get-gtdb refactor

[mikerobeson]

Fixes #167 and #168.

This refactor:

• File parsing is different.
  • We now pull taxonomy from the FASTA files directly, not the taxonomy files.
• Download Options:
  • GTDB SSU Species Representatives, as we've been doing
    • Both
    • Bacteria
    • Archaea
  • Non-species representative GTDB SSUs
    • Contains all Bacteria and Archaea that have passed GTDBs QA/QC But are not clustered into species representatives.

To do:

• update test code
• break up code into simpler functions where possible.

[nbokulich]

Hey @mikerobeson looks good, thanks! I pulled down and tested and it works like a charm 😉 . I just have some super minor questions/comments on the code inline.

[nbokulich]

theoretical question: should we keep the exact same nomenclature as used by GTDB? Or introduce the decimals for all versions as you have done here? I don't have an opinion on this but just pausing for a moment's thought. Would any users be confused that we support a version 207.0 but this does not exist in GTDB (strictly speaking)?

[mikerobeson]

I was pondering this myself. The issue is that the http paths all contain a folder with a decimal:
https://data.gtdb.ecogenomic.org/releases/releases/release202/202.0/
https://data.gtdb.ecogenomic.org/releases/releases/release207/207.0/
https://data.gtdb.ecogenomic.org/releases/releases/release214/214.0/
https://data.gtdb.ecogenomic.org/releases/releases/release214/214.1/

At the time I figured I'd just use that folder label and then parse out the base version, e.g. 214 from 214.1 rather than split off the decimal and rebuild it for the http string.

All I'd need to do is check for a decimal within the version string e.g. 214.1 and then use that for the folder. Then append a .0 to those versions that do not have a a decimal, e.g. 207, etc...

[nbokulich]

ah makes sense... let's use the decimals then.

[mikerobeson]

The PR should be good to go then. :-)

[nbokulich]

🙌