fix(extract): prevent duplicate Function nodes for Java interface methods (#1234)

[harshitaajoshi]

What does this PR do?

Fixes #1234.

Java interface methods were emitted as both a Method node (correct, via extract_class_methods) and a duplicate Function node (incorrect, via the walk_defs fallback). The root cause is that push_class_body_children() in extract_defs.c did not recognize interface_body as a class body container, even though find_class_body() already did. When the body type was unrecognized, the fallback path pushed all children (including method_declaration nodes) back onto the walk stack, where extract_func_def() minted a spurious Function node for each one.

This change adds interface_body and enum_body to the body container check in push_class_body_children(), consistent with find_class_body(). Both node types are used across multiple tree-sitter grammars (Java, TypeScript, Apex, Dart, and others).

A reproduce-first regression test verifies that a Java interface with two method declarations produces one Method node per method, zero Function nodes, and one Interface node for the type itself. The test fails without the fix (count_defs_with_label(r, "Function") == 2, expected 0) and passes with it.

Verification

• build/c/test-runner extraction -- 254 passed
• build/c/test-runner extraction extraction_inheritance extraction_imports pipeline registry grammar_labels grammar_regression -- 569 passed, 0 failed

Checklist

• Every commit is signed off (git commit -s) -- required, CI rejects
  unsigned commits (DCO, see CONTRIBUTING.md)
• Tests pass locally (make -f Makefile.cbm test)
  ASan hangs on macOS Tahoe 26.5 (Apple Clang 17) for any binary, including a trivial int main(){return 0;}. All 569 focused suite tests pass when built without sanitizers (SANITIZE="").
• Lint passes (make -f Makefile.cbm lint-ci)
• New behavior is covered by a test (reproduce-first for bug fixes)