Set cognitive complexity to industry default (25), tag 168 functions

Lower thresholds to best-practice values:
  cognitive-complexity: 250 -> 25 (industry default)
  statements: 400 -> 200
  lines: 800 -> 400

Add NOLINTNEXTLINE to 168 functions that currently exceed these thresholds.
Each suppression will be removed as the function is split in subsequent
commits. Zero net behavioral changes.

Also: extract execute_with_clause from cypher execute_single.

Author: DeusData

.clang-tidy

@@ -76,9 +76,9 @@ CheckOptions:
   readability-implicit-bool-conversion.AllowPointerConditions: true
   readability-magic-numbers.IgnoredIntegerValues: "1;2;3;4;5;6;7;8;9;10;11;12;14;15;16;20;24;32;48;64;100;128;200;256;493;512;755;1000;1024;1040;2048;4096;8192;16384;32768;65536;500000;1000000;1000000000"
   readability-magic-numbers.IgnoredFloatingPointValues: "0.0;0.25;0.5;0.7;0.75;0.8;0.85;0.9;0.95;1.0;2.0;100.0;1e308"
-  readability-function-cognitive-complexity.Threshold: 250
-  readability-function-size.StatementThreshold: 400
-  readability-function-size.LineThreshold: 800
+  readability-function-cognitive-complexity.Threshold: 25
+  readability-function-size.StatementThreshold: 200
+  readability-function-size.LineThreshold: 400
 
 HeaderFilterRegex: '^(src|tests)/.*\.h$'
 

internal/cbm/ac.c

@@ -72,6 +72,7 @@ static void queue_free(Queue *q) {
 //   alpha_size  — alphabet size (256 if alpha_map is NULL)
 //
 // Returns a heap-allocated automaton. Caller must call cbm_ac_free().
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 CBMAutomaton *cbm_ac_build(const char **patterns, const int *lengths, int count,
                            const uint8_t *alpha_map, int alpha_size) {
     if (count <= 0) {

internal/cbm/extract_calls.c

@@ -88,6 +88,7 @@ static bool lean_is_in_type_position(TSNode node) {
 }
 
 // Extract callee name from a call node
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static char *extract_callee_name(CBMArena *a, TSNode node, const char *source, CBMLanguage lang) {
     // Lean 4: apply — name field is callee. Skip if in a type annotation position.
     // Must be checked before the generic "name" field handler below.
@@ -278,6 +279,7 @@ static char *extract_callee_name(CBMArena *a, TSNode node, const char *source, C
 
 // Walk AST for call nodes (iterative)
 #define CALLS_STACK_CAP 512
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_calls(CBMExtractCtx *ctx, TSNode root, const CBMLangSpec *spec) {
     TSNode stack[CALLS_STACK_CAP];
     int top = 0;
@@ -385,6 +387,7 @@ void cbm_extract_calls(CBMExtractCtx *ctx) {
 /* Extract all arguments from a call expression into call->args[].
  * Captures expression text, resolved constants, keyword names, and
  * dotted field chains (member_expression → "payload.info.url"). */
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_call_args(CBMExtractCtx *ctx, TSNode args, CBMCall *call) {
     uint32_t argc = ts_node_named_child_count(args);
     int positional_idx = 0;
@@ -429,6 +432,7 @@ static void extract_call_args(CBMExtractCtx *ctx, TSNode args, CBMCall *call) {
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void handle_calls(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec, WalkState *state) {
     if (!spec->call_node_types || !spec->call_node_types[0]) {
         return;

internal/cbm/extract_defs.c

@@ -40,6 +40,7 @@ static TSNode func_name_node(TSNode node) {
 }
 
 // Resolve the name node for a function, handling language-specific quirks
+// NOLINTNEXTLINE(readability-function-cognitive-complexity,readability-function-size)
 static TSNode resolve_func_name(TSNode node, CBMLanguage lang, const char *source) {
     const char *kind = ts_node_type(node);
 
@@ -398,6 +399,7 @@ static bool is_js_exported(TSNode node) {
 }
 
 // Extract docstring from the node's leading comment
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char *extract_docstring(CBMArena *a, TSNode node, const char *source,
                                      CBMLanguage lang) {
     // Go: type_spec is inside type_declaration; comment is before type_declaration
@@ -498,6 +500,7 @@ static const char *decorator_method_name(const char *attr_text) {
  * Pure AST approach: walks the decorator node's call children to find:
  * 1. The function/attribute name → infer HTTP method
  * 2. The first string argument → route path */
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_route_from_decorators(CBMArena *a, TSNode func_node, const char *source,
                                           const CBMLangSpec *spec, const char **out_path,
                                           const char **out_method) {
@@ -583,6 +586,7 @@ static void extract_route_from_decorators(CBMArena *a, TSNode func_node, const c
 }
 
