Text this: Tokenization efficiency of current foundational large language models for the Ukrainian language