fix for numerical overflow during enumerate

when performing the enumerate command the integer used to count isoforms can potentially overflow, resulting in the -max_isoforms argument not being correctly respected (see #10). I've replaced the integer variable with a BigInteger to avoid these overflow issues.

There were also some lingering "null" labeled junctions relating to issue #10, these remaining cases appear to be due to 1nt long introns implied by aligned reads. While a 1nt intron doesn't make much sense (a 1nt deletion is probably more likely), IsoSCM delegates to the aligner to identify intron positions correctly. As such, if 1nt introns are reported by the aligner, they are now handled by IsoSCM incorporated into resulting gene models without null labels.