X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/d1d6ea8ed2889f871b6d89d4fae6021f286f921a..a3b153ea6e30da275edf3e34b8d2f8ca1bb436cc:/src/catalogue/views.py?ds=inline diff --git a/src/catalogue/views.py b/src/catalogue/views.py index 1f2db19c6..817b3ada5 100644 --- a/src/catalogue/views.py +++ b/src/catalogue/views.py @@ -1,73 +1,61 @@ -# -*- coding: utf-8 -*- # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later. # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information. # from collections import OrderedDict import random +import re +from urllib.parse import quote_plus from django.conf import settings -from django.http.response import HttpResponseForbidden from django.template.loader import render_to_string from django.shortcuts import get_object_or_404, render, redirect from django.http import HttpResponse, HttpResponseRedirect, Http404, HttpResponsePermanentRedirect -from django.core.urlresolvers import reverse +from django.urls import reverse from django.db.models import Q, QuerySet from django.contrib.auth.decorators import login_required, user_passes_test -from django.utils.http import urlquote_plus from django.utils import translation -from django.utils.translation import ugettext as _, ugettext_lazy +from django.utils.translation import gettext as _, gettext_lazy +from django.views.decorators.cache import never_cache +from django.views.generic import TemplateView from ajaxable.utils import AjaxableFormView +from club.forms import DonationStep1Form +from club.models import Club +from annoy.models import DynamicTextInsert from pdcounter import views as pdcounter_views -from paypal.rest import user_is_subscribed from picture.models import Picture, PictureArea -from ssify import ssi_included, ssi_expect, SsiVariable as Var +from wolnelektury.utils import is_ajax from catalogue import constants from catalogue import forms from catalogue.helpers import get_top_level_related_tags from catalogue.models import Book, Collection, Tag, Fragment +from catalogue.models.tag import TagRelation from catalogue.utils import split_tags from catalogue.models.tag import prefetch_relations -from wolnelektury.utils import is_crawler staff_required = user_passes_test(lambda user: user.is_staff) def catalogue(request): return render(request, 'catalogue/catalogue.html', { - 'books': Book.objects.filter(parent=None), + 'books': Book.objects.filter(findable=True, parent=None), 'pictures': Picture.objects.all(), - 'collections': Collection.objects.all(), + 'collections': Collection.objects.filter(listed=True), 'active_menu_item': 'all_works', }) -def book_list(request, filters=None, template_name='catalogue/book_list.html', - nav_template_name='catalogue/snippets/book_list_nav.html', - list_template_name='catalogue/snippets/book_list.html'): - """ generates a listing of all books, optionally filtered """ - books_by_author, orphans, books_by_parent = Book.book_list(filters) - books_nav = OrderedDict() - for tag in books_by_author: - if books_by_author[tag]: - books_nav.setdefault(tag.sort_key[0], []).append(tag) - return render(request, template_name, { - 'rendered_nav': render_to_string(nav_template_name, {'books_nav': books_nav}), - 'rendered_book_list': render_to_string(list_template_name, { - 'books_by_author': books_by_author, - 'orphans': orphans, - 'books_by_parent': books_by_parent, - }) - }) - - def daisy_list(request): - return book_list(request, Q(media__type='daisy'), template_name='catalogue/daisy_list.html') + return object_list(request, Book.objects.filter(media__type='daisy')) def collection(request, slug): coll = get_object_or_404(Collection, slug=slug) - return render(request, 'catalogue/collection.html', {'collection': coll}) + template_name = 'catalogue/2022/collection.html' + return render(request, template_name, { + 'collection': coll, + 'active_menu_item': 'collections', + }) def differentiate_tags(request, tags, ambiguous_slugs): @@ -81,10 +69,250 @@ def differentiate_tags(request, tags, ambiguous_slugs): }) return render( request, - 