Add compiled-language extraction coverage

packages/rbtr/src/rbtr/languages/c.py

@@ -1,7 +1,10 @@
 """C language plugin.
 
-Provides symbol extraction (functions, structs, enums) and
-include directive capture.
+Provides symbol extraction (functions, function prototypes, structs,
+unions, enums, typedefs, global variables, and function/object-like
+macros) and include directive capture.  Object-like macros and
+pointer-declared globals are variables; function-like macros and
+prototypes are functions.
 
 Extracted chunks::
 
@@ -34,18 +37,55 @@
   path: (string_literal) @_import_module) @import
 
 (struct_specifier
-  name: (type_identifier) @_cls_name) @class
+  name: (type_identifier) @_cls_name
+  body: (field_declaration_list)) @class
+
+(union_specifier
+  name: (type_identifier) @_cls_name
+  body: (field_declaration_list)) @class
 
 (enum_specifier
-  name: (type_identifier) @_cls_name) @class
+  name: (type_identifier) @_cls_name
+  body: (enumerator_list)) @class
+
+(enumerator
+  name: (identifier) @_var_name) @variable
 
 (type_definition
   declarator: (type_identifier) @_cls_name) @class
 
+(type_definition
+  declarator: (function_declarator
+    declarator: (parenthesized_declarator
+      (pointer_declarator
+        declarator: (type_identifier) @_cls_name)))) @class
+
+(preproc_function_def
+  name: (identifier) @_fn_name) @function
+
+(preproc_def
+  name: (identifier) @_var_name) @variable
+
+(translation_unit
+  (declaration
+    declarator: (function_declarator
+      declarator: (identifier) @_fn_name)) @function)
+
 (translation_unit
   (declaration
     declarator: (init_declarator
       declarator: (identifier) @_var_name)) @variable)
+
+(translation_unit
+  (declaration
+    declarator: (pointer_declarator
+      declarator: (identifier) @_var_name)) @variable)
+
+(translation_unit
+  (declaration
+    declarator: (init_declarator
+      declarator: (pointer_declarator
+        declarator: (identifier) @_var_name))) @variable)
 """
 
 # ── Plugin ───────────────────────────────────────────────────────────
@@ -68,6 +108,6 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 doc_comment_node_types=frozenset({"comment"}),
                 source_roots=("", "include", "src"),
                 test_prefix="test_",
-                language_plugin_version=2,
+                language_plugin_version=3,
             ),
         ]

packages/rbtr/src/rbtr/languages/cpp.py

@@ -1,7 +1,10 @@
 """C++ language plugin.
 
-Provides symbol extraction (functions, classes, structs, enums,
-methods) and include directive capture.
+Provides symbol extraction (functions, prototypes, classes, structs,
+unions, enums, type aliases, namespaces, methods including operator
+overloads, namespace/global variables, and function/object-like
+macros) and include directive capture.  Namespaces are both a symbol
+(class) and a naming scope.
 
 Extracted chunks::
 
@@ -31,31 +34,101 @@ class Foo {
   declarator: (function_declarator
     declarator: (field_identifier) @_fn_name)) @function
 
+(function_definition
+  declarator: (function_declarator
+    declarator: (operator_name) @_fn_name)) @function
+
+(field_declaration
+  declarator: (function_declarator
+    declarator: (field_identifier) @_method_name)) @method
+
+(field_declaration_list
+  (declaration
+    declarator: (function_declarator
+      declarator: (identifier) @_method_name)) @method)
+
 (preproc_include
   path: (system_lib_string) @_import_module) @import
 
 (preproc_include
   path: (string_literal) @_import_module) @import
 
+(preproc_function_def
+  name: (identifier) @_fn_name) @function
+
+(preproc_def
+  name: (identifier) @_var_name) @variable
+
 (class_specifier
-  name: (type_identifier) @_cls_name) @class
+  name: (type_identifier) @_cls_name
+  body: (field_declaration_list)) @class
 
 (struct_specifier
-  name: (type_identifier) @_cls_name) @class
+  name: (type_identifier) @_cls_name
+  body: (field_declaration_list)) @class
+
+(union_specifier
+  name: (type_identifier) @_cls_name
+  body: (field_declaration_list)) @class
 
