Merge pull request #239 from nephila/feature/instant_articles_improved

Improve instant articles support

README.rst

@@ -455,6 +455,10 @@ Global Settings
 * BLOG_TAGS_PLUGIN_NAME: Blog tags plugin name (default: ``Tags``)
 * BLOG_CATEGORY_PLUGIN_NAME: Blog categories plugin name (default: ``Categories``)
 * BLOG_ARCHIVE_PLUGIN_NAME: Blog archive plugin name (default: ``Archive``)
+* BLOG_FEED_CACHE_TIMEOUT: Cache timeout for RSS feeds
+* BLOG_FEED_INSTANT_ITEMS: Number of items in Instant Article feed
+* BLOG_FEED_LATEST_ITEMS: Number of items in latest items feed
+* BLOG_FEED_TAGS_ITEMS: Number of items in per tags feed
 
 Read-only settings
 ++++++++++++++++++

djangocms_blog/cms_menus.py

@@ -3,7 +3,6 @@ from __future__ import absolute_import, print_function, unicode_literals
 
 from cms.apphook_pool import apphook_pool
 from cms.menu_bases import CMSAttachMenu
-from django.contrib.sites.shortcuts import get_current_site
 from django.core.urlresolvers import resolve
 from django.db.models.signals import post_delete, post_save
 from django.utils.translation import get_language_from_request, ugettext_lazy as _
@@ -14,6 +13,11 @@ from .cms_appconfig import BlogConfig
 from .models import BlogCategory, Post
 from .settings import MENU_TYPE_CATEGORIES, MENU_TYPE_COMPLETE, MENU_TYPE_POSTS, get_setting
 
+try:
+    from django.contrib.sites.shortcuts import get_current_site
+except ImportError:
+    from django.contrib.sites.models import get_current_site
+
 
 class BlogCategoryMenu(CMSAttachMenu):
     """

djangocms_blog/feeds.py

@@ -7,18 +7,31 @@ from django.contrib.syndication.views import Feed
 from django.core.cache import cache
 from django.core.urlresolvers import reverse
 from django.utils.encoding import force_text
-from django.utils.feedgenerator import Rss201rev2Feed, rfc2822_date
+from django.utils.feedgenerator import Rss201rev2Feed
+from django.utils.html import strip_tags
 from django.utils.safestring import mark_safe
+from django.utils.six import BytesIO
 from django.utils.translation import get_language_from_request, ugettext as _
+from lxml import etree
 
 from djangocms_blog.settings import get_setting
 from djangocms_blog.views import PostDetailView
 
 from .models import Post
 
+try:
+    import HTMLParser
+
+    h = HTMLParser.HTMLParser()
+except ImportError:
+    from html.parser import HTMLParser
+
+    h = HTMLParser()
+
 
 class LatestEntriesFeed(Feed):
     feed_type = Rss201rev2Feed
+    feed_items_number = get_setting('FEED_LATEST_ITEMS')
 
     def __call__(self, request, *args, **kwargs):
         self.request = request
@@ -29,27 +42,48 @@ class LatestEntriesFeed(Feed):
         return reverse('%s:posts-latest' % self.namespace, current_app=self.namespace)
 
     def title(self):
+        return Site.objects.get_current().name
+
+    def description(self):
         return _('Blog articles on %(site_name)s') % {'site_name': Site.objects.get_current().name}
 
     def items(self, obj=None):
-        return Post.objects.namespace(self.namespace).published().order_by('-date_published')[:10]
+        return Post.objects.namespace(
+            self.namespace
+        ).published().order_by('-date_published')[:self.feed_items_number]
 
     def item_title(self, item):
-        return item.safe_translation_getter('title')
+        return mark_safe(item.safe_translation_getter('title'))
 
     def item_description(self, item):
         if item.app_config.use_abstract:
-            return item.safe_translation_getter('abstract')
+            return mark_safe(item.safe_translation_getter('abstract'))
-        return item.safe_translation_getter('post_text')
+        return mark_safe(item.safe_translation_getter('post_text'))
+
+    def item_updateddate(self, item):
+        return item.date_modified
+
+    def item_pubdate(self, item):
+        return item.date_published
+
+    def item_guid(self, item):
+        return item.guid
+
+    def item_author_name(self, item):
+        return item.get_author_name()
+
+    def item_author_url(self, item):
+        return item.get_author_url()
 
 
 class TagFeed(LatestEntriesFeed):
+    feed_items_number = get_setting('FEED_TAGS_ITEMS')
 
     def get_object(self, request, tag):
         return tag  # pragma: no cover
 
     def items(self, obj=None):
-        return Post.objects.published().filter(tags__slug=obj)[:10]
+        return Post.objects.published().filter(tags__slug=obj)[:self.feed_items_number]
 
 
 class FBInstantFeed(Rss201rev2Feed):
@@ -61,31 +95,51 @@ class FBInstantFeed(Rss201rev2Feed):
         }
 
