code-gauge 3.1.0 → 4.0.0

package/native/src/lib.rs

@@ -0,0 +1,101 @@
+#![deny(clippy::all)]
+
+use napi::bindgen_prelude::*;
+use napi_derive::napi;
+
+use crate::duplication::DuplicationSettings;
+
+mod complexity;
+mod dep_degree;
+mod duplication;
+mod functions;
+mod languages;
+mod measure;
+mod ncss;
+mod types;
+mod util;
+
+/// Version of the native payload schema. The TypeScript wrapper refuses a binding whose version
+/// differs from the one it expects, so a stale prebuilt addon fails with a clear rebuild message
+/// instead of silently returning an incompatible payload. Bump on every payload-shape change,
+/// together with `expectedPayloadVersion` in src/nativeMetrics.ts.
+#[napi]
+pub fn payload_version() -> u32 {
+    4
+}
+
+/// Measures code metrics for the given source, returning the NativeMetrics payload as JSON.
+/// The TypeScript wrapper derives the remaining float metrics (Halstead volume/effort/...): they
+/// involve transcendental functions whose last-bit results can differ between V8 and Rust's libm,
+/// and results must not depend on which side computes them.
+#[napi]
+pub fn measure_code_native(
+    code: String,
+    language: String,
+    include_syntax_tree: Option<bool>,
+    min_tokens: Option<u32>,
+    max_gap_tokens: Option<u32>,
+    min_similarity_percent: Option<u32>,
+) -> Result<String> {
+    let definition = find_language(&language)?;
+    let settings = to_duplication_settings(min_tokens, max_gap_tokens, min_similarity_percent);
+    let metrics = measure::measure(
+        &code,
+        definition,
+        include_syntax_tree.unwrap_or(false),
+        &settings,
+    )
+    .map_err(Error::from_reason)?;
+    serde_json::to_string(&metrics).map_err(|error| Error::from_reason(error.to_string()))
+}
+
+/// Collects one file's cross-file clone-detection contribution (candidates, normalized token
+/// stream, statement structure, and code line numbers) as JSON; see CrossFileFileData.
+#[napi]
+pub fn collect_cross_file_data_native(
+    code: String,
+    language: String,
+    min_tokens: Option<u32>,
+) -> Result<String> {
+    let definition = find_language(&language)?;
+    let min_tokens = min_tokens
+        .map(|value| value as usize)
+        .unwrap_or(DuplicationSettings::default().min_tokens);
+    let data = measure::collect_cross_file_data(&code, definition, min_tokens)
+        .map_err(Error::from_reason)?;
+    serde_json::to_string(&data).map_err(|error| Error::from_reason(error.to_string()))
+}
+
+/// Collects normalized token hash sequences of every function as JSON (number[][]),
+/// index-parallel to the functions array of measure_code_native.
