GitLab

Uses the dc:subject, ‘Tag’ now uses ctag:Tag

and start encoding for themes

BTW on VJ14, we were getting a result:

{'date': u'2013-12-14T00:00:00',
  'href': u'http://video.constantvzw.org/VJ14/videoarchive/wolke/Wolke-JuliaRone.ogv',
  'license': u'Free Art Licence'}

Still haven’t figured out how that got encoded…

Long term: maybe have some meta info. that determine if a page can be sniffed?

Python 3 seems to be worth it just because of these errors…

Using the natural_join template tag,

    Authors: Royer, Élodie
    Authors: Gourmel, Yoann

Renders as:

    Yoann Gourmel & Élodie Royer

(sorting in django templates: http://blog.ionelmc.ro/2013/10/26/django-dictsort-template-tag/ )

Would be more classy if this was in the SPARQL directly

FIXME to change so that it works in dev and on the server

`python manage.py index` scrapes the site for RDF/A and than generate an index from it

What’s nice is that this part doesn’t need a database—
we create the graph each time from scratch.

We then convert the RDF/A to a form the template can handle, store it in a json,
and the index view uses this.