fix(export): In export schema, allow cds name with nuc prefix, as long as not equal to nuc

[corneliusroemer]

Fixes #1433

Description of proposed changes

Fixes bug #1433 introduced in v23.1.0, that causes validation to fail when gene names start with nuc, e.g. nucleocapsid.

Regex excluded all strings that start with nuc not just nuc. Adding a trailing $ to the lookahead fixes that.

Checklist

• Checks pass
• If making user-facing changes, add a message in CHANGES.md summarizing the changes in this PR

It seems like we don't really test export validation much. I had a look whether I could quickly add a test case for this bug to prevent regression but couldn't see a neat way. I think we should not block this on lacking tests, but rather add an issue to improve testing of validation. Would be nice if there was a way to do more unit tests for export here: tests/test_validate.py - and potentially also test validation in an E2E test.

[jameshadfield]

Can confirm that this fixes the case where keys in the branch_attrs.mutations couldn't start with "nuc" ✅

There's a related bug in the part of the schema which checks the meta.genome_annotations, where we use the same regex but fail to disallow additional properties (which is why it wasn't flagged up in validation). Could you add the following patch to this PR?

diff --git a/augur/data/schema-annotations.json b/augur/data/schema-annotations.json
index 37a9075a..e064a34b 100644
--- a/augur/data/schema-annotations.json
+++ b/augur/data/schema-annotations.json
@@ -24,8 +24,9 @@
         }
     },
     "required": ["nuc"],
+    "additionalProperties": false,
     "patternProperties": {
-        "^(?!nuc)[a-zA-Z0-9*_-]+$": {
+        "^(?!nuc$)[a-zA-Z0-9*_-]+$": {
             "$comment": "Each object here defines a single CDS",
             "type": "object",
             "oneOf": [{ "$ref": "#/$defs/startend" }, { "$ref": "#/$defs/segments" }],

P.S. The reason for the negative look-ahead is that we use a different schema for "nuc" objects vs "CDS" objects.

[corneliusroemer]

Sorry for the delay @jameshadfield - I've added the patch now