     def add_root_elements(self, handler):
-        handler.addQuickElement("title", self.feed['title'])
+        handler.addQuickElement('title', self.feed['title'])
-        handler.addQuickElement("link", self.feed['link'])
+        handler.addQuickElement('link', self.feed['link'])
-        handler.addQuickElement("description", self.feed['description'])
+        handler.addQuickElement('description', self.feed['description'])
         if self.feed['language'] is not None:
-            handler.addQuickElement("language", self.feed['language'])
+            handler.addQuickElement('language', self.feed['language'])
         for cat in self.feed['categories']:
-            handler.addQuickElement("category", cat)
+            handler.addQuickElement('category', cat)
         if self.feed['feed_copyright'] is not None:
-            handler.addQuickElement("copyright", self.feed['feed_copyright'])
+            handler.addQuickElement('copyright', self.feed['feed_copyright'])
-        handler.addQuickElement("lastBuildDate", rfc2822_date(self.latest_post_date()))
+        handler.addQuickElement('lastBuildDate', self.latest_post_date().isoformat())
         if self.feed['ttl'] is not None:
-            handler.addQuickElement("ttl", self.feed['ttl'])
+            handler.addQuickElement('ttl', self.feed['ttl'])
 
     def add_item_elements(self, handler, item):
         super(FBInstantFeed, self).add_item_elements(handler, item)
+        if item['author']:
+            handler.addQuickElement('author', item['author'])
+        if item['date_mod'] is not None:
+            handler.addQuickElement('pubDate', item['date'].isoformat())
+        if item['date_pub'] is not None:
+            handler.addQuickElement('modDate', item['date'].isoformat())
+        handler.startElement('description', {})
+        handler._write('<![CDATA[{0}]]>'.format(h.unescape(force_text(item['abstract']))))
+        handler.endElement('description')
         handler.startElement('content:encoded', {})
         handler._write('<![CDATA[')
-        handler._write(force_text(item['content']))
+        handler._write('<!doctype html>')
+        handler._write(h.unescape(force_text(item['content'])))
         handler._write(']]>')
         handler.endElement('content:encoded')
-        handler.addQuickElement('guid', item['guid'])
 
 
 class FBInstantArticles(LatestEntriesFeed):
     feed_type = FBInstantFeed
+    feed_items_number = get_setting('FEED_INSTANT_ITEMS')
+
+    def _clean_html(self, content):
+        body = BytesIO(content)
+        document = etree.iterparse(body, html=True)
+        for a, e in document:
+            if not (e.text and e.text.strip()) and len(e) == 0 and e.tag == 'p':
+                e.getparent().remove(e)
+            if e.tag in ('h3', 'h4', 'h5', 'h6') and 'op-kicker' not in e.attrib.get('class', ''):
+                e.tag = 'h2'
+        return etree.tostring(document.root)
 
     def item_extra_kwargs(self, item):
         if not item:
@@ -97,10 +151,32 @@ class FBInstantArticles(LatestEntriesFeed):
             view = PostDetailView.as_view(instant_article=True)
             response = view(self.request, slug=item.safe_translation_getter('slug'))
             response.render()
-            content = mark_safe(response.content)
+            content = self._clean_html(response.content)
             cache.set(key, content, timeout=get_setting('FEED_CACHE_TIMEOUT'))
+        if item.app_config.use_abstract:
+            abstract = strip_tags(item.safe_translation_getter('abstract'))
+        else:
+            abstract = strip_tags(item.safe_translation_getter('post_text'))
         return {
+            'author': item.get_author_name(),
             'content': content,
-            'slug': item.safe_translation_getter('slug'),
+            'date': item.date_modified,
-            'guid': item.guid,
+            'date_pub': item.date_modified,
+            'date_mod': item.date_modified,
+            'abstract': abstract
         }
+
+    def item_categories(self, item):
+        return [category.safe_translation_getter('name') for category in item.categories.all()]
+
+    def item_author_name(self, item):
+        return ''
+
+    def item_author_url(self, item):
+        return ''
+
+    def item_description(self, item):
+        return None
+
+    def item_pubdate(self, item):
+        return None

djangocms_blog/models.py

@@ -205,7 +205,7 @@ class Post(KnockerModel, ModelMeta, TranslatableModel):
     def guid(self, language=None):
         if not language:
             language = self.get_current_language()
-        base_string = '{0}-{1}-{2}'.format(
+        base_string = '{0}{2}{1}'.format(
             language, self.app_config.namespace,
             self.safe_translation_getter('slug', language_code=language, any_language=True)
         )

djangocms_blog/settings.py

@@ -123,6 +123,12 @@ def get_setting(name):
             settings, 'BLOG_ARCHIVE_PLUGIN_NAME', _('Archive')),
         'BLOG_FEED_CACHE_TIMEOUT': getattr(
             settings, 'BLOG_FEED_CACHE_TIMEOUT', 3600),
+        'BLOG_FEED_INSTANT_ITEMS': getattr(
+            settings, 'BLOG_FEED_INSTANT_ITEMS', 50),
+        'BLOG_FEED_LATEST_ITEMS': getattr(
+            settings, 'BLOG_FEED_LATEST_ITEMS', 10),
+        'BLOG_FEED_TAGS_ITEMS': getattr(
+            settings, 'BLOG_FEED_TAGS_ITEMS', 10),
 
