Better try and weed out silly tags
Some people seem to not understand the "tags" field in the new media form, and just type a description sentence.
There are a few things we could do to limit the problems caused by that:
-
separate punctuation from words
We have some tags like
foo;
, which we should tokenize intofoo
and;
. Then we can just remove tags that are purely punctuation. -
try lower-casing the tag name
This could help de-duplicate some tags.
A bigger effort might be needed on the UX of this field, but that's for another ticket.
Edited by Matthieu Gautier