'catalogue/differentiate_tags.html', {'tags': tags, 'options': options, 'unparsed': ambiguous_slugs[1:]}) + 'catalogue/differentiate_tags.html', + {'tags': tags, 'options': options, 'unparsed': ambiguous_slugs[1:]} + ) + + +from django.db.models import FilteredRelation, Q +from django.views.decorators.cache import cache_control +from django.views.decorators.vary import vary_on_headers +from django.utils.decorators import method_decorator + + +@method_decorator([ + vary_on_headers('X-Requested-With'), + cache_control(max_age=1), +], 'dispatch') +class ObjectListView(TemplateView): + page_size = 100 + themed_page_size = 10 + item_template_name = '' + orderings = {} + default_ordering = None + + def analyse(self): + self.is_themed = False + self.ctx = ctx = {} + ctx['tags'] = [] + + def dispatch(self, *args, **kwargs): + try: + self.analyse() + except ResponseInstead as e: + return e.response + return super().dispatch(*args, **kwargs) + + def get_orderings(self): + order = self.get_order() + return [ + { + "slug": k, + "name": v[1], + "active": k == order, + "default": v[0] is None, + } + for k, v in self.orderings.items() + ] + + def get_order(self): + order = self.request.GET.get('order') + if order not in self.orderings: + order = self.default_ordering + return order + + def order(self, qs): + order_tag = self.get_order() + if order_tag: + order = self.orderings[order_tag] + order_by = order[0] + if order_by: + qs = qs.order_by(order_by) + return qs + + def search(self, qs): + return qs + + def get_template_names(self): + if is_ajax(self.request) or self.request.GET.get('dyn'): + if self.is_themed: + return self.dynamic_themed_template_name + else: + return self.dynamic_template_name + else: + if self.is_themed: + return self.themed_template_name + else: + return self.template_name + + def get_context_data(self, **kwargs): + ctx = super().get_context_data() + ctx.update(self.ctx) + qs = self.get_queryset() + qs = self.search(qs) + qs = self.order(qs) + + ctx['object_list'] = qs + ctx['suggested_tags'] = self.get_suggested_tags(qs) + ctx['suggested_tags_by_category'] = split_tags(ctx['suggested_tags']) + return ctx + + +class BookList(ObjectListView): + title = gettext_lazy('Literature') + list_type = 'books' + template_name = 'catalogue/2022/book_list.html' + dynamic_template_name = 'catalogue/2022/dynamic_book_list.html' + themed_template_name = 'catalogue/2022/themed_book_list.html' + dynamic_themed_template_name = 'catalogue/2022/dynamic_themed_book_list.html' + + orderings = { + 'pop': ('-popularity__count', 'najpopularniejsze'), + 'alpha': (None, 'alfabetycznie'), + } + default_ordering = 'alpha' + + def get_queryset(self): + return Book.objects.filter(parent=None, findable=True) + + def search(self, qs): + term = self.request.GET.get('search') + if term: + meta_rels = TagRelation.objects.filter(tag__category='author') + # TODO: search tags in currently displaying language + if self.is_themed: + rels = meta_rels.filter(tag__name_pl__icontains=term) + qs = qs.filter( + Q(book__title__icontains=term) | + Q(tag_relations__in=rels) | + Q(text__icontains=term) + ).distinct() + else: + qs = qs.annotate( + meta=FilteredRelation('tag_relations', condition=Q(tag_relations__in=meta_rels)) + ) + qs = qs.filter(Q(title__icontains=term) | Q(meta__tag__name_pl__icontains=term)).distinct() + return qs + + +class ArtList(ObjectListView): + template_name = 'catalogue/2022/book_list.html' + dynamic_template_name = 'catalogue/2022/dynamic_book_list.html' + title = gettext_lazy('Art') + list_type = 'gallery' + + def get_queryset(self): + return Picture.objects.all() + + def search(self, qs): + term = self.request.GET.get('search') + if term: + qs = qs.filter(Q(title__icontains=term) | Q(tag_relations__tag__name_pl__icontains=term)).distinct() + return qs + + +class LiteratureView(BookList): + def get_suggested_tags(self, queryset): + tags = list(get_top_level_related_tags([])) + tags.sort(key=lambda t: -t.count) + if self.request.user.is_authenticated: + tags.extend(list(Tag.objects.filter(user=self.request.user).exclude(name=''))) + return