 (enum_specifier
+  name: (type_identifier) @_cls_name
+  body: (enumerator_list)) @class
+
+(alias_declaration
   name: (type_identifier) @_cls_name) @class
 
+(namespace_definition
+  name: (namespace_identifier) @_cls_name) @class
+
+(concept_definition
+  name: (identifier) @_cls_name) @class
+
+(translation_unit
+  (declaration
+    declarator: (function_declarator
+      declarator: (identifier) @_fn_name)) @function)
+
+(namespace_definition
+  body: (declaration_list
+    (declaration
+      declarator: (function_declarator
+        declarator: (identifier) @_fn_name)) @function))
+
 (translation_unit
   (declaration
     declarator: (init_declarator
       declarator: (identifier) @_var_name)) @variable)
 
+(translation_unit
+  (declaration
+    declarator: (pointer_declarator
+      declarator: (identifier) @_var_name)) @variable)
+
+(translation_unit
+  (declaration
+    declarator: (init_declarator
+      declarator: (pointer_declarator
+        declarator: (identifier) @_var_name))) @variable)
+
 (namespace_definition
   body: (declaration_list
     (declaration
       declarator: (init_declarator
         declarator: (identifier) @_var_name)) @variable))
+
+(namespace_definition
+  body: (declaration_list
+    (declaration
+      declarator: (pointer_declarator
+        declarator: (identifier) @_var_name)) @variable))
+
+(namespace_definition
+  body: (declaration_list
+    (declaration
+      declarator: (init_declarator
+        declarator: (pointer_declarator
+          declarator: (identifier) @_var_name))) @variable))
 """
 
 # ── Plugin ───────────────────────────────────────────────────────────
@@ -80,6 +153,6 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 doc_comment_node_types=frozenset({"comment"}),
                 source_roots=("", "include", "src"),
                 test_prefix="test_",
-                language_plugin_version=4,
+                language_plugin_version=5,
             ),
         ]

packages/rbtr/src/rbtr/languages/go.py

@@ -27,20 +27,35 @@
   name: (identifier) @_fn_name) @function
 
 (method_declaration
+  receiver: (parameter_list
+    (parameter_declaration
+      type: [
+        (type_identifier) @_scope
+        (pointer_type (type_identifier) @_scope)
+      ]))
   name: (field_identifier) @_method_name) @method
 
 (type_declaration
   (type_spec
     name: (type_identifier) @_cls_name)) @class
 
+(type_spec
+  (interface_type
+    (method_elem
+      name: (field_identifier) @_method_name) @method))
+
+(type_declaration
+  (type_alias
+    name: (type_identifier) @_cls_name)) @class
+
 (import_declaration
   (import_spec
-    path: (interpreted_string_literal) @_import_module)) @import
+    path: (interpreted_string_literal) @_import_module) @import)
 
 (import_declaration
   (import_spec_list
     (import_spec
-      path: (interpreted_string_literal) @_import_module))) @import
+      path: (interpreted_string_literal) @_import_module) @import))
 
 (source_file
   (var_declaration
@@ -85,6 +100,6 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 # type for both line and block forms.
                 doc_comment_node_types=frozenset({"comment"}),
                 test_suffix="_test",
-                language_plugin_version=2,
+                language_plugin_version=3,
             ),
         ]

packages/rbtr/src/rbtr/languages/java.py

@@ -25,9 +25,43 @@ class Service {
 (method_declaration
   name: (identifier) @_method_name) @method
 
+(constructor_declaration
+  name: (identifier) @_method_name) @method
+
+(field_declaration
+  declarator: (variable_declarator
+    name: (identifier) @_var_name)) @variable
+
 (class_declaration
   name: (identifier) @_cls_name) @class
 
