fix: cds IDs to be unique

[xvazquezc]

This commit fixes an issue with the gff writer where all CDS sequences from the same gene would have the same ID, e.g.:

##gff-version 3
chr1    Helixer gene    4681    5263    .       -       .       ID=a5d42bed_chr1_001101;
chr1    Helixer mRNA    4681    5263    .       -       .       ID=a5d42bed_chr1_001101.1;Parent=a5d42bed_chr1_001101;product=hypothetical protein;
chr1    Helixer five_prime_UTR  4681    4691    .       -       .       ID=a5d42bed_chr1_001101.1.utr5p1;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer exon    5063    5263    .       -       .       ID=a5d42bed_chr1_001101.1.exon1;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer exon    4917    5012    .       -       .       ID=a5d42bed_chr1_001101.1.exon2;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer exon    4777    4860    .       -       .       ID=a5d42bed_chr1_001101.1.exon3;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer exon    4681    4726    .       -       .       ID=a5d42bed_chr1_001101.1.exon4;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer three_prime_UTR 5241    5263    .       -       .       ID=a5d42bed_chr1_001101.1.utr3p1;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer CDS     5063    5240    .       -       0       ID=a5d42bed_chr1_001101.1.cds;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer CDS     4917    5012    .       -       2       ID=a5d42bed_chr1_001101.1.cds;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer CDS     4777    4860    .       -       2       ID=a5d42bed_chr1_001101.1.cds;Parent=a5d42bed_chr1_001101.1;
chr1    Helixer CDS     4692    4726    .       -       2       ID=a5d42bed_chr1_001101.1.cds;Parent=a5d42bed_chr1_001101.1;

I noticed this when taking a look at the HelixerLite output used as external predictions for funannotate2 predict.

I have no idea if this constitutes an issue for funannotate2 predict itself but tools using strict GFF3 rules are likely to fail or misbehave when parsing CDS information due to this.