Add schema def for unaligned sequences

[victorlin]

Description of proposed changes

Some workflows use unaligned sequences as input to augur subsample. In this case, proximal samples will not work properly. This schema def makes it easier for workflows to guard against configs with proximal samples.

Related issue(s)

nextstrain/rsv#103 (comment)

Checklist

• Manually tested in Use augur subsample rsv#103
• Automated checks pass
• Check if you need to add a changelog message
• Check if you need to add tests
• Check if you need to update docs

[joverlee521]

I wasn't sure how allOf works here, so I just tested a couple configs.
Looks like this works as expected if just trying to use a proximalSampleProperties

Validating schema of 'results/run_config.yaml'...
  .subsample.'a/genome/6y'.samples.background failed: Unexpected property 'focal_sample'
ERROR: Validation of 'results/run_config.yaml' failed.

There's an edge case where I mix filterSampleProperties and proximalSampleProperties, then I get an error message that suggests proximalSampleProperties are accepted

Validating schema of 'results/run_config.yaml'...
  .subsample.'a/genome/6y'.samples.background failed: {"group_by": ["year", "country"], "focal_sample"…} did not match one of the acceptable options below.
    Option 1: {"$ref": "#/$defs/filterSampleProperties"}
      .subsample.'a/genome/6y'.samples.background failed: Unexpected property 'focal_sample'
    Option 2: {"$ref": "#/$defs/proximalSampleProperties"}
      .subsample.'a/genome/6y'.samples.background failed: Unexpected property 'group_by'
  .subsample.'a/genome/6y'.samples.background failed: Unexpected property 'focal_sample'
ERROR: Validation of 'results/run_config.yaml' failed.

I'm not sure there's a way to work around this since JSON schema doesn't really support overriding subschemas.

[victorlin]

Currently the schema defaults to allowing proximal samples. We could flip it around and make it opt-in such that nextstrain/rsv#103 references https://nextstrain.org/schemas/augur/subsample-config/v1, while measles references https://nextstrain.org/schemas/augur/subsample-config/v1#/$defs/schemaForAlignedSequences. This would work better with allOf since it's additive.

[joverlee521]

Hmm, based on allOf docs, flipping it around won't work either

allOf can not be used to "extend" a schema to add more details to it in the sense of object-oriented inheritance. Instances must independently be valid against "all of" the schemas in the allOf. See the section on Extending Closed Schemas for more information.

[victorlin]

e19abfa works:

Validating schema of 'results/run_config.yaml'...
  .subsample.'a/genome/all-time'.samples.sample failed: Unexpected property 'focal_sample'
ERROR: Validation of 'results/run_config.yaml' failed.

But validation is also done at augur subsample run time. If proximal samples are opt-in, we'd need to make it explicit using something like --aligned-sequences instead of --sequences...