tags + + +class AudiobooksView(LiteratureView): + title = gettext_lazy('Audiobooks') + list_type = 'audiobooks' + + def get_queryset(self): + return Book.objects.filter(findable=True, media__type='mp3').distinct() + + +class GalleryView(ArtList): + def get_suggested_tags(self, queryset): + return Tag.objects.usage_for_queryset( + queryset, + counts=True + ).exclude(pk__in=[t.id for t in self.ctx['tags']]).order_by('-count') + + +class TaggedObjectList(BookList): + def analyse(self): + super().analyse() + self.ctx['tags'] = analyse_tags(self.request, self.kwargs['tags']) + self.ctx['fragment_tags'] = [t for t in self.ctx['tags'] if t.category in ('theme', 'object')] + self.ctx['work_tags'] = [t for t in self.ctx['tags'] if t not in self.ctx['fragment_tags']] + self.is_themed = self.ctx['has_theme'] = bool(self.ctx['fragment_tags']) + if self.is_themed: + self.ctx['main_tag'] = self.ctx['fragment_tags'][0] + elif self.ctx['tags']: + self.ctx['main_tag'] = self.ctx['tags'][0] + else: + self.ctx['main_tag'] = None + self.ctx['filtering_tags'] = [ + t for t in self.ctx['tags'] + if t is not self.ctx['main_tag'] + ] + + def get_queryset(self): + qs = Book.tagged.with_all(self.ctx['work_tags']).filter(findable=True) + qs = qs.exclude(ancestor__in=qs) + if self.is_themed: + fqs = Fragment.tagged.with_all(self.ctx['fragment_tags']) + if self.ctx['work_tags']: + fqs = fqs.filter( + Q(book__in=qs) | Q(book__ancestor__in=qs) + ) + qs = fqs + return qs + + def get_suggested_tags(self, queryset): + tag_ids = [t.id for t in self.ctx['tags']] + if self.is_themed: + related_tags = [] + current_books = self.get_queryset().values_list('book', flat=True).distinct() + containing_books = Book.objects.filter(Q(id__in=current_books) | Q(children__in=current_books)) + + related_tags.extend(list( + Tag.objects.usage_for_queryset( + containing_books, + ).exclude(category='set').exclude(pk__in=tag_ids) + )) + if self.request.user.is_authenticated: + related_tags.extend(list( + Tag.objects.usage_for_queryset( + containing_books + ).filter( + user=self.request.user + ).exclude(name='').exclude(pk__in=tag_ids) + )) + else: + related_tags = list(get_top_level_related_tags(self.ctx['tags'])) + if self.request.user.is_authenticated: + qs = Book.tagged.with_all(self.ctx['tags']).filter(findable=True) + related_tags.extend(list( + Tag.objects.usage_for_queryset( + qs + ).filter( + user=self.request.user + ).exclude(name='').exclude(pk__in=tag_ids) + )) + + fragments = Fragment.objects.filter( + Q(book__in=queryset) | Q(book__ancestor__in=queryset) + ) + related_tags.extend( + Tag.objects.usage_for_queryset( + fragments, counts=True + ).filter(category__in=('theme', 'object')).exclude(pk__in=tag_ids) + .only('name', 'sort_key', 'category', 'slug')) + return related_tags -def object_list(request, objects, fragments=None, related_tags=None, tags=None, list_type='books', extra=None): + + +def object_list(request, objects, fragments=None, related_tags=None, tags=None, + list_type='books', extra=None): if not tags: tags = [] tag_ids = [tag.pk for tag in tags] @@ -94,7 +322,17 @@ def object_list(request, objects, fragments=None, related_tags=None, tags=None, related_tag_lists.append(related_tags) else: related_tag_lists.append( - Tag.objects.usage_for_queryset(objects, counts=True).exclude(category='set').exclude(pk__in=tag_ids)) + Tag.objects.usage_for_queryset( + objects, counts=True + ).exclude(category='set').exclude(pk__in=tag_ids)) + if request.user.is_authenticated: + related_tag_lists.append( + Tag.objects.usage_for_queryset( + objects, counts=True + ).filter( + user=request.user + ).exclude(name='').exclude(pk__in=tag_ids) + ) if not (extra and extra.get('theme_is_set')): if fragments is None: if list_type == 'gallery': @@ -102,12 +340,18 @@ def object_list(request, objects, fragments=None, related_tags=None, tags=None, else: fragments = Fragment.objects.filter(book__in=objects) related_tag_lists.append( - Tag.objects.usage_for_queryset(fragments, counts=True).filter(category='theme').exclude(pk__in=tag_ids) + Tag.objects.usage_for_queryset( + fragments, counts=True + ).filter(category='theme').exclude(pk__in=tag_ids) .only('name', 'sort_key', 'category', 'slug')) if isinstance(objects, QuerySet): objects = prefetch_relations(objects, 'author') + categories = split_tags(*related_tag_lists) + suggest = [] + for c in ['set', 'author', 'epoch', 'kind', 'genre']: + suggest.extend(sorted(categories[c], key=lambda t: -t.count)) objects = list(objects) @@ -123,8 +367,10 @@ def object_list(request, objects, fragments=None, related_tags=None, tags=None, result = { 'object_list': objects, 'categories': categories, + 'suggest': suggest, 'list_type': list_type, 'tags': tags, + 'main_tag': tags[0] if tags else None, 'formats_form': forms.DownloadFormatsForm(), 'best': best, @@ -132,28 +378,19 @@ def object_list(request, objects, fragments=None, related_tags=None, tags=None, } if extra: result.update(extra) + + has_theme = any(((theme := x).category == 'theme' for x in tags)) + if has_theme: + result['main_tag'] = theme + template = 'catalogue/2022/theme_detail.html' + else: + template = 'catalogue/2022/author_detail.html' + return render( - request, - 'catalogue/tagged_object_list.html', result, + request, template, result, ) -def literature(request): - books = Book.objects.filter(parent=None) - return object_list(request, books, related_tags=get_top_level_related_tags([])) - - -def gallery(request): - return object_list(request, Picture.objects.all(), list_type='gallery') - - -def audiobooks(request): - audiobooks = Book.objects.filter(media__type__in=('mp3', 'ogg')).distinct() - return object_list(request, audiobooks, list_type='audiobooks', extra={ - 'daisy': Book.objects.filter(media__type='daisy').distinct(), - }) - - class ResponseInstead(Exception): def __init__(self, response): super(ResponseInstead, self).__init__() @@ -169,15 +406,16 @@ def analyse_tags(request, tag_str): chunks = tag_str.split('/') if len(chunks) == 2 and chunks[0] == 'autor': raise ResponseInstead(pdcounter_views.author_detail(request, chunks[1])) - else: - raise Http404 - except Tag.MultipleObjectsReturned, e: + raise Http404 + except Tag.MultipleObjectsReturned as e: # Ask the user to disambiguate raise ResponseInstead(differentiate_tags(request, e.tags, e.ambiguous_slugs)) - except Tag.UrlDeprecationWarning, e: + except Tag.UrlDeprecationWarning as e: raise ResponseInstead(HttpResponsePermanentRedirect( reverse('tagged_object_list', args=['/'.join(tag.url_chunk for tag in e.tags)]))) + if not tags: + raise Http404 try: if len(tags) > settings.MAX_TAG_LIST: raise Http404 @@ -199,6 +437,8 @@ def theme_list(request, tags, list_type): # TODO: Pictures on shelves not supported yet. books = Book.tagged.with_all(shelf_tags).order_by() fragments = fragments.filter(Q(book__in=books) | Q(book__ancestor__in=books)) + elif list_type == 'books': + fragments = fragments.filter(book__findable=True) if not fragments and len(tags) == 1 and list_type == 'books': if PictureArea.tagged.with_any(tags).exists() or Picture.tagged.with_any(tags).exists(): @@ -211,14 +451,14 @@ def theme_list(request, tags, list_type): def tagged_object_list(request, tags, list_type): + if list_type in ('books', 'audiobooks'): + return TaggedObjectList.as_view()(request, tags=tags) + try: tags = analyse_tags(request, tags) except ResponseInstead as e: return e.response - if is_crawler(request) and len(tags) > 1: - return HttpResponseForbidden('address removed from crawling. check robots.txt') - if list_type == 'gallery' and any(tag.category == 'set' for tag in tags): raise Http404 @@ -231,19 +471,22 @@ def tagged_object_list(request, tags, list_type): if any(tag.category == 'set' for tag in tags): params = {'objects': books} else: + books = books.filter(findable=True) params = { - 'objects': Book.tagged_top_level(tags), + 'objects': Book.tagged_top_level(tags).filter(findable=True), 'fragments': Fragment.objects.filter(book__in=books), 'related_tags': get_top_level_related_tags(tags), } elif list_type == 'gallery': params = {'objects': Picture.tagged.with_all(tags)} elif list_type == 'audiobooks': - audiobooks = Book.objects.filter(media__type__in=('mp3', 'ogg')).distinct() + audiobooks = Book.objects.filter(findable=True, media__type__in=('mp3', 'ogg')).distinct() params = { 'objects': Book.tagged.with_all(tags, audiobooks), 'extra': { - 'daisy': Book.tagged.with_all(tags, audiobooks.filter(media__type='daisy').distinct()), + 'daisy': Book.tagged.with_all( + tags, audiobooks.filter(media__type='daisy').distinct() + ), } } else: @@ -258,9 +501,10 @@ def book_fragments(request, slug, theme_slug): fragments = Fragment.tagged.with_all([theme]).filter( Q(book=book) | Q(book__ancestor=book)) + template_name = 'catalogue/2022/book_fragments.html' return render( request, - 'catalogue/book_fragments.html', + template_name, { 'book': book, 'theme': theme, @@ -269,6 +513,7 @@ def book_fragments(request, slug, theme_slug): }) +@never_cache def book_detail(request, slug): try: book = Book.objects.get(slug=slug) @@ -277,42 +522,33 @@ def book_detail(request, slug): return render( request, - 'catalogue/book_detail.html', + 'catalogue/2022/book_detail.html', { 'book': book, + 'accessible': book.is_accessible_to(request.user), 'book_children': book.children.all().order_by('parent_number', 'sort_key'), 'active_menu_item': 'books', - }) - - -# używane w publicznym interfejsie -def player(request, slug): - book = get_object_or_404(Book, slug=slug) - if not book.has_media('mp3'): - raise Http404 - - audiobooks, projects = book.get_audiobooks() - - return render( - request, - 'catalogue/player.html', - { - 'book': book, - 'audiobook': '', - 'audiobooks': audiobooks, - 'projects': projects, + 'club': Club.objects.first() if book.preview else None, + 'donation_form': DonationStep1Form(), }) def book_text(request, slug): book = get_object_or_404(Book, slug=slug) - if book.preview and not user_is_subscribed(request.user): + if not book.is_accessible_to(request.user): return HttpResponseRedirect(book.get_absolute_url()) if not book.has_html_file(): raise Http404 - return render(request, 'catalogue/book_text.html', {'book': book}) + with book.html_file.open('r') as f: + book_text = f.read() + + return render(request, 'catalogue/book_text.html', { + 'book': book, + 'book_text': book_text, + 'inserts': DynamicTextInsert.get_all(request) + }) # ========= @@ -334,11 +570,13 @@ def import_book(request): exception = pprint.pformat(info[1]) tb = '\n'.join(traceback.format_tb(info[2])) return HttpResponse( - _("An error occurred: %(exception)s\n\n%(tb)s") % {'exception': exception, 'tb': tb}, - mimetype='text/plain') + _("An error occurred: %(exception)s\n\n%(tb)s") % { + 'exception': exception, 'tb': tb + }, + content_type='text/plain' + ) return HttpResponse(_("Book imported successfully")) - else: - return HttpResponse(_("Error importing file: %r") % book_import_form.errors) + return HttpResponse(_("Error importing file: %r") % book_import_form.errors) # info views for API @@ -355,33 +593,34 @@ def tag_info(request, tag_id): return HttpResponse(tag.description) -def embargo_link(request, format_, slug): +@never_cache +def embargo_link(request, key, format_, slug): book = get_object_or_404(Book, slug=slug) if format_ not in Book.formats: raise Http404 + if key != book.preview_key: + raise Http404 media_file = book.get_media(format_) if not book.preview: return HttpResponseRedirect(media_file.url) - if not user_is_subscribed(request.user): - return HttpResponseRedirect(book.get_absolute_url()) return HttpResponse(media_file, content_type=constants.EBOOK_CONTENT_TYPES[format_]) -def download_zip(request, format, slug=None): - if format in Book.ebook_formats: - url = Book.zip_format(format) - elif format in ('mp3', 'ogg') and slug is not None: +def download_zip(request, file_format=None, media_format=None, slug=None): + if file_format: + url = Book.zip_format(file_format) + elif media_format and slug is not None: book = get_object_or_404(Book, slug=slug) - url = book.zip_audiobooks(format) + url = book.zip_audiobooks(media_format) else: raise Http404('No format specified for zip package') - return HttpResponseRedirect(urlquote_plus(settings.MEDIA_URL + url, safe='/?