     }
     return default['BLOG_%s' % name]

djangocms_blog/templates/djangocms_blog/post_instant_article.html

@@ -1,6 +1,6 @@
 {% load thumbnail cms_tags %}
 <!doctype html>
-<html lang="en" prefix="op: http://media.facebook.com/op#">
+<html lang="{{ post.get_current_language }}" prefix="op: http://media.facebook.com/op#">
   <head>
       <meta charset="utf-8">
       {% block canonical_url %}<link rel="canonical" href="{{ meta.url }}"/>{% endblock canonical_url %}
@@ -16,19 +16,17 @@
         <time class="op-modified" dateTime="{{ post.date_modified.isoformat }}">{{ post.date_modified|date:"DATE_FORMAT" }}</time>
 
         <address>
-            {% if og_author_url %}<a rel="facebook" href="{{ og_author_url }}">{% endif %}
+            <a {% if og_author_url %}rel="facebook" href="{{ og_author_url }}"{% endif %}>{{ post.get_author_name }}</a>
-            {{ post.author }}
-            {% if og_author_url %}</a>{% endif %}
         </address>
 
         <figure>
-            <img src="{{ meta.image }}" alt="{{ post.main_image.default_alt_text }}" />
+            <img src="{{ meta.image }}" alt="{{ post.main_image.default_alt_text|default:'' }}" />
             {% if post.main_image.default_caption %}
                 <figcaption>{{ post.main_image.default_caption }}</figcaption>{% endif %}
         </figure>
 
         <h3 class="op-kicker">
-          {{ post.abstract|striptags }}
+          {{ post.abstract|striptags|safe }}
         </h3>
 
       </header>

tests/test_views.py

@@ -4,6 +4,7 @@ from __future__ import absolute_import, print_function, unicode_literals
 import os.path
 
 from aldryn_apphooks_config.utils import get_app_instance
+from cms.api import add_plugin
 from cms.toolbar.items import ModalItem
 from django.contrib.auth.models import AnonymousUser
 from django.core.exceptions import ImproperlyConfigured
@@ -315,9 +316,13 @@ class ViewTest(BaseTest):
             self.assertEqual(context['post_list'][0].title, 'First post')
 
     def test_feed(self):
+        self.user.first_name = 'Admin'
+        self.user.last_name = 'User'
+        self.user.save()
         posts = self.get_posts()
         pages = self.get_pages()
         posts[0].tags.add('tag 1', 'tag 2', 'tag 3', 'tag 4')
+        posts[0].author = self.user
         posts[0].save()
         posts[1].tags.add('tag 6', 'tag 2', 'tag 5', 'tag 8')
         posts[1].save()
@@ -335,6 +340,7 @@ class ViewTest(BaseTest):
                 xml = feed(request)
                 self.assertContains(xml, posts[0].get_absolute_url())
                 self.assertContains(xml, 'Blog articles on example.com')
+                self.assertContains(xml, 'Admin User</dc:creator>')
 
         with smart_override('it'):
             with switch_language(posts[0], 'it'):
@@ -352,8 +358,18 @@ class ViewTest(BaseTest):
                 self.assertEqual(list(feed.items('tag-2')), [posts[0]])
 
     def test_instant_articles(self):
+        self.user.first_name = 'Admin'
+        self.user.last_name = 'User'
+        self.user.save()
         posts = self.get_posts()
         pages = self.get_pages()
+        posts[0].tags.add('tag 1', 'tag 2', 'tag 3', 'tag 4')
+        posts[0].categories.add(self.category_1)
+        posts[0].author = self.user
+        posts[0].save()
+        add_plugin(
+            posts[0].content, 'TextPlugin', language='en', body='<h3>Ciao</h3><p></p><p>Ciao</p>'
+        )
 
         with smart_override('en'):
             with switch_language(posts[0], 'en'):
@@ -373,6 +389,9 @@ class ViewTest(BaseTest):
                 self.assertContains(xml, '<link rel="canonical" href="{0}"/>'.format(
                     posts[0].get_full_url()
                 ))
+                # Assert text transformation
+                self.assertContains(xml, '<h2>Ciao</h2><p>Ciao</p>')
+                self.assertContains(xml, '<a>Admin User</a>')
 
     def test_sitemap(self):
         posts = self.get_posts()