 // Extract decorator names from preceding decorator/annotation nodes
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char **extract_decorators(CBMArena *a, TSNode node, const char *source,
                                        CBMLanguage lang, const CBMLangSpec *spec) {
     if (!spec->decorator_node_types || !spec->decorator_node_types[0]) {
@@ -657,6 +661,7 @@ static const char **extract_decorators(CBMArena *a, TSNode node, const char *sou
 }
 
 // Extract base class names from a class node
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char **extract_base_classes(CBMArena *a, TSNode node, const char *source,
                                          CBMLanguage lang) {
     (void)lang;
@@ -841,6 +846,7 @@ static char *clean_type_name(CBMArena *a, const char *raw) {
 
 // Extract param_names from a parameter list node.
 // Returns NULL-terminated arena-allocated array.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char **extract_param_names(CBMArena *a, TSNode params, const char *source,
                                         CBMLanguage lang) {
     (void)lang;
@@ -910,6 +916,7 @@ static const char **extract_param_names(CBMArena *a, TSNode params, const char *
 // Extract return_types from a return type node.
 // Parses Go-style multi-return (T1, T2) and single return types.
 // Returns NULL-terminated arena-allocated array.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char **extract_return_types(CBMArena *a, TSNode rt_node, const char *source,
                                          CBMLanguage lang) {
     (void)lang;
@@ -979,6 +986,7 @@ static const char **extract_return_types(CBMArena *a, TSNode rt_node, const char
 
 // Extract param_types from a parameter list node.
 // Returns NULL-terminated arena-allocated array.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char **extract_param_types(CBMArena *a, TSNode params, const char *source,
                                         CBMLanguage lang) {
     if (ts_node_is_null(params)) {
@@ -1131,6 +1139,7 @@ static const char **extract_param_types(CBMArena *a, TSNode params, const char *
 
 // --- Function definition extraction ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_func_def(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec) {
     CBMArena *a = ctx->arena;
 
@@ -1272,6 +1281,7 @@ static void extract_func_def(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec
 
 // --- Class definition extraction ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity,readability-function-size)
 static void extract_class_def(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec) {
     CBMArena *a = ctx->arena;
     const char *kind = ts_node_type(node);
@@ -1587,6 +1597,7 @@ static TSNode find_class_body(TSNode class_node, CBMLanguage lang) {
 }
 
 // Helper: try to extract method name from a node, with fallbacks
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static TSNode resolve_method_name(TSNode child, CBMLanguage lang) {
     TSNode name_node = func_name_node(child);
     if (!ts_node_is_null(name_node)) {
@@ -1661,6 +1672,7 @@ static TSNode resolve_method_name(TSNode child, CBMLanguage lang) {
 }
 
 // Push a single method definition
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void push_method_def(CBMExtractCtx *ctx, TSNode child, const char *class_qn,
                             const CBMLangSpec *spec, TSNode name_node) {
     CBMArena *a = ctx->arena;
@@ -1741,6 +1753,7 @@ static void push_method_def(CBMExtractCtx *ctx, TSNode child, const char *class_
 }
 
 // Extract methods inside a class body
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_class_methods(CBMExtractCtx *ctx, TSNode class_node, const char *class_qn,
                                   const CBMLangSpec *spec) {
     TSNode body = find_class_body(class_node, ctx->language);
@@ -1872,6 +1885,7 @@ static void extract_rust_impl(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec
 
 // --- Elixir def/defp/defmodule ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_elixir_call(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec) {
     CBMArena *a = ctx->arena;
 
@@ -2049,6 +2063,7 @@ static const char *extract_java_field_name(CBMArena *a, TSNode field, const char
 
