ncbi_fetch: Require "gb" identifier

[tyleraland]

Allegedly GI identifiers are going away in June and switching to "accessioned" gb identifiers: https://www.ncbi.nlm.nih.gov/Sitemap/sequenceIDs.html

So ncbi_fetch should require gb's

[crosenth]

FYI Noah and I have a version of bioy ncbi_fetch called deenurp ncbi_extract_genbank that maybe we can combine efforts on: https://github.com/fhcrc/deenurp/blob/master/deenurp/subcommands/ncbi_extract_genbank.py

[tyleraland]

I'll look into using that instead, since ncbi_fetch has a pretty limited scope anyway. Thanks.

[tyleraland]

deenurp has enough dependencies that it seems like a significant hassle to include it in my pipeline. So I'd prefer one of:

• porting the essential features from deenurp to bioy
• solving a limited subset of the problem in bioy
• stealing the essential pieces as a standalone script and using that for my purposes

After some tinkering I got deenurp installed with very few dependencies and enough functionality to run ncbi_extract_genbank. I'll monkey around with it.

[tyleraland]

Seems to work and I like it. I'd vote to just ditch bioy ncbi_fetch entirely and consolidate efforts into deenurp.