phylo: Replace `set_final_strain_name.py`

[joverlee521]

Context

Part of nextstrain/public#5, recently discussed in #94

@jameshadfield commented:

If "accession" is in the metadata TSV (as it currently is) then augur export v2 will always export it. Additionally it's being used as the node "name" as the current workflow (and parent PR, and current HEAD) use accession as the ID / newick-names.

PPX data has a well-formed strain name, so I'd consider using "strain" as the ID throughout both ingest & phylo, and thus having "PPX_accession" and "INSDC_accession" but no "accession" key. There may be reasons to not want to do this that I am oblivious to.

If we did use a display-default of "tip_label": "strain" then we would still want to delete the automatically-added "accession" somehow, although it wouldn't be a deal-breaker.

[joverlee521]

PPX data has a well-formed strain name, so I'd consider using "strain" as the ID throughout both ingest & phylo, and thus having "PPX_accession" and "INSDC_accession" but no "accession" key. There may be reasons to not want to do this that I am oblivious to.

The strain name includes the PPX versioned accession, so it can change and cannot be considered a stable id. Ingest requires stable id to link manual annotations and phylo requires a stable id for excluding/including specific strains, so I think we still need to stick with the non-versioned accession as the ID.

[jameshadfield]

That's really clarifying actually - as part of #94 I tried out using PPX versioned accession as the ID for a while but thought there must be some reason we use non-versioned accessions. And there is!