fix(extract): emit Java annotation type references

[oleksii-tumanov]

Summary

• emit references edges for class literals inside Java type and method annotation arguments
• emit return-type references for Java annotation members, including array component types
• cover the reciprocal @Repeatable relationship with cross-file resolution to source-backed nodes

Fixes #2426

Result

Label-normalized edges from the issue reproducer:

[
  {
    "source": "RubricFor",
    "target": "RubricsFor",
    "relation": "references",
    "context": "attribute",
    "confidence": "EXTRACTED"
  },
  {
    "source": "RubricsFor",
    "target": "RubricFor",
    "relation": "references",
    "context": "return_type",
    "confidence": "EXTRACTED"
  }
]

Testing

• pytest tests/test_languages.py tests/test_java_type_resolution.py tests/test_java_member_calls.py -q --tb=short (341 passed, 13 skipped)
• pytest tests/ -q --tb=short (3902 passed, 36 skipped; 7 DNS-resolution-only failures in tests/test_security.py)
• pytest tests/test_security.py -q --tb=short with normal DNS resolution (60 passed)
• python -m tools.skillgen --check (134 artifacts matched)
• ruff check graphify/extractors/engine.py tests/test_languages.py (passed)

[graphify-labs]

Looks safe to merge — no coupling regressions and no blocking issues, checked against the code graph (not a self-assessment).

Graphify reviewed this change.

Looks safe to merge — no coupling regressions and no blocking issues, checked against the code graph (not a self-assessment).

---

Graphify review — findings

This PR extends the Java extraction logic in graphify/extractors/engine.py to capture type references from annotation arguments. It adds a helper (_java_annotation_nodes) to enumerate annotation nodes and refactors the existing _java_annotation_names to use it, then introduces _java_annotation_class_literal_refs to collect type names used in class-literal annotation arguments (e.g. @UsesRules(Foo.class)). New extraction branches wire these class-literal references into references/attribute edges for classes and functions, and add handling for annotation_type_element_declaration return-type references. The test file adds two new tests covering class-literal annotation arguments and repeatable-annotation container/element type references across multiple files. I've described the intent and surface area only, without evaluating correctness.

No blocking issues surfaced.

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 808 functions depend on the 608 functions this change touches.

Health — grade A; 10 existing hotspot(s) in the area this change touches (pre-existing, not introduced here):

• _extract_generic() — 18 callers, 15 callees (high)
• extract_xaml() — 19 callers, 12 callees (high)
• extract_objc() — 27 callers, 5 callees (high)
• extract_julia() — 16 callers, 4 callees (high)
• extract_vue() — 10 callers, 6 callees (high)
• extract_groovy() — 14 callers, 3 callees (high)
• extract_astro() — 6 callers, 5 callees (medium)
• _js_extra_walk() — 2 callers, 6 callees (medium)
• …and 2 more

Verification — 808 functions in the blast radius were not formally verified this run (proofs are advisory here).

Gate & verification

graphify gate

PASS — objectively clean (no health regressions, tests not run — proofs not run this pass (advisory)). Grounded, not self-assessed.

Advisory (not blocking):

• verification_scope: 766 function(s) in the blast radius were not formally verified this run