Currently, the post.can_translate attribute is determined in the post serializer itself, by checking if the detected_language of the post is translatable by the selected translator (e.g. Google, Microsoft)

post.can_translate determines if the translate globe is visible on a post.

This PR moves the check into the translator itself using the language_supported? method, and the translator can determine if a language is translatable rather than doing the check outside the translator. This is needed because Discourse AI does not have the limitation of translatable languages, and the return result of language_supported? will always be true, putting the check inside the post serializer has limitations since all translators will be required to define a SUPPORTED_LANG_MAPPING when there doesn't need to be one always.