
 210             chunk.number = number
 211
 212             if titles is None:
 213                 # try some title guessing
 214                 if other.title.startswith(self.title):
 215                     other_title_part = other.title[len(self.title):].lstrip(' /')
 216                 else:
 217                     other_title_part = other.title
 218
 219                 if single:
 220                     # special treatment for appending one-parters:
 221                     # just use the guessed title and original book slug
 222                     chunk.title = other_title_part
 223                     if other.slug.startswith(self.slug):
 224                         chunk.slug = other.slug[len(self.slug):].lstrip('-_')
 225                     else:
 226                         chunk.slug = other.slug
 227                 else:
 228                     chunk.title = ("%s, %s" % (other_title_part, chunk.title))[:255]
 229             else:
 230                 chunk.slug = slugs[i]
 231                 chunk.title = titles[i]
 232
 233             chunk.slug = self.make_chunk_slug(chunk.slug)
 234             chunk.save()
 235             number += 1
 236         assert not other.chunk_set.exists()
 237
 238         gm = GalleryMerger(self.gallery, other.gallery)
 239         self.gallery = gm.merge()
 240
 241         # and move the gallery starts
 242         if gm.was_merged:
 243                 for chunk in self[len(self) - len_other:]:
 244                         old_start = chunk.gallery_start or 1
 245                         chunk.gallery_start = old_start + gm.dest_size - gm.num_deleted
 246                         chunk.save()
 247
 248         other.delete()
 249
 250
 251     @transaction.atomic
 252     def prepend_history(self, other):
 253         """Prepend history from all the other book's chunks to own."""
 254         assert self != other
 255
 256         for i in range(len(self), len(other)):
 257             title = u"pusta część %d" % i
 258             chunk = self.add(slugify(title), title)
 259             chunk.commit('')
 260
 261         for i in range(len(other)):
 262             self[i].prepend_history(other[0])
 263
 264         assert not other.chunk_set.exists()
 265         other.delete()
 266
 267     def split(self):
 268         """Splits all the chunks into separate books."""
 269         self.title
 270         for chunk in self:
 271             book = Book.objects.create(title=chunk.title, slug=chunk.slug,
 272                     public=self.public, gallery=self.gallery)
 273             book[0].delete()
 274             chunk.book = book
 275             chunk.number = 1
 276             chunk.save()
 277         assert not self.chunk_set.exists()
 278         self.delete()
 279
 280     # State & cache
 281     # =============
 282
 283     def last_published(self):
 284         try:
 285             return self.publish_log.all()[0].timestamp
 286         except IndexError:
 287             return None
 288
 289     def assert_publishable(self):
 290         assert self.chunk_set.exists(), _('No chunks in the book.')
 291         try:
 292             changes = self.get_current_changes(publishable=True)
 293         except self.NoTextError:
 294             raise AssertionError(_('Not all chunks have approved revisions.'))
 295
 296         from librarian import NoDublinCore, ParseError, ValidationError
 297
 298         try:
 299             bi = self.wldocument(changes=changes, strict=True).book_info
 300         except ParseError as e:
 301             raise AssertionError(_('Invalid XML') + ': ' + str(e))
 302         except NoDublinCore:
 303             raise AssertionError(_('No Dublin Core found.'))
 304         except ValidationError as e:
 305             raise AssertionError(_('Invalid Dublin Core') + ': ' + str(e))
 306
 307         valid_about = self.correct_about()
 308         assert bi.about == valid_about, _("rdf:about is not") + " " + valid_about
 309
 310     def publishable_error(self):
 311         try:
 312             return self.assert_publishable()
 313         except AssertionError as e:
 314             return e
 315         else:
 316             return None
 317
 318     def hidden(self):
 319         return self.slug.startswith('.')
 320
 321     def is_new_publishable(self):
 322         """Checks if book is ready for publishing.
 323
 324         Returns True if there is a publishable version newer than the one
 325         already published.
