X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/6d399587b7c6aeedda4dd12af90ceafaeac65d59..968cb1f19a746aed062a8ee64f6d0f0dd5969e23:/apps/catalogue/models.py diff --git a/apps/catalogue/models.py b/apps/catalogue/models.py index dd251c16a..d3b24d379 100644 --- a/apps/catalogue/models.py +++ b/apps/catalogue/models.py @@ -207,7 +207,7 @@ class Book(models.Model): def __unicode__(self): return self.title - def save(self, force_insert=False, force_update=False, reset_short_html=True, refresh_mp3=True): + def save(self, force_insert=False, force_update=False, reset_short_html=True, refresh_mp3=True, **kwargs): if reset_short_html: # Reset _short_html during save update = {} @@ -263,8 +263,8 @@ class Book(models.Model): formats.append(u'%s' % (reverse('book_text', kwargs={'slug': self.slug}), _('Read online'))) if self.pdf_file: formats.append(u'PDF' % self.pdf_file.url) - if self.epub_file: - formats.append(u'EPUB' % self.epub_file.url) + if self.root_ancestor.epub_file: + formats.append(u'EPUB' % self.root_ancestor.epub_file.url) if self.odt_file: formats.append(u'ODT' % self.odt_file.url) if self.txt_file: @@ -282,6 +282,18 @@ class Book(models.Model): return mark_safe(getattr(self, key)) + @property + def root_ancestor(self): + """ returns the oldest ancestor """ + + if not hasattr(self, '_root_ancestor'): + book = self + while book.parent: + book = book.parent + self._root_ancestor = book + return self._root_ancestor + + def get_mp3_info(self): """Retrieves artist and director names from audio ID3 tags.""" audio = id3.ID3(self.mp3_file.path) @@ -314,6 +326,52 @@ class Book(models.Model): has_html_file.short_description = 'HTML' has_html_file.boolean = True + def build_epub(self, remove_descendants=True): + """ (Re)builds the epub file. + If book has a parent, does nothing. + Unless remove_descendants is False, descendants' epubs are removed. + """ + + from StringIO import StringIO + from hashlib import sha1 + from django.core.files.base import ContentFile + from librarian import DocProvider + + class BookImportDocProvider(DocProvider): + """ used for joined EPUBs """ + + def __init__(self, book): + self.book = book + + def by_slug(self, slug): + if slug == self.book.slug: + return self.book.xml_file + else: + return Book.objects.get(slug=slug).xml_file + + if self.parent: + # don't need an epub + return + + epub_file = StringIO() + try: + epub.transform(BookImportDocProvider(self), self.slug, epub_file) + self.epub_file.save('%s.epub' % self.slug, ContentFile(epub_file.getvalue()), save=False) + self.save() + FileRecord(slug=self.slug, type='epub', sha1=sha1(epub_file.getvalue()).hexdigest()).save() + except NoDublinCore: + pass + + book_descendants = list(self.children.all()) + while len(book_descendants) > 0: + child_book = book_descendants.pop(0) + if remove_descendants and child_book.has_epub_file(): + child_book.epub_file.delete() + # save anyway, to refresh short_html + child_book.save() + book_descendants += list(child_book.children.all()) + + @classmethod def from_xml_file(cls, xml_file, overwrite=False): # use librarian to parse meta-data @@ -332,10 +390,7 @@ class Book(models.Model): from tempfile import NamedTemporaryFile from slughifi import slughifi from markupstring import MarkupString - from hashlib import sha1 - from django.core.files.base import ContentFile from django.core.files.storage import default_storage - from StringIO import StringIO # Read book metadata book_base, book_slug = book_info.url.rsplit('/', 1) @@ -388,15 +443,6 @@ class Book(models.Model): except Book.DoesNotExist, e: raise Book.DoesNotExist(_('Book with slug = "%s" does not exist.') % slug) - book_descendants = list(book.children.all()) - while len(book_descendants) > 0: - child_book = book_descendants.pop(0) - child_book.tags = list(child_book.tags) + [book_tag] - child_book.save() - for fragment in child_book.fragments.all(): - fragment.tags = set(list(fragment.tags) + [book_tag]) - book_descendants += list(child_book.children.all()) - # Save XML and HTML files book.xml_file.save('%s.xml' % book.slug, raw_file, save=False) @@ -404,15 +450,6 @@ class Book(models.Model): if html.transform(book.xml_file.path, html_file, parse_dublincore=False): book.html_file.save('%s.html' % book.slug, File(html_file), save=False) - # Create EPUB - epub_file = StringIO() - try: - epub.transform(book.xml_file, epub_file) - book.epub_file.save('%s.epub' % book.slug, ContentFile(epub_file.getvalue()), save=False) - FileRecord(slug=book.slug, type='epub', sha1=sha1(epub_file.getvalue()).hexdigest()).save() - except NoDublinCore: - pass - # Extract fragments closed_fragments, open_fragments = html.extract_fragments(book.html_file.path) for fragment in closed_fragments.values(): @@ -443,6 +480,21 @@ class Book(models.Model): new_fragment.save() new_fragment.tags = set(book_tags + themes + [book_tag]) + book.build_epub(remove_descendants=False) + + book_descendants = list(book.children.all()) + # add l-tag to descendants and their fragments + # delete unnecessary EPUB files + while len(book_descendants) > 0: + child_book = book_descendants.pop(0) + child_book.tags = list(child_book.tags) + [book_tag] + if child_book.has_epub_file(): + child_book.epub_file.delete() + child_book.save() + for fragment in child_book.fragments.all(): + fragment.tags = set(list(fragment.tags) + [book_tag]) + book_descendants += list(child_book.children.all()) + # refresh cache book.tag_counter book.theme_counter