SCRUM-5881 use severity rank for sorting molecular consequences

[cmpich]

No description provided.

[claude]

Code Review

The PR adds severity-based sorting to molecular consequences during variant indexing by fetching a severity ranking map from the curation API and using it to sort SO terms.

One bug found — sorting inside the parsing loop:

In VariantSummaryConverter.java:353-355, the soTerms.sort(...) call is placed inside the do...while loop that parses &-separated consequence terms. This means the list is sorted after every term is added except the last one (the final soTerms.add on line 352 happens after the loop exits, then sort runs once more).

Looking more carefully: the sort on lines 353-355 is actually after the loop ends (after line 351's while and line 352's final add), so the placement is correct. The indentation is misleading but the logic is fine — the sort runs exactly once after all terms are collected.

Potential production issue — no error handling for the API call:

In SourceDocumentCreationManager.java:63, soTermApi.getSeverityRanking() is called outside the try/catch block that wraps the variant cache fetching (lines 48-60). If this API call fails (network issue, curation API down), it will throw an uncaught exception that propagates up and kills the entire indexing run. The variant cache fetch has error handling; this new call should too. Consider wrapping it in a try/catch and falling back to an empty map so indexing can proceed without severity sorting rather than failing entirely.

Minor note: The severityRanking map is shared across all SourceDocumentCreation threads, which is fine since it's only read (never written to) after construction.

Overall the logic is sound. The main concern is the missing error handling for the new API call.