Removed fish from GO terms projections. Also double-checked order of execution code and added some explanatory comments.

ENSCORESW-160. Committed with test cases checked in. Also manually verified the code works with transcripts (live as of v67 of Ensembl) ENST00000330803, ENST00000366472 and ENST00000318244

ENSCORESW-171. Bringing in support for splice site work. We also had to change the intron table SQL as it was out of date.

Patching issue with there being a difference between NULL and and empty string WRT description and also plugged a gap in the code when we were trying to query using a display_label which the api allows to be NULL but the DB does not

ENSCORESW-165. Committing the second attempt at the IntronSupportingEvidence API for genebuild RNASeq data.

Patches suggested by Wormbase to deal with scenarios where NULL != 0 and this was our aim. Patch works better now

Auto-increment IDs means we cannot guarentee/reset the DB every time so assert that we were given an id

Execution order now defined in an array, and species pairs now defined through a hash.

If it was already a test db we were dumping then the user and test_db prefix got added onto the species name.