GitLab contributor name parsing is unstable when encountering strange display names

[zyzzyxdonta]

Our GitLab instance gives users display names that contain a department and a number in addition to their lastname, firstname. They look something like this:

Doe, John (DEP) - 123456

The currently used algorithm for deconstructing this into a first and last name can not really handle this and produces this result:

Given name: Doe, John John John Family name: (DEP) - 123456

I wouldn't expect SMECS to handle this case perfectly but the result I get is a bit bizarre nonetheless.

[sferenz]

Done within 2dade71532929df075c62dfc2365da433a759458