add gene annotation to midas_merge_snps

[zhaoc1]

In this PR, we added the gene annotation of representative genome for given {species_id} to the midas_merge_snps workflow. To be specific, four columns were added to the one of the output file: {species_id}.snps_info.tsv:

• locus_type: IGR, CDS, tRNA and etc
• `gene_id: gene id generated by Prokka if locus_type is CDS
• site_type: degeneracy
• amino_acids: amino acids encoded by 4 possible alleles

We also added a subcommand build_gene_features to convert the Prokka GFF format into genes features desired by MIDAS . For each genome, the generated output is saved in s3://microbiome-igg/2.0/genes_annotations/{species_id}/{genome_id}.genes}. The build_gene_featurescommand can be merge intoannotate_genes` issue 45.

For a given genomic position, in order to locate the gene carrying that position, we first converted the list of gene ranges into a list of half-open gene boundaries, and used binary search to find the range that given position fell into. If return even index, then the given position fell between two genes, return "IGR"; if return odd, then the given position fell within one gene, return corresponding gene_type as the locus_type.

Passed the unit test.

python -m iggtools midas_merge_snps --samples_list samples_list.tsv --num_cores 4 merged_midas_output