model.base.LangStringSet: change class structure to fit the explanation in DotAASv2.0 better

According to DotAASv2.0, the LangStringSet is "A set of strings, each annotated by the language of the string. The meaning of the string in each language shall be the same."

Right now, its content is just a set of LangStrings.

Wouldn't it make more sense, if its content was a set of pairs of a LangString, together with a string that "annotates the language" (or even an Enum of the different language IDs or sth) instead?

The way I understand the DotAASv2.0, is that LangStrings are nothing but strings (in a specified language) but it doesn't seem like there is a way for anyone to know which language it is written in.