The incorrect parsing is a known issue already for Japanese and Chinese. (not sure about other languages)
Is this being worked on?
Love using Language Reactor for my Japanese Immersion but without proper parsing, it is heavily handicapped and I have to use other tools to actually learn the language.
Chrome extensions like Rikaikun or Yomichan do it properly, so I have to rely on those sadly for that task.
In the meantime maybe a function to save custom words would be good, where we the users can select what the actual words are so that it makes sense when we save the sentences.