The current GMOD method using Stag is quite slow - an hour or two for GO and FYPO. (Loading the cvtermpath table takes another 10-ish hours after that).
If no one has written a new loader, maybe we should.
Notes to self:
try reading all the current cvterm table contents into memory about then read all the OBO files into memory - we have lots of memory, let's use it.
needs to update cvterm_relationship too
don't use DBIx::Class
use the COPY FROM support in DBD::Pg to make loading fast: http://search.cpan.org/~turnstep/DBD-Pg/Pg.pm#pg_putcopydata
We should load the replaced_by tags from the OBO files for this ticket:
https://sourceforge.net/p/pombase/chado/271/