+#[napi]
+pub fn collect_function_token_sequences_native(code: String, language: String) -> Result<String> {
+    let definition = find_language(&language)?;
+    let sequences =
+        measure::collect_function_token_sequences(&code, definition).map_err(Error::from_reason)?;
+    serde_json::to_string(&sequences).map_err(|error| Error::from_reason(error.to_string()))
+}
+
+fn find_language(language: &str) -> Result<&'static languages::LanguageDefinition> {
+    languages::find_language(language)
+        .ok_or_else(|| Error::from_reason(format!("Unsupported language: {language}")))
+}
+
+fn to_duplication_settings(
+    min_tokens: Option<u32>,
+    max_gap_tokens: Option<u32>,
+    min_similarity_percent: Option<u32>,
+) -> DuplicationSettings {
+    let defaults = DuplicationSettings::default();
+    DuplicationSettings {
+        min_tokens: min_tokens
+            .map(|value| value as usize)
+            .unwrap_or(defaults.min_tokens),
+        max_gap_tokens: max_gap_tokens
+            .map(|value| value as usize)
+            .unwrap_or(defaults.max_gap_tokens),
+        min_similarity_percent: min_similarity_percent
+            .map(|value| value as usize)
+            .unwrap_or(defaults.min_similarity_percent),
+    }
+}

package/native/src/measure.rs

@@ -0,0 +1,590 @@
+use std::collections::{HashMap, HashSet};
+use std::sync::OnceLock;
+use tree_sitter::Node;
+
+use crate::complexity::{
+    is_lambda_body_block, measure_complexity, measure_function_body_metrics, LanguageSets,
+};
+use crate::dep_degree::measure_dep_degree;
+use crate::duplication::{
+    collect_cross_file_file_data, hash_text, measure_duplication, DuplicationSettings,
+};
+use crate::functions::{
+    collect_nodes, count_parameters, find_function_name, is_implemented_function,
+};
+use crate::languages::LanguageDefinition;
+use crate::types::{
+    CrossFileFileData, FunctionMetrics, HalsteadCounts, LineMetrics, NativeMetrics,
+};
+use crate::util::{all_children, is_js_whitespace, named_children, node_text, split_lines, Source};
+
+pub fn measure(
+    code: &str,
+    language: &LanguageDefinition,
+    include_syntax_tree: bool,
+    duplication_settings: &DuplicationSettings,
+) -> Result<NativeMetrics, String> {
+    let source = Source::new(code);
+    let tree = parse_source(&source, language)?;
+    let root = tree.root_node();
+    let code = &source;
+    let sets = LanguageSets::new(language);
+
+    let functions: Vec<Node<'_>> = collect_nodes(root, &sets.function_nodes)
+        .into_iter()
+        .filter(|node| !is_lambda_body_block(*node) && is_implemented_function(*node))
+        .collect();
+
+    let body_metrics_by_node_id = measure_function_body_metrics(root, &sets, code);
+    let function_metrics: Vec<FunctionMetrics> = functions
+        .iter()
+        .map(|node| {
+            let body_metrics = body_metrics_by_node_id
+                .get(&node.id())
+                .expect("every collected function node opens a frame in the body-metrics pass");
+            FunctionMetrics {
+                name: find_function_name(*node, code),
+                node_type: node.kind().to_string(),
+                start_line: node.start_position().row + 1,
+                // The tree is parsed from UTF-16, so columns are UTF-16 code units x 2 — halving
+                // yields the JavaScript string (UTF-16 code unit) column.
+                start_column: node.start_position().column / 2,
+                end_line: node.end_position().row + 1,
+                // Sonar's written spec adds +1 cognitive complexity per function in a recursion
+                // cycle, but this is intentionally not implemented (issue #22): mainstream
+                // implementations (PMD, SonarQube analyzers) omit it.
+                cognitive_complexity: body_metrics.cognitive_complexity,
+                nesting_depth: body_metrics.nesting_depth,
+                ncss: body_metrics.ncss,
+                parameter_count: count_parameters(*node, code),
+                halstead_counts: measure_halstead(*node, code),
+                dep_degree: measure_dep_degree(*node, code, &sets.function_nodes),
+            }
+        })
+        .collect();
+
+    let global_complexity = measure_complexity(root, &sets, code);
+    let (lines, code_line_numbers) = classify_lines(code, root);
+    let halstead_counts = measure_halstead(root, code);
+
+    Ok(NativeMetrics {
+        language: language.name.to_string(),
+        bytes: code.code.len(),
+        lines,
+        cognitive_complexity: global_complexity.cognitive_complexity,
+        max_cognitive_complexity: function_metrics
+            .iter()
+            .map(|function| function.cognitive_complexity)
+            .max()
+            .unwrap_or(0),
+        nesting_depth: global_complexity.nesting_depth,
+        ncss_count: crate::ncss::count_ncss(root, &sets.ncss_nodes, &sets.ncss_containers),
+        duplication: measure_duplication(root, &code_line_numbers, code, duplication_settings),
+        halstead_counts,
+        functions: function_metrics,
+        syntax_tree: if include_syntax_tree {
+            Some(root.to_sexp())
+        } else {
+            None
+        },
+    })
+}
+
+/// Collects one file's cross-file clone-detection contribution; see CrossFileFileData.
