Term-collection language resources import worker should run only one at a time

Problem

The current language resources import worker allows 3 term-collection to run in the same time. This leads to a tbx parsing problem when  additional term collection is re-imported in the already running tbx import.

The language-resources worker should make difference of how many parallel workers are allowed to run by language resource (term collection import should be limited to 1, all other to 3 for example.)

Additional: disable the term-collection re-import when an import is already running.