fix(extractors): extract Kotlin annotations and constructor properties (#3835)

[hopstreax]

Summary

Fixes #3835 by restoring Kotlin annotation and property dependency extraction while preserving Graphify's existing class-level JVM graph model.

Changes

• Extract Kotlin declaration annotations from classes, constructor properties, functions, and properties.
• Support Kotlin use-site targets such as @get:Transient.
• Support bracketed annotations such as @get:[Transient VisibleForTesting].
• Extract annotation class literals such as Customer::class.
• Restore field references for primary-constructor val/var properties.
• Restore generic_arg references for generic constructor and class-body properties.
• Handle annotations on property getters/setters.
• Preserve the distinction between constructor properties and plain constructor parameters.
• Avoid introducing Kotlin property/member nodes.

Tests

• pytest tests/test_kotlin_grammar.py tests/test_kotlin_object_literal.py — 42 passed
• pytest tests/test_languages.py -k "kotlin or java" — 46 passed
• git diff --check — clean

Architectural Notes

This implementation intentionally uses Graphify's existing class-level dependency representation:

Order --references(context="field")--> Customer
Order --references(context="generic_arg")--> OrderLine
Order --references(context="attribute")--> ManyToOne
Order --references(context="attribute")--> Customer

It does not introduce Kotlin-specific property/member nodes, avoiding inconsistencies with the existing Java/Kotlin JVM graph model and type-resolution infrastructure.

[github-actions]

Thanks for the pull request, @hopstreax. A maintainer will review it soon.

Want to talk it through while it is in review? Come join us on our Discord server. For longer-form discussion there is also GitHub Discussions.

A couple of things that speed up review: make sure the test suite passes on Python 3.10 and 3.13, and that the change keeps extraction deterministic.

[graphify-labs]

Graphify reviewed this change.

Worth a look — the grounded gate found no coupling regressions or blocking issues, but 1 advisory finding(s) below merit a look before merge.

Formal verification. No changes could be formally verified in this run.

---

Graphify review — findings

Extends Kotlin annotation handling for JPA-style code: _kotlin_annotation_names now also reads annotations on property getters and setters and dedupes repeated names. A new _kotlin_annotation_class_literal_refs adds references edges with context="attribute" for class literals in annotation arguments, such as Customer::class, skipping Kotlin and Java builtins. These edges are emitted for classes, properties, and primary-constructor val/var parameters, and constructor-parameter annotations now produce edges even when the parameter's type node isn't found.

Worth a look

• Class-literal detection likely never matches Foo::class in the Kotlin grammar — graphify/extractors/engine.py:820 · Escalate · medium
  • agreed by 2 of 2 members but NOT verified (no proof, no reproducing execution) — consensus is not a verdict; needs human review

Analysis details — impact, health, verification

Impact & health

Graphify review

Impact — 774 functions depend on the 299 functions this change touches.

Health — this change adds coupling hotspots:

• new: _extract_generic() — 18 callers, 29 callees
• new: extract_js() — 87 callers, 4 callees
• new: extract_xaml() — 19 callers, 17 callees
• new: extract_objc() — 27 callers, 9 callees
• new: extract_julia() — 19 callers, 7 callees
• new: extract_cpp() — 29 callers, 3 callees
• new: extract_vue() — 10 callers, 7 callees
• new: walk() — 1 callers, 65 callees
• …and 9 more — each is listed as a finding

Verification — 774 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: 711 function(s) in the blast radius were not formally verified this run

Test selection

Test selection

26 of 308 test file(s) selected (8%) via static blast radius.

• tests/test_astro_extraction.py — impact
• tests/test_build.py — impact
• tests/test_cjs_module_extension.py — impact
• tests/test_cpp_nested_and_cli.py — impact
• tests/test_dotnet.py — impact
• tests/test_extract.py — impact
• tests/test_extract_php_closures.py — impact
• tests/test_import_extension_resolution.py — impact
• tests/test_indirect_call_block_scoped_shadow.py — impact
• tests/test_indirect_dispatch.py — impact
• tests/test_indirect_dispatch_assign_return.py — impact
• tests/test_indirect_dispatch_getattr.py — impact
• tests/test_js_exported_scalar_bindings.py — impact
• tests/test_kotlin_grammar.py — impact, changed-test
• tests/test_languages.py — impact
• tests/test_multilang.py — impact
• tests/test_python_underscore_resolution.py — impact
• tests/test_rationale.py — impact
• tests/test_ruby_resolution.py — impact
• tests/test_scala_self_type.py — impact
• tests/test_swift_computed_properties.py — impact
• tests/test_swift_protocol_requirements.py — impact
• tests/test_trailing_newline_not_a_syntax_error.py — impact
• tests/test_ts_new_expression_calls.py — impact
• tests/test_typescript_module_extensions.py — impact
• tests/test_vue_extraction.py — impact

Selection is safe under the controlled-regression assumption; always-run tests + a periodic full run are the backstops. Advisory — it never changes the check verdict.

Formal verification

Could not verify: Could not verify \_extract\_generic.

The verifier did not have enough to check \_extract\_generic, so it is saying so rather than guessing. No false assurance is the whole point.

Guarantee: No guarantee either way, this is an honest abstention, not a pass.

Note: Reason: no capturable inputs from the test suite; property tier: parameter `config` is annotated `LanguageConfig` — outside the synthesizable primitive/collection set

Could not verify: Could not verify \_kotlin\_annotation\_names.

The verifier did not have enough to check \_kotlin\_annotation\_names, so it is saying so rather than guessing. No false assurance is the whole point.

Guarantee: No guarantee either way, this is an honest abstention, not a pass.

Note: Reason: no capturable inputs from the test suite; property tier: not verifiable: all 200 sampled inputs raised on both versions — the function never executed, so 'no divergence' would be vacuous (mostly AttributeError — names the real obstacle, not a sampling gap)

· 17 more finding(s) on lines outside this diff (see the check run).

[safishamsi]

Thanks @hopstreax. Heads-up that most of this PR's scope already landed on v8: #3848 (0.9.68) added Kotlin annotations (class/function/property, use-site targets, bracketed lists) as references/context=attribute edges and val/var primary-constructor properties as field refs, and #3915 (0.9.72) fixed the annotated-inferred-property UnboundLocalError. So the annotation + ctor-property extraction this PR describes is redundant now, and a cherry-pick would conflict heavily in the _kotlin_annotation_names / ctor-param region.

There is one genuinely net-new piece worth having on its own: annotation-argument class-literal extraction — mining X::class inside annotation arguments (e.g. @ManyToOne(targetEntity = Customer::class) -> Order --references(attribute)--> Customer) via your _kotlin_annotation_class_literal_refs helper. Nothing in v8 does that today, it's well-tested, and it correctly uses context=attribute, filters builtins, strips FQ prefixes, and de-dups.

Could you resubmit just that slice as a focused PR against current v8 (the ::class helper + its emit sites + test_kotlin_annotation_class_literal), dropping the now-redundant annotation/ctor refactor and its duplicate tests? That'll rebase cleanly and land easily. (Optional: if you can confirm the inferred-type ctor-param annotation case actually produces a different graph on current v8 — tree-sitter-kotlin tends to parse @Volatile var x = 0 in a ctor as an ERROR node — that hoist could come along too.) Marking changes-requested for now.