Start a job that uploads a CSV file to a glossary. Check the job's status
with the returned job's self link.

The file uses the format the glossary download produces: term and pos
columns are required; notes, is_case_sensitive, translation_<lang>
and notes_<lang> columns are optional; a notes_<lang> column needs the
translation_<lang> column of its language. A translation column for a
language the glossary doesn't have yet adds that language; a column whose
language code isn't a known language is ignored.

By default the upload is a merge: it adds and updates terms and
translations and removes nothing. With replace set to true it also
removes the terms the file leaves out and, in each language the file has a
translation column for, the translations the file leaves out. The whole
file is saved or nothing is.

Only organization administrators can upload glossary files, and only on a
plan that supports glossary upload.

A glossary that is archived, has no resource, or whose resource isn't the
fileless kind glossaries are created with, can't take an upload: 409.
A glossary takes one upload at a time: while an earlier upload to it is
pending or processing, another one gets 409; try again once it has
finished.