bioconvert genbank2gff3 erroneous tRNA parent if not gene feature before tRNA feature

[tuspjo]

Hi,
Thank you for making and maintaining this very useful tool.

Working with genbank output from antiSMASH6.1.1, bioconvert genbank2gff3 fails when a tRNA feature is found as the first feature on the genome. The same genome directly from PGAP does not fail, because the /gene feature is listed before the /tRNA feature. This led me to look at the files which don't fail, and it turns out that the genbank2gff3 displays the wrong parent locus_tag for tRNAs when the order of features is unexpected (like tRNA before gene), without causing errors.
I believe that genbank format has freedom of feature order, so it could be an issue to rely on /gene being found first for the gff3 conversion, not only for tRNA and not only for antiSMASH output.

Version of bioconvert

1.0.0.post0

Command or code used

bioconvert genbank2gff3 input.gbk test.gff3 --force

genbank input

 tRNA            243644..243717
                 /anticodon=(pos:243678..243680,aa:Pro,seq:ggg)
                 /inference="COORDINATES: profile:tRNAscan-SE:2.0.9"
                 /locus_tag="OIE51_01175"
                 /note="Derived by automated computational analysis using
                 gene prediction method: tRNAscan-SE."
                 /product="tRNA-Pro"
 gene            243644..243717
                 /locus_tag="OIE51_01175"

Expected behaviour

CP109073 GenBank tRNA 243644 243717 . + . ID=OIE51_01175.tRNA.1;Parent=OIE51_01175;anticodon=tRNA-Pro
CP109073 GenBank gene 243644 243717 . + . ID=OIE51_01175;locus_tag=OIE51_01175

Your results

CP109073 GenBank tRNA 243644 243717 . + . ID=OIE51_01175.tRNA.1;Parent=OIE51_01170;anticodon=tRNA-Pro
CP109073 GenBank gene 243644 243717 . + . ID=OIE51_01175;locus_tag=OIE51_01175

What you think might have happened

using /gene feature rather than locus_tag or similar, genbank2gff3 picks up the previous (and wrong) locus tag when /gene is not found before other feature.