 /* ── Variable name extractors by language group ─────────────────── */
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_vars_mainstream(CBMExtractCtx *ctx, TSNode node, CBMArena *a,
                                     const char *kind) {
     (void)kind;
@@ -2153,6 +2168,7 @@ static void extract_vars_mainstream(CBMExtractCtx *ctx, TSNode node, CBMArena *a
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_vars_dynamic(CBMExtractCtx *ctx, TSNode node, CBMArena *a, const char *kind) {
     switch (ctx->language) {
     case CBM_LANG_PHP: {
@@ -2288,6 +2304,7 @@ static void extract_vars_dynamic(CBMExtractCtx *ctx, TSNode node, CBMArena *a, c
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_vars_jvm(CBMExtractCtx *ctx, TSNode node, CBMArena *a) {
     switch (ctx->language) {
     case CBM_LANG_SCALA: {
@@ -2344,6 +2361,7 @@ static void extract_vars_jvm(CBMExtractCtx *ctx, TSNode node, CBMArena *a) {
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_vars_config(CBMExtractCtx *ctx, TSNode node, CBMArena *a, const char *kind) {
     switch (ctx->language) {
     case CBM_LANG_YAML: {
@@ -2652,6 +2670,7 @@ static void extract_variables(CBMExtractCtx *ctx, TSNode root, const CBMLangSpec
 // Extract typed struct/class fields for cross-file LSP resolution (C/C++/CUDA/Go/Java/Rust etc.)
 // Creates "Field" label definitions with return_type set to the field's type text.
 // These are later collected by DefsToLSPDefs to build FieldDefs pipe-separated strings.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_class_fields(CBMExtractCtx *ctx, TSNode class_node, const char *class_qn,
                                  const CBMLangSpec *spec) {
     if (!spec->field_node_types || !spec->field_node_types[0]) {
@@ -2819,6 +2838,7 @@ static void push_nested_class_nodes(TSNode body, const CBMLangSpec *spec, walk_d
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_defs(CBMExtractCtx *ctx, TSNode root, const CBMLangSpec *spec, int depth_unused) {
     (void)depth_unused;
     walk_defs_frame_t stack[CBM_WALK_DEFS_STACK_CAP];

internal/cbm/extract_imports.c

@@ -52,6 +52,7 @@ static const char *path_last(CBMArena *a, const char *path) {
 // --- Go imports ---
 // import_declaration -> import_spec_list -> import_spec -> (name, path)
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void parse_go_imports(CBMExtractCtx *ctx) {
     CBMArena *a = ctx->arena;
 
@@ -133,6 +134,7 @@ static void parse_go_imports(CBMExtractCtx *ctx) {
 // import_statement: import X, import X as Y
 // import_from_statement: from X import Y, from X import Y as Z
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void parse_python_imports(CBMExtractCtx *ctx) {
     CBMArena *a = ctx->arena;
 
@@ -244,6 +246,7 @@ static void parse_python_imports(CBMExtractCtx *ctx) {
 // import X from "Y"; import {A, B} from "Y"; import * as X from "Y"
 // const X = require("Y")
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_es_imports(CBMExtractCtx *ctx, TSNode node) {
     CBMArena *a = ctx->arena;
     const char *kind = ts_node_type(node);
@@ -424,6 +427,7 @@ static void parse_rust_imports(CBMExtractCtx *ctx) {
 // --- C/C++ imports ---
 // preproc_include -> path or string_literal
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void parse_c_imports(CBMExtractCtx *ctx) {
     CBMArena *a = ctx->arena;
 
@@ -477,6 +481,7 @@ static void parse_c_imports(CBMExtractCtx *ctx) {
 // --- Ruby imports ---
 // call nodes: require("X"), require_relative("X")
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void parse_ruby_imports(CBMExtractCtx *ctx) {
     CBMArena *a = ctx->arena;
 
@@ -606,6 +611,7 @@ static void parse_lua_imports(CBMExtractCtx *ctx) {
 
 // --- Generic import parsing for languages with simple import_declaration ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void parse_generic_imports(CBMExtractCtx *ctx, const char *node_type) {
     CBMArena *a = ctx->arena;
 
@@ -667,6 +673,7 @@ static void parse_generic_imports(CBMExtractCtx *ctx, const char *node_type) {
 // get_top: << "package" (Get["file"])
 // apply where first child is builtin_symbol "Needs" with string arg
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_wolfram_imports(CBMExtractCtx *ctx, TSNode node) {
     CBMArena *a = ctx->arena;
     const char *kind = ts_node_type(node);

internal/cbm/extract_k8s.c

@@ -91,6 +91,7 @@ static void emit_kustomize_sequence(CBMExtractCtx *ctx, TSNode seq_node, const c
     }
 }
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_kustomize(CBMExtractCtx *ctx) {
     CBMArena *a = ctx->arena;
 
@@ -156,6 +157,7 @@ static void extract_kustomize(CBMExtractCtx *ctx) {
 
 // Descend into the first block_mapping of a document and extract apiVersion,
 // kind, and metadata.name. Returns void; fills kind_buf and meta_name_buf.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void extract_k8s_scalars(CBMExtractCtx *ctx, TSNode mapping, char *kind_buf, size_t kind_sz,
                                 char *meta_name_buf, size_t meta_sz) {
     CBMArena *a = ctx->arena;

internal/cbm/extract_semantic.c

@@ -13,6 +13,7 @@
 // --- Throw/Raise extraction ---
 
 // Process a single node for throw extraction (called from iterative walker).
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void process_throw_node(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec) {
     const char *kind = ts_node_type(node);
 
@@ -105,6 +106,7 @@ static void walk_throws(CBMExtractCtx *ctx, TSNode root, const CBMLangSpec *spec
 // --- Read/Write detection (iterative) ---
 
