Use ngettext where plurals are shown #1668

This looks like a fundamental change in how we do translations, with no prior discussion. Please bring me up to speed.

Gettext is a family of functions, of which the two main ones are gettext (often aliased to `_`) and ngettext. The latter takes an English singular, an English plural, and a count of something, depending on which a plural is selected. Every PO file specifies the formula for selecting a plural depending on the count. During message extraction, source lines marked with ngettext rather than gettext automatically receive [multiple plural form](https://www.gnu.org/software/gettext/manual/html_node/Entries-with-Plural-Forms.html] inputs for each language according to its rules in Poedit, Weblate or another editor. When a plural isn't found, the corresponding English string is shown. This way, the programmer doesn't have to hardcode the pluralization rules for each language or get weirdly-looking combinations like "1 Users". It's documented in [Flask-Babel hello world](https://python-babel.github.io/flask-babel/index.html#using-translations), so I didn't consider it a fundamental change. However, you're right, I should have brought this up in Zulip or an issue. I'd appreciate if someone with more Python experience then myself reviewed if I connected the plurals correctly.