 326
 327         """
 328         new_publishable = False
 329         if not self.chunk_set.exists():
 330             return False
 331         for chunk in self:
 332             change = chunk.publishable()
 333             if not change:
 334                 return False
 335             if not new_publishable and not change.publish_log.exists():
 336                 new_publishable = True
 337         return new_publishable
 338     new_publishable = cached_in_field('_new_publishable')(is_new_publishable)
 339
 340     def is_published(self):
 341         return self.publish_log.exists()
 342     published = cached_in_field('_published')(is_published)
 343
 344     def get_on_track(self):
 345         if self.published:
 346             return -1
 347         stages = [ch.stage.ordering if ch.stage is not None else 0
 348                     for ch in self]
 349         if not len(stages):
 350             return 0
 351         return min(stages)
 352     on_track = cached_in_field('_on_track')(get_on_track)
 353
 354     def is_single(self):
 355         return len(self) == 1
 356     single = cached_in_field('_single')(is_single)
 357
 358     def book_info(self, publishable=True):
 359         try:
 360             book_xml = self.materialize(publishable=publishable)
 361         except self.NoTextError:
 362             pass
 363         else:
 364             from librarian.dcparser import BookInfo
 365             from librarian import NoDublinCore, ParseError, ValidationError
 366             try:
 367                 return BookInfo.from_bytes(book_xml.encode('utf-8'))
 368             except (self.NoTextError, ParseError, NoDublinCore, ValidationError):
 369                 return None
 370
 371     def refresh_dc_cache(self):
 372         update = {
 373             'dc_slug': None,
 374             'dc_cover_image': None,
 375         }
 376
 377         info = self.book_info()
 378         if info is not None:
 379             update['dc_slug'] = info.url.slug
 380             if info.cover_source:
 381                 try:
 382                     image = Image.objects.get(pk=int(info.cover_source.rstrip('/').rsplit('/', 1)[-1]))
 383                 except:
 384                     pass
 385                 else:
 386                     if info.cover_source.rstrip('/') == image.get_full_url().rstrip('/'):
 387                         update['dc_cover_image'] = image
 388             update['dc'] = info.to_dict()
 389         Book.objects.filter(pk=self.pk).update(**update)
 390
 391     def touch(self):
 392         update = {
 393             "_new_publishable": self.is_new_publishable(),
 394             "_published": self.is_published(),
 395             "_single": self.is_single(),
 396             "_on_track": self.get_on_track(),
 397         }
 398         Book.objects.filter(pk=self.pk).update(**update)
 399         self.refresh_dc_cache()
 400         self.build_cover()
 401
 402     def build_cover(self):
 403         width, height = 212, 300
 404         try:
 405             xml = self.materialize(publishable=True).encode('utf-8')
 406             info = BookInfo.from_bytes(xml)
 407             kwargs = {}
 408             if self.project is not None:
 409                 if self.project.logo_mono or self.project.logo:
 410                     kwargs['cover_logo'] = (self.project.logo_mono or self.project.logo).path
 411             cover = make_cover(info, width=width, height=height, **kwargs)
 412             out = BytesIO()
 413             ext = cover.ext()
 414             cover.save(out)
 415             self.cover.save(f'{self.slug}.{ext}', out, save=False)
 416             type(self).objects.filter(pk=self.pk).update(cover=self.cover)
 417         except:
 418             type(self).objects.filter(pk=self.pk).update(cover='')
 419
 420     # Materializing & publishing
 421     # ==========================
 422
 423     def get_current_changes(self, publishable=True):
 424         """
 425             Returns a list containing one Change for every Chunk in the Book.
 426             Takes the most recent revision (publishable, if set).
 427             Throws an error, if a proper revision is unavailable for a Chunk.
 428         """
 429         if publishable:
 430             changes = [chunk.publishable() for chunk in self]
 431         else:
 432             changes = [chunk.head for chunk in self if chunk.head is not None]
 433         if None in changes:
 434             raise self.NoTextError('Some chunks have no available text.')
 435         return changes
 436
 437     def materialize(self, publishable=False, changes=None):
 438         """
 439             Get full text of the document compiled from chunks.
 440             Takes the current versions of all texts
 441             or versions most recently tagged for publishing,
 442             or a specified iterable changes.
 443         """
 444         if changes is None:
 445             changes = self.get_current_changes(publishable)
 446         return compile_text(change.materialize() for change in changes)
 447
 448     def wldocument(self, publishable=True, changes=None,
 449                    parse_dublincore=True, strict=False, librarian2=False):
 450         from documents.ebook_utils import RedakcjaDocProvider
 451         from librarian.parser import WLDocument
 452         from librarian.document import WLDocument as WLDocument2
 453
 454         provider = RedakcjaDocProvider(publishable=publishable)
 455         xml = self.materialize(publishable=publishable, changes=changes).encode('utf-8')
 456
 457         if librarian2:
 458             return WLDocument2(
 459                 BytesIO(xml),
 460                 provider=provider)
 461         return WLDocument.from_bytes(
 462                 xml,
 463                 provider=provider,
 464                 parse_dublincore=parse_dublincore,
 465                 strict=strict)
 466
 467     def publish(self, user, fake=False, host=None, days=0, beta=False, hidden=False):
 468         """
 469             Publishes a book on behalf of a (local) user.
 470         """
 471         self.assert_publishable()
 472         changes = self.get_current_changes(publishable=True)
 473         if not fake:
 474             book_xml = self.materialize(changes=changes)
 475             data = {"book_xml": book_xml, "days": days, "hidden": hidden}
 476             if self.project is not None:
 477                 if self.project.logo:
 478                     data['logo'] = urljoin(
 479                         'https://' + Site.objects.get_current().domain,
 480                         self.project.logo.url,
 481                     )
 482                 if self.project.logo_mono:
 483                     data['logo_mono'] = urljoin(
 484                         'https://' + Site.objects.get_current().domain,
 485                         self.project.logo_mono.url,
 486                     )
 487                 if self.project.logo_alt:
 488                     data['logo_alt'] = self.project.logo_alt
 489             if host:
 490                 data['gallery_url'] = host + self.gallery_url()
 491             apiclient.api_call(user, "books/", data, beta=beta)
 492         if not beta:
 493             # record the publish
 494             br = BookPublishRecord.objects.create(book=self, user=user)
 495             for c in changes:
 496                 ChunkPublishRecord.objects.create(book_record=br, change=c)
 497             if not self.public and days == 0:
 498                 self.public = True
 499                 self.save()
 500             if self.public and days > 0:
 501                 self.public = False
 502                 self.save()
 503             post_publish.send(sender=br)
 504
 505     def latex_dir(self):
 506         doc = self.wldocument()
 507         return doc.latex_dir(cover=True, ilustr_path=self.gallery_path())