 #define READWRITE_STACK_CAP 512
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_readwrites(CBMExtractCtx *ctx, TSNode root, const CBMLangSpec *spec) {
     TSNode stack[READWRITE_STACK_CAP];
     int top = 0;
@@ -159,6 +161,7 @@ void cbm_extract_semantic(CBMExtractCtx *ctx) {
 
 // --- Unified handlers ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void handle_throws(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec, WalkState *state) {
     bool has_throws = spec->throw_node_types && spec->throw_node_types[0];
     bool has_clause = spec->throws_clause_field && spec->throws_clause_field[0];

internal/cbm/extract_type_assigns.c

@@ -10,6 +10,7 @@
 
 // Extract class/type name from a constructor expression.
 // e.g., new Foo() -> "Foo", Foo() -> "Foo" (if uppercase), Foo{} -> "Foo"
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char *extract_constructor_type(CBMArena *a, TSNode rhs, const char *source,
                                             CBMLanguage lang) {
     const char *kind = ts_node_type(rhs);
@@ -77,6 +78,7 @@ static const char *extract_constructor_type(CBMArena *a, TSNode rhs, const char
 }
 
 // Walk AST for assignment patterns where RHS is a constructor call.
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_type_assigns(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec) {
     const char *kind = ts_node_type(node);
 
@@ -189,6 +191,7 @@ void cbm_extract_type_assigns(CBMExtractCtx *ctx) {
 
 // --- Unified handler ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void handle_type_assigns(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec,
                          WalkState *state) {
     const char *kind = ts_node_type(node);

internal/cbm/extract_type_refs.c

@@ -142,6 +142,7 @@ static void extract_return_type_refs(CBMExtractCtx *ctx, TSNode func_node, const
 }
 
 // Walk function body for type references (casts, type assertions, local var types, generics).
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_body_type_refs(CBMExtractCtx *ctx, TSNode node, const char *func_qn) {
     const char *kind = ts_node_type(node);
 
@@ -296,6 +297,7 @@ void cbm_extract_type_refs(CBMExtractCtx *ctx) {
 // The cursor visits both, so this single handler replaces the old
 // walk_type_refs + walk_body_type_refs split.
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void handle_type_refs(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec, WalkState *state) {
     if (!spec->function_node_types || !spec->function_node_types[0]) {
         return;

internal/cbm/extract_unified.c

@@ -54,6 +54,7 @@ static void recompute_state(WalkState *state, const char *module_qn) {
 }
 
 // Compute function QN for scope tracking (mirrors cbm_enclosing_func_qn logic).
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static const char *compute_func_qn(CBMExtractCtx *ctx, TSNode node, const CBMLangSpec *spec,
                                    WalkState *state) {
     (void)spec;
@@ -251,6 +252,7 @@ static void handle_string_refs(CBMExtractCtx *ctx, TSNode node, const WalkState
 /* Recursively walk YAML block_mapping_pair nodes, building dotted key paths.
  * Emits string_refs with key_path for leaf values that are URLs or config values.
  * Example: body.operational_info.post_url → "https://..." */
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void walk_yaml_mapping(CBMExtractCtx *ctx, TSNode node, const char *prefix) {
     uint32_t nc = ts_node_named_child_count(node);
     for (uint32_t i = 0; i < nc; i++) {
@@ -386,6 +388,7 @@ static const char *infer_broker(const char *file_path, const char *source_key) {
 
 /* Scan a YAML mapping for source+target key pairs.
  * Collects all key-value pairs at this level and one level deep (for nested config:). */
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void scan_mapping_for_bindings(CBMExtractCtx *ctx, TSNode mapping) {
     const char *sources[8] = {NULL};
     const char *source_keys[8] = {NULL};
@@ -504,6 +507,7 @@ static void scan_yaml_for_infra_bindings(CBMExtractCtx *ctx, TSNode node) {
  * where one is a source key (topic, queue_name) and another is a
  * target key (uri, push_endpoint). Handles nested blocks like
  * push_config { push_endpoint = "..." }. */
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 static void scan_hcl_block_for_bindings(CBMExtractCtx *ctx, TSNode block) {
     const char *sources[8] = {NULL};
     const char *source_keys[8] = {NULL};
@@ -630,6 +634,7 @@ static void handle_yaml_nested(CBMExtractCtx *ctx, TSNode node) {
 
 // --- Main unified cursor walk ---
 
+// NOLINTNEXTLINE(readability-function-cognitive-complexity)
 void cbm_extract_unified(CBMExtractCtx *ctx) {
     const CBMLangSpec *spec = cbm_lang_spec(ctx->language);
     if (!spec) {