Code Review

Summary: This PR consistently replaces getTaxon().getName() with getTaxon().getSpecies().getFullName() across 8 files to use species full name in downloads and API responses. The change is mechanical and the intent is clear.

Potential NullPointerException risk (medium severity):

The new call chain getTaxon().getSpecies().getFullName() adds an extra dereference. Existing indexer code in DiseaseAnnotationCurationIndexer.java explicitly null-checks getSpecies() (lines 207 and 415):

if (gene.getTaxon().getSpecies() != null) {
    gdad.setPhylogeneticSortingIndex(gene.getTaxon().getSpecies().getPhylogeneticOrder());
}

This confirms getSpecies() can return null. None of the call sites in this PR guard against that. If a record has a taxon but no associated species, each of these locations will throw an NPE at runtime:

• ExpressionRibbonESService.java:156
• DiseaseAnnotationToTdfTranslator.java:194, 200, 206
• AlleleToTdfTranslator.java:166
• ExpressionToTdfTranslator.java:62
• GeneGeneticInteractionToTdfTranslator.java:59
• GeneMolecularInteractionToTdfTranslator.java:59
• VariantSearchResultConverter.java:35 (already checks getTaxon() != null but not getSpecies())

Suggestion: Add null checks for getSpecies(), or if species is guaranteed non-null for all records reaching these code paths, document that assumption. The VariantSearchResultConverter is the easiest to fix since it already has a null-guard pattern — just extend it:

if (variantLocation.getVariantAssociationSubject().getTaxon() != null
    && variantLocation.getVariantAssociationSubject().getTaxon().getSpecies() != null) {

Everything else (logic, test updates, import reordering) looks correct.