FIX: update `get-ncbi-genomes` to use genome accession IDs as taxonomy IDs

[misialq]

This PR fixes the issue where the taxonomies generated by the get-ncbi-genomes action would contain IDs that do not match the IDs used in the corresponding FASTA file to represent genomic sequences.

• Previously, genome assembly IDs (GCF) were used in the taxonomy but genome accession numbers (e.g., NC) were used in sequence headers obtained from NCBI.
• The changes introduced here will allow finding the accession ID in the sequence metadata obtained with the sequences and use this one instead of the assembly ID.
• An additional challenge which is solved by this approach is the potential existence of multiple chromosomes in a single assembly (hence, single ID would be have been used to represent multiple sequences): in case of bacterial, the typical scenario is existence of additional plasmid molecules; in eukaryotes it may be gDNA molecule comprising multiple chromosomes + mitochondrial/chloroplast DNA
• Additionally, a new flag is being introduced: --p-only-genomic which will allow users to only keep the genomic DNA vs. keeping all the molecules.

[misialq]

Hey @nbokulich, thanks for review! I removed the README file - the zip one had already been removed (you meant the one in the tests/data dir, right?)

[nbokulich]

Hey @misialq I did a bit of user testing with my favorite taxa and all LGTM, thanks!

Yes you are right, I was not reading carefully enough. I now see that this PR removes the unwanted file.

[misialq]

Hey @nbokulich, thanks a lot for testing! In that case, I'm merging this with great excitement 😆