#8 Revised handling of non-canonical splicing to be consistent with database documentation.

[Jeffrey Hoover]

Revised handling of non-canonical splicing to be consistent with database documentation.
Tested with validation scripts and results are actually better now.
1. added subroutine get_ref_splicepairs
2. changed extract_spliced_exons to call get_ref_splicepairs instead of get_gene_splicepairs