Fix #303 and add regression test

[Koeng101]

Fixes #303

Basically, there were two weird features in the JCVI-Syn3a genome:

     CDS             3207..4007
                     /gene="ksgA"
                     /locus_tag="JCVISYN3A_0004"
                     /inference="EXISTENCE: similar to AA
                     sequence:RefSeq:WP_011166215.1"
                     /codon_start=1
                     /transl_table=4
                     /product="16S rRNA
                     (adenine(1518)-N(6)/adenine(1519)-N(6))-
                     dimethyltransferase"
                     /protein_id="AVX54572.1"
                     /translation="MKAKKYYGQNFISDLNLINKIVDVLDQNKDQLIIEIGPGKGALT
                     KELVKRFDKVVVIEIDKDMVEILKTKFNHSNLEIIQADVLEIDLKQLISKYDYKNISI
                     ISNTPYYITSEILFKTLQISDLLTKAVFMLQKEVALRICSNKNENNYNNLSIACQFYS
                     QRNFEFVVNKKMFYPIPKVDSAIISLTFNDIYKKQVNNDKKFIDFVRLLFNNKRKTIL
                     NNLNNIIQNKNKALEYLNTLNISSNLRPEQLDIDQYIKLFNLIYNSNF"

The product line (adenine(1518)-N(6)/adenine(1519)-N(6))- messed up because it had an internal /, which we were not checking for. Additionally, there are genes like:

CDS             complement(44397..45011)
                     /locus_tag="JCVISYN3A_0051"
                     /inference="EXISTENCE: similar to AA
                     sequence:RefSeq:WP_017698191.1"
                     /pseudo
                     /codon_start=1
                     /transl_table=4

that messed up because we assumed all qualifiers would have a = associated with them. Turns out, assumption false!

[TimothyStiles]

@Koeng101 can you make the test file smaller?

[Koeng101]

@TimothyStiles fixed.

[Koeng101]

Turns out this is also needed to read https://www.ncbi.nlm.nih.gov/nuccore/NC_001133.9