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