+(interface_declaration
+  name: (identifier) @_cls_name) @class
+
+(enum_declaration
+  name: (identifier) @_cls_name) @class
+
+(enum_declaration
+  body: (enum_body
+    (enum_constant
+      name: (identifier) @_var_name) @variable))
+
+(record_declaration
+  name: (identifier) @_cls_name) @class
+
+(record_declaration
+  parameters: (formal_parameters
+    (formal_parameter
+      name: (identifier) @_var_name) @variable))
+
+(annotation_type_declaration
+  name: (identifier) @_cls_name) @class
+
+(annotation_type_declaration
+  body: (annotation_type_body
+    (annotation_type_element_declaration
+      name: (identifier) @_method_name) @method))
+
 (import_declaration
   (scoped_identifier) @_import_module) @import
 """
@@ -53,14 +87,22 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 extensions=frozenset({".java"}),
                 grammar_module="tree_sitter_java",
                 query=_QUERY,
-                scope_types=frozenset({"class_declaration"}),
+                scope_types=frozenset(
+                    {
+                        "class_declaration",
+                        "interface_declaration",
+                        "enum_declaration",
+                        "record_declaration",
+                        "annotation_type_declaration",
+                    }
+                ),
                 # Javadoc uses `block_comment`; `//` runs use
                 # `line_comment`.  Attach either when they sit
                 # directly above a method or class.
                 doc_comment_node_types=frozenset({"block_comment", "line_comment"}),
                 source_roots=("", "src/main/java"),
                 test_suffix="Test",
                 module_style=ModuleStyle.DOTTED,
-                language_plugin_version=2,
+                language_plugin_version=3,
             ),
         ]

packages/rbtr/src/rbtr/languages/rust.py

@@ -1,7 +1,9 @@
 """Rust language plugin.
 
-Provides full support: functions, structs, enums, impl blocks,
-and `use` declaration extraction.
+Provides full support: functions, structs, enums, unions, type
+aliases, traits, modules, impl blocks, `macro_rules!` definitions,
+and `use` declaration extraction. Traits, modules, and impl blocks
+form naming scopes; trait and impl members are methods.
 
 Extracted chunks::
 
@@ -36,12 +38,35 @@
 (function_item
   name: (identifier) @_fn_name) @function
 
+(function_signature_item
+  name: (identifier) @_fn_name) @function
+
 (struct_item
   name: (type_identifier) @_cls_name) @class
 
 (enum_item
   name: (type_identifier) @_cls_name) @class
 
+(enum_item
+  body: (enum_variant_list
+    (enum_variant
+      name: (identifier) @_var_name) @variable))
+
+(union_item
+  name: (type_identifier) @_cls_name) @class
+
+(type_item
+  name: (type_identifier) @_cls_name) @class
+
+(trait_item
+  name: (type_identifier) @_cls_name) @class
+
+(mod_item
+  name: (identifier) @_cls_name) @class
+
+(macro_definition
+  name: (identifier) @_fn_name) @function
+
 (impl_item
   type: (type_identifier) @_cls_name) @class
 
@@ -160,8 +185,10 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 grammar_module="tree_sitter_rust",
                 query=_QUERY,
                 import_extractor=extract_import_meta,
-                scope_types=frozenset({"impl_item", "struct_item", "mod_item"}),
-                class_scope_types=frozenset({"impl_item", "struct_item"}),
+                scope_types=frozenset(
+                    {"impl_item", "struct_item", "trait_item", "mod_item", "enum_item"}
+                ),
+                class_scope_types=frozenset({"impl_item", "struct_item", "trait_item"}),
                 # Rust splits doc comments (`///`, `//!`) from
                 # regular line/block comments at parse time by
                 # wrapping them in distinct grammar rules, but
@@ -172,6 +199,6 @@ def rbtr_register_languages(self) -> list[LanguageRegistration]:
                 index_files=frozenset({"mod.rs"}),
                 path_substitutions=(("crate/", "src/"),),
                 test_prefix="test_",
-                language_plugin_version=3,
+                language_plugin_version=4,
             ),
         ]

packages/rbtr/src/rbtr/tests/languages/__snapshots__/test_samples/test_sample_edges_match_snapshot[c].json

@@ -0,0 +1,6 @@
+[
+  "c.c::#include \"greeter.h\" -> greeter.h::GREETER_H [imports]",
+  "c.c::#include \"greeter.h\" -> greeter.h::Greeter [imports]",
+  "c.c::#include \"greeter.h\" -> greeter.h::Greeter [imports]",
+  "c.c::#include \"greeter.h\" -> greeter.h::MAX_NAME [imports]"
+]