+pub fn collect_cross_file_data(
+    code: &str,
+    language: &LanguageDefinition,
+    min_tokens: usize,
+) -> Result<CrossFileFileData, String> {
+    let source = Source::new(code);
+    let tree = parse_source(&source, language)?;
+    let root = tree.root_node();
+    let (candidates, tokens, container_statements) =
+        collect_cross_file_file_data(root, &source, min_tokens);
+    let (_, code_line_numbers) = classify_lines(&source, root);
+    let mut code_line_numbers: Vec<usize> = code_line_numbers.into_iter().collect();
+    code_line_numbers.sort_unstable();
+    Ok(CrossFileFileData {
+        candidates,
+        tokens,
+        container_statements,
+        code_line_numbers,
+    })
+}
+
+/// Name-carrying leaf types anonymized by tokenize_function so consistent renames still match.
+const IDENTIFIER_LEAF_NODE_TYPES: &[&str] = &[
+    "identifier",
+    "property_identifier",
+    "field_identifier",
+    "type_identifier",
+    "constant",
+    "instance_variable",
+    "class_variable",
+    "global_variable",
+];
+
+/// Normalized token hash sequences of every function, index-parallel to the functions array of
+/// measure(); a faithful port of tokenizeFunction in src/metrics.ts.
+pub fn collect_function_token_sequences(
+    code: &str,
+    language: &LanguageDefinition,
+) -> Result<Vec<Vec<i32>>, String> {
+    let source = Source::new(code);
+    let tree = parse_source(&source, language)?;
+    let root = tree.root_node();
+    let sets = LanguageSets::new(language);
+    Ok(collect_nodes(root, &sets.function_nodes)
+        .into_iter()
+        .filter(|node| !is_lambda_body_block(*node) && is_implemented_function(*node))
+        .map(|node| {
+            let mut symbols = Vec::new();
+            let mut id_index_by_name: HashMap<String, usize> = HashMap::new();
+            collect_token_symbols(node, &source, &mut symbols, &mut id_index_by_name);
+            symbols
+        })
+        .collect())
+}
+
+fn collect_token_symbols(
+    node: Node<'_>,
+    code: &Source<'_>,
+    symbols: &mut Vec<i32>,
+    id_index_by_name: &mut HashMap<String, usize>,
+) {
+    if matches!(node.kind(), "comment" | "line_comment" | "block_comment") {
+        return;
+    }
+    if atomic_operand_node_types().contains(node.kind()) {
+        symbols.push(hash_text(node.kind()));
+        return;
+    }
+    if node.child_count() > 0 {
+        for child in all_children(node) {
+            collect_token_symbols(child, code, symbols, id_index_by_name);
+        }
+        return;
+    }
+    if IDENTIFIER_LEAF_NODE_TYPES.contains(&node.kind()) {
+        let next_index = id_index_by_name.len();
+        let index = *id_index_by_name
+            .entry(node_text(node, code).to_string())
+            .or_insert(next_index);
+        symbols.push(hash_text(&format!("id{index}")));
+        return;
+    }
+    // Remaining operand leaves are literals, normalized by kind; everything else (keywords,
+    // operators, punctuation) is kept verbatim.
+    symbols.push(hash_text(if operand_node_types().contains(node.kind()) {
+        node.kind()
+    } else {
+        node_text(node, code)
+    }));
+}
+
+/// Parses the source from UTF-16 (matching node-tree-sitter's JavaScript string semantics:
+/// tree-sitter's error recovery differs between input encodings for malformed non-ASCII source)
+/// and refuses pathologically deep trees: the metric passes recurse per tree level and would
+/// overflow the native stack (a process-killing SIGSEGV, not a catchable error) around depth ~20k.
