export and validate: Substantially improve validation errors

[tsibley]

All errors are reported in the case of subschemas inside a "oneOf" validator, such as used by our "tree" property. Previously only the top-level error was reported and it was worse than useless as it was incredibly long and contained no actionable .

Suberrors are grouped by the validator arm they failed, making it clear why each element of a "oneOf" failed.

Errors avoid being overly long by consistently shortening output that might be long. Different shortening strategies are used for different values in an attempt to preserve the most useful information.

Paths in error messages now index into the given JSON being validated, not the JSON Schema. This disconnect was confusing as the schema path refers to properties that don't exist in the JSON being validated.

Resolves #1044.

Testing

• Manually tested various error message situations
• Added doctests
• CI passes

Checklist

• Add a message in CHANGES.md summarizing the changes in this PR that are end user focused. Keep headers and formatting consistent with the rest of the file.

[codecov]

Codecov Report

Base: 68.08% // Head: 68.14% // Increases project coverage by +0.06% 🎉

Coverage data is based on head (59b5505) compared to base (85d2edd).
Patch coverage: 80.00% of modified lines in pull request are covered.

❗ Current head 59b5505 differs from pull request most recent head a8ea3dd. Consider uploading reports for the commit a8ea3dd to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1134      +/-   ##
==========================================
+ Coverage   68.08%   68.14%   +0.06%     
==========================================
  Files          62       62              
  Lines        6690     6725      +35     
  Branches     1641     1646       +5     
==========================================
+ Hits         4555     4583      +28     
- Misses       1829     1836       +7     
  Partials      306      306              

Impacted Files	Coverage Δ
augur/validate.py	72.67% <75.00%> (+0.24%)	⬆️
augur/__version__.py	100.00% <100.00%> (ø)
augur/io/json.py	86.76% <100.00%> (+1.76%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[rneher]

thanks, @trs! this is much more helpful:

      .tree.children[…].node_attrs.accession "England/Qatar/2012" failed pattern validation for "^[0-9A-Za-z-_.]+$"
      .tree.children[…].node_attrs.accession "D998/15" failed pattern validation for "^[0-9A-Za-z-_.]+$"
      .tree.children[…].node_attrs.accession "Jordan-N3/2012" failed pattern validation for "^[0-9A-Za-z-_.]+$"
    validation for arm 1: {"type": "array", "minItems": 1, "items": {"$ref": "#/$defs/tree"}}
      .tree {"name": "NODE_0000000", "node_attrs": {"div": 0…} failed type validation for "array"
Validation of 'auspice/mers2023.json' failed.

[emmahodcroft]

Thanks so much for taking this up Tom! I can remember a time when our validation errors were helpful (this is a long time ago), so it's great to get them back to something that one can actually interpret! 🙌

[joverlee521]

Thanks for fixing up all the things I missed when I copied over the json module 🙏