=')) + return HttpResponseRedirect(quote_plus(settings.MEDIA_URL + url, safe='/?=')) class CustomPDFFormView(AjaxableFormView): form_class = forms.CustomPDFForm - title = ugettext_lazy('Download custom PDF') - submit = ugettext_lazy('Download') + title = gettext_lazy('Download custom PDF') + submit = gettext_lazy('Download') template = 'catalogue/custom_pdf_form.html' honeypot = True @@ -396,7 +635,7 @@ class CustomPDFFormView(AjaxableFormView): def validate_object(self, obj, request): book = obj - if book.preview and not user_is_subscribed(request.user): + if not book.is_accessible_to(request.user): return HttpResponseRedirect(book.get_absolute_url()) return super(CustomPDFFormView, self).validate_object(obj, request) @@ -408,83 +647,6 @@ class CustomPDFFormView(AjaxableFormView): return obj.pretty_title() -#### -# Includes -#### - - -@ssi_included -def book_mini(request, pk, with_link=True): - # book = get_object_or_404(Book, pk=pk) - try: - book = Book.objects.only('cover_thumb', 'title', 'language', 'slug').get(pk=pk) - except Book.DoesNotExist: - raise Http404 - return render(request, 'catalogue/book_mini_box.html', { - 'book': book, - 'no_link': not with_link, - }) - - -@ssi_included(get_ssi_vars=lambda pk: (lambda ipk: ( - ('ssify.get_csrf_token',), - ('social_tags.likes_book', (ipk,)), - ('social_tags.book_shelf_tags', (ipk,)), - ))(ssi_expect(pk, int))) -def book_short(request, pk): - book = get_object_or_404(Book, pk=pk) - - return render(request, 'catalogue/book_short.html', { - 'book': book, - }) - - -@ssi_included( - get_ssi_vars=lambda pk: book_short.get_ssi_vars(pk) + - (lambda ipk: ( - ('social_tags.choose_cite', [ipk]), - ('catalogue_tags.choose_fragment', [ipk], { - 'unless': Var('social_tags.choose_cite', [ipk])}), - ))(ssi_expect(pk, int))) -def book_wide(request, pk): - book = get_object_or_404(Book, pk=pk) - - return render(request, 'catalogue/book_wide.html', { - 'book': book, - }) - - -@ssi_included -def fragment_short(request, pk): - fragment = get_object_or_404(Fragment, pk=pk) - return render(request, 'catalogue/fragment_short.html', {'fragment': fragment}) - - -@ssi_included -def fragment_promo(request, pk): - fragment = get_object_or_404(Fragment, pk=pk) - return render(request, 'catalogue/fragment_promo.html', {'fragment': fragment}) - - -@ssi_included -def tag_box(request, pk): - tag = get_object_or_404(Tag, pk=pk) - assert tag.category != 'set' - - return render(request, 'catalogue/tag_box.html', { - 'tag': tag, - }) - - -@ssi_included -def collection_box(request, pk): - collection = get_object_or_404(Collection, pk=pk) - - return render(request, 'catalogue/collection_box.html', { - 'collection': collection, - }) - - def tag_catalogue(request, category): if category == 'theme': tags = Tag.objects.usage_for_model( @@ -499,7 +661,8 @@ def tag_catalogue(request, category): else: best = described_tags - return render(request, 'catalogue/tag_catalogue.html', { + template_name = 'catalogue/2022/tag_catalogue.html' + return render(request, template_name, { 'tags': tags, 'best': best, 'title': constants.CATEGORIES_NAME_PLURAL[category], @@ -509,16 +672,18 @@ def tag_catalogue(request, category): def collections(request): - objects = Collection.objects.all() + objects = Collection.objects.filter(listed=True) if len(objects) > 3: - best = random.sample(objects, 3) + best = random.sample(list(objects), 4) else: best = objects - return render(request, 'catalogue/collections.html', { + template_name = 'catalogue/2022/collections.html' + return render(request, template_name, { 'objects': objects, 'best': best, + 'active_menu_item': 'collections' }) @@ -533,4 +698,4 @@ def ridero_cover(request, slug): def get_isbn(request, book_format, slug): book = Book.objects.get(slug=slug) - return HttpResponse(book.extra_info.get('isbn_%s' % book_format)) + return HttpResponse(book.get_extra_info_json().get('isbn_%s' % book_format))