+fn parse_source(
+    source: &Source<'_>,
+    language: &LanguageDefinition,
+) -> Result<tree_sitter::Tree, String> {
+    let mut parser = tree_sitter::Parser::new();
+    parser
+        .set_language(&language.grammar())
+        .map_err(|error| error.to_string())?;
+    let tree = parser
+        .parse_utf16(source.to_utf16(), None)
+        .ok_or_else(|| "parse failed".to_string())?;
+    if tree_depth(tree.root_node()) > MAX_TREE_DEPTH {
+        return Err(format!("tree depth exceeds {MAX_TREE_DEPTH}"));
+    }
+    Ok(tree)
+}
+
+/// See the depth check in parse_source(); computed iteratively so the check itself cannot overflow.
+const MAX_TREE_DEPTH: usize = 5_000;
+
+fn tree_depth(root: Node<'_>) -> usize {
+    let mut cursor = root.walk();
+    let mut depth = 0;
+    let mut max_depth = 0;
+    loop {
+        if cursor.goto_first_child() {
+            depth += 1;
+            max_depth = max_depth.max(depth);
+            continue;
+        }
+        loop {
+            if cursor.goto_next_sibling() {
+                break;
+            }
+            if !cursor.goto_parent() {
+                return max_depth;
+            }
+            depth -= 1;
+        }
+    }
+}
+
+struct CommentSpan {
+    line: usize,
+    start_column: usize,
+    end_column: usize,
+}
+
+/// 1-based numbers of lines that are neither blank nor comment-only, matching classifyLines in
+/// metrics.ts so duplication line coverage and its code-line denominator agree.
+fn classify_lines(code: &Source<'_>, root: Node<'_>) -> (LineMetrics, HashSet<usize>) {
+    let source_lines = split_lines(code.code);
+    // Spans are bucketed by line so classification stays linear.
+    let mut comment_spans_by_line: HashMap<usize, Vec<CommentSpan>> = HashMap::new();
+    for span in collect_comment_spans(root) {
+        comment_spans_by_line
+            .entry(span.line)
+            .or_default()
+            .push(span);
+    }
+    let mut blank = 0;
+    let mut comment = 0;
+    let mut code_line_numbers = HashSet::new();
+
+    for (index, line) in source_lines.iter().enumerate() {
+        if line.chars().all(is_js_whitespace) {
+            blank += 1;
+            continue;
+        }
+        let empty_spans = Vec::new();
+        let relevant_spans = comment_spans_by_line.get(&index).unwrap_or(&empty_spans);
+        if is_comment_only_line(line, relevant_spans) {
+            comment += 1;
+        } else {
+            code_line_numbers.insert(index + 1);
+        }
+    }
+
+    (
+        LineMetrics {
+            total: source_lines.len(),
+            code: code_line_numbers.len(),
+            comment,
+            blank,
+        },
+        code_line_numbers,
+    )
+}
+
+fn collect_comment_spans(root: Node<'_>) -> Vec<CommentSpan> {
+    let mut spans = Vec::new();
+
+    fn visit(node: Node<'_>, spans: &mut Vec<CommentSpan>) {
+        if matches!(node.kind(), "comment" | "line_comment" | "block_comment") {
+            for row in node.start_position().row..=node.end_position().row {
+                // Node columns are UTF-16 code units x 2 (the tree is parsed from UTF-16);
+                // halving matches the code-unit columns the line scan below counts.
+                spans.push(CommentSpan {
+                    line: row,
+                    start_column: if row == node.start_position().row {
+                        node.start_position().column / 2
+                    } else {
+                        0
+                    },
+                    end_column: if row == node.end_position().row {
+                        node.end_position().column / 2
+                    } else {
+                        usize::MAX
+                    },
+                });
+            }
+        }
+
+        for child in named_children(node) {
+            visit(child, spans);
+        }
+    }
+
+    visit(root, &mut spans);
+    spans
+}
+
+fn is_comment_only_line(line: &str, relevant_spans: &[CommentSpan]) -> bool {
+    if relevant_spans.is_empty() {
+        return false;
+    }
+
+    // A line may hold several comments (`/* one */ /* two */`), so every non-whitespace column must
+    // be covered by the UNION of spans, not by a single span. Columns are UTF-16 code units,
+    // matching the span columns derived from the UTF-16 parse.
