Make it possible to filter for unused media elements

[timobrembeck]

Motivation

In the media library, it should be possible to filter for all media elements which are not embedded in any pages/events/pois and are not used as icon of any object.

Proposed Solution

Evaluate possible approaches for this task, e.g. whether we should store this information somewhere in the database and update it every time a media element is inserted into a page, or whether this should be "calculated" at runtime by scanning the whole content for occurrences of the media file's absolute paths.

Alternatives

Additional Context

[ulliholtgrave]

I agree that this should be come in handy, but as for the migration we can focus on only migrating used media files. Therefore, I would leave it on the Version 2.0 milestone.

[JanErikHuehne]

Is there any consent on how this should be implemented? Scanning at runtime sounds very slow to me. So should we only perform this one time scan during migration and add fields to media files that get set if they are added / removed somewhere? Maybe a decorator implementation to all functions involving media files might be a good idea to check it? Otherwise, there will a lot of redundant code I assume.

[timobrembeck]

Actually, the information is already stored in the database implicitly by our broken link checker.
So all we would have to do is count how many Link objects exist with the URL to the media element, e.g.

$ pipenv run integreat-cms-cli shell --settings=integreat_cms.core.docker_settings
Python 3.9.15 (main, Oct 26 2022, 16:20:53)
Type 'copyright', 'credits' or 'license' for more information
IPython 8.5.0 -- An enhanced Interactive Python. Type '?' for help.

In [1]: from integreat_cms.cms.models import *

In [2]: from linkcheck.models import *

In [3]: m = MediaFile.objects.first()

In [4]: m.url
Out[4]: 'http://localhost:8000/media/regions/1/2022/11/php.png'

In [5]: url = Url.objects.get(url=m.url)

In [6]: url
Out[6]: <Url: http://localhost:8000/media/regions/1/2022/11/php.png>

In [7]: {link.content_object for link in url.links.all()}
Out[7]: {<PageTranslation (id: 731, page_id: 21, language: de, slug: alltag)>}

This gets all usages in the content, and then we probably want to additionally count the usages as icon, but this is even simpler:

In [8]: m.icon_organizations.all()
Out[8]: <QuerySet []>

In [9]: m.icon_regions.all()
Out[9]: <QuerySet []>

In [10]: m.events.all()
Out[10]: <EventQuerySet []>

In [11]: m.pages.all()
Out[11]: <PageQuerySet []>

In [12]: m.events.all()
Out[12]: <EventQuerySet [<Event (id: 1, region: augsburg, slug: test-veranstaltung)>]>

In [13]: m.pois.all()
Out[13]: <ContentQuerySet []>

And since it's so easy, it probably also makes sense to directly show the first 1-2 objects where it's used in the sidebar of the media library, in case users want to directly edit the page where the image is used.

[JanErikHuehne]

I see. I´m on it. Thanks for the input 👍