This was based on libraries that were never ported to python3. If this
is still needed, it would have to be reimplemented with other newer
technologies.

some conversion (TextDocument) depend on the content type, so it's
necessary to first update the content type before trying to convert to
base format