+    let mut column = 0;
+    for character in line.chars() {
+        if !is_js_whitespace(character)
+            && !relevant_spans
+                .iter()
+                .any(|span| span.start_column <= column && column < span.end_column)
+        {
+            return false;
+        }
+        column += character.len_utf16();
+    }
+    true
+}
+
+const OPERATOR_TEXTS: &[&str] = &[
+    "+",
+    "-",
+    "*",
+    "/",
+    "%",
+    "**",
+    "=",
+    "+=",
+    "-=",
+    "*=",
+    "/=",
+    "%=",
+    "==",
+    "!=",
+    "===",
+    "!==",
+    "<",
+    "<=",
+    ">",
+    ">=",
+    "!",
+    "~",
+    "&",
+    "|",
+    "^",
+    "++",
+    "--",
+    "<<",
+    ">>",
+    ">>>",
+    "=>",
+    "**=",
+    "<<=",
+    ">>=",
+    ">>>=",
+    "&=",
+    "|=",
+    "^=",
+    "&&=",
+    "||=",
+    "??=",
+    "??",
+    "?.",
+    "?",
+    "//",
+    "//=",
+    "@",
+    "@=",
+    ":=",
+    "<-",
+    "<=>",
+    "=~",
+    "..",
+    "...",
+    "..=",
+    "&&",
+    "||",
+    "!~",
+    "&^",
+    "&^=",
+    "&.",
+    // Member access/qualification are classical Halstead operators; `->` also captures
+    // Python/Rust return-type arrows, consistent with the counted `=>`.
+    ".",
+    "->",
+    "::",
+    "->*",
+    ".*",
+    "sizeof",
+    "alignof",
+    "defined?",
+    "as",
+    // C++ alternative operator tokens parse as anonymous leaves like their symbolic forms.
+    "bitand",
+    "bitor",
+    "xor",
+    "compl",
+    "and_eq",
+    "or_eq",
+    "xor_eq",
+    "not_eq",
+    "and",
+    "or",
+    "not",
+    "in",
+    "is",
+    "instanceof",
+    "typeof",
+    "new",
+    "delete",
+    "return",
+    "throw",
+    "raise",
+    "yield",
+    "await",
+    "co_await",
+    "co_yield",
+    "co_return",
+    "break",
+    "continue",
+];
+
+const OPERAND_NODE_TYPES: &[&str] = &[
+    "identifier",
+    "property_identifier",
+    "field_identifier",
+    "type_identifier",
+    "constant",
+    "instance_variable",
+    "class_variable",
+    "global_variable",
+    "simple_symbol",
+    "self",
+    "this",
+    "super",
+    // C/C++/Rust built-in types are leaves of their own node type, unlike Go's `type_identifier`.
+    "primitive_type",
+    "boolean_type",
+    "void_type",
+    "auto",
+    "number",
+    "integer",
+    "float",
+    "integer_literal",
+    "float_literal",
+    "int_literal",
+    "rune_literal",
+    "imaginary_literal",
+    "number_literal",
+    "decimal_integer_literal",
+    "hex_integer_literal",
+    "octal_integer_literal",
+    "binary_integer_literal",
+    "decimal_floating_point_literal",
+    "hex_floating_point_literal",
+    "string",
+    "string_literal",
+    // Go raw strings are leaves with no content child, unlike Rust/C++ `raw_string_literal`s.
+    "raw_string_literal",
+    "string_fragment",
+    "multiline_string_fragment",
+    "string_content",
+    "raw_string_content",
+    "template_string",
+    "character_literal",
+    "char_literal",
+    "character",
+    "true",
+    "false",
+    "null",
+    "null_literal",
+    "undefined",
+    "nil",
+    "none",
+];
+
+/// Non-leaf literals counted as one Halstead operand without descending; see metrics.ts.
