Provide a generic pattern for including additional user data alongside curated data

[genehack]

Context

From the Priorities meeting "maybe work on this" section, drafted by (I believe) @trvrb :

Also relevant early step if we are doing more as “workflows-as-programs” and I think would be generally super helpful
Augur merge should make this possible, though maybe we need sequence support first?

Useful summary in the augur repo

[jameshadfield]

Existing implementations (please edit the message to fill in missing parts):

ncov

(the original implementation, as far as I remember). If multiple origins are specified in the config then we combine metadata (via this script) and sequences (script). The merging behaviour is (I believe) maintained in augur merge. This functionality is mentioned in the docs although there used to be an entire page focusing on just this concept. The workflow is somewhat complicated due to the nature and history of ncov.

mpox

Additional metadata/sequences are merged via this script, keeping the same merging behaviour as ncov. Triggered if the config defines (a path to) private metadata and sequences.

zika

Additional USVI metadata/sequences are merged via this rule, where sequences are cat'd together and metadata are merged using csvtk but could probably be merged via augur merge.

metadata curation?

I presume the remit of this issue (the generic pattern) does not touch on metadata curation, but that will be an ongoing issue once data merging becomes possible. For context see nextstrain/augur#1475 or the curation we do for the INRB mpox builds.

[jameshadfield]

I implemented this functionality for avian flu recently, although using cat while we wait for sequence support in augur merge. I think the interface in that work is nice - it allows swapping out of the canonical inputs and/or additional inputs.