Can I access raw fragment match data in generated translation model, either directly through Studio 2021 or through an API? Where would I find instructions? My purpose is to build a domain specific phrase (and word) glossary for an under-resourced language.