+const ATOMIC_OPERAND_NODE_TYPES: &[&str] = &[
+    "interpreted_string_literal",
+    "regex",
+    "user_defined_literal",
+    "integral_type",
+    "floating_point_type",
+    "sized_type_specifier",
+    "placeholder_type_specifier",
+];
+
+fn operator_texts() -> &'static HashSet<&'static str> {
+    static SET: OnceLock<HashSet<&'static str>> = OnceLock::new();
+    SET.get_or_init(|| OPERATOR_TEXTS.iter().copied().collect())
+}
+
+fn operand_node_types() -> &'static HashSet<&'static str> {
+    static SET: OnceLock<HashSet<&'static str>> = OnceLock::new();
+    SET.get_or_init(|| OPERAND_NODE_TYPES.iter().copied().collect())
+}
+
+fn atomic_operand_node_types() -> &'static HashSet<&'static str> {
+    static SET: OnceLock<HashSet<&'static str>> = OnceLock::new();
+    SET.get_or_init(|| ATOMIC_OPERAND_NODE_TYPES.iter().copied().collect())
+}
+
+fn measure_halstead(root: Node<'_>, code: &Source<'_>) -> HalsteadCounts {
+    let mut operators: HashMap<String, u64> = HashMap::new();
+    let mut operands: HashMap<String, u64> = HashMap::new();
+
+    fn visit(
+        node: Node<'_>,
+        code: &Source<'_>,
+        operators: &mut HashMap<String, u64>,
+        operands: &mut HashMap<String, u64>,
+    ) {
+        if matches!(node.kind(), "comment" | "line_comment" | "block_comment") {
+            return;
+        }
+
+        if atomic_operand_node_types().contains(node.kind()) {
+            *operands
+                .entry(node_text(node, code).to_string())
+                .or_insert(0) += 1;
+            return;
+        }
+
+        // Operators are counted from leaf tokens only: keyword-named nodes always contain a
+        // same-text anonymous keyword leaf, so counting the named node as well would double-count.
+        if node.child_count() == 0 {
+            let text = node_text(node, code);
+            // Operands win over text matches so identifiers spelled like word operators stay operands.
+            if operand_node_types().contains(node.kind()) {
+                *operands.entry(text.to_string()).or_insert(0) += 1;
+            } else if (operator_texts().contains(text) || operator_texts().contains(node.kind()))
+                && is_countable_contextual_token(node, text)
+            {
+                let key = if text.is_empty() { node.kind() } else { text };
+                *operators.entry(key.to_string()).or_insert(0) += 1;
+            }
+            return;
+        }
+
+        for child in all_children(node) {
+            visit(child, code, operators, operands);
+        }
+    }
+
+    visit(root, code, &mut operators, &mut operands);
+
+    HalsteadCounts {
+        distinct_operators: operators.len(),
+        distinct_operands: operands.len(),
+        total_operators: operators.values().sum(),
+        total_operands: operands.values().sum(),
+    }
+}
+
+/// Ternary/conditional and Rust try parents make `?` an operator; TS optional markers do not.
+const QUESTION_OPERATOR_PARENT_TYPES: &[&str] = &[
+    "ternary_expression",
+    "conditional_expression",
+    "conditional",
+    "try_expression",
+    // TypeScript conditional types (`T extends U ? X : Y`) select like a ternary.
+    "conditional_type",
+];
+
+fn is_countable_contextual_token(node: Node<'_>, text: &str) -> bool {
+    if text == "@" {
+        // Python matrix multiplication only; decorator/annotation `@` marks are not operators.
+        let parent_type = node.parent().map(|parent| parent.kind());
+        return parent_type == Some("binary_operator")
+            || parent_type == Some("augmented_assignment");
+    }
+    if text != "?" {
+        return true;
+    }
+    node.parent()
+        .is_some_and(|parent| QUESTION_OPERATOR_PARENT_TYPES.contains(&parent.kind()))
+}