pi-read-map - npm Package Compare versions

Comparing version 1.1.0 to 1.2.0

CHANGELOG.md

+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+## [1.2.0] - 2026-02-14
+
+### Added
+
+- **Clojure mapper** — tree-sitter-based parser for `.clj`, `.cljs`, `.cljc`, and `.edn` files. Extracts `ns`, `defn`, `defn-`, `def`, `defonce`, `defmacro`, `defmulti`, `defmethod`, `defprotocol`, `defrecord`, and `deftype` forms with docstrings, signatures, modifiers, and protocol method children. Supports reader conditionals (`#?`) with per-platform annotations. Contributed by [Baishampayan Ghose](https://github.com/ghoseb). ([#2](https://github.com/Whamp/pi-read-map/pull/2))
+- Clojure demo asset (`clojure/core.clj` from the official Clojure repo)
+
+### Changed
+
+- Tree-sitter tests use `describe.runIf` for conditional execution
+
+## [1.1.0] - 2026-02-10
+
+### Added
+
+- Docstring extraction (`FileSymbol.docstring`) across all mappers
+- Export flag (`FileSymbol.isExported`) across all mappers
+- Required imports (`FileMap.imports`) across all mappers
+- Skipped read recovery — detects reads cancelled by steering queue and re-issues them
+- JSONL session-aware maps for pi session files
+- Directory read handling with EISDIR error and `ls` fallback
+
+### Fixed
+
+- Symbol duplication in file map output
+- oxlint warnings and errors resolved across codebase
+
+### Changed
+
+- Test helpers refactored; `models.ts` renamed to `constants.ts`
+
+## [1.0.0] - 2026-02-09
+
+Initial release.
+
+### Added
+
+- Structural file maps for large files (>2,000 lines or >50 KB)
+- 14 language mappers: TypeScript, JavaScript, Python, Go, Rust, C, C++, SQL, JSON, JSONL, YAML, TOML, CSV, Markdown
+- Budget-aware formatting with progressive detail reduction (10 KB full → 100 KB truncated)
+- In-memory caching by file path and modification time
+- Fallback chain: language mapper → universal-ctags → grep
+- Custom `file-map` messages delivered after `tool_result` events
+- E2E test infrastructure via tmux
+- Demo assets from 10 major open-source projects

src/mappers/clojure.ts

+import { readFile, stat } from "node:fs/promises";
+/**
+ * Clojure mapper using tree-sitter for AST extraction.
+ *
+ * The tree-sitter-clojure grammar parses at the S-expression level:
+ * all forms are `list_lit` nodes. We identify def forms by matching
+ * the first `sym_lit` child against known Clojure special forms.
+ */
+import { createRequire } from "node:module";
+
+import type { FileMap, FileSymbol } from "../types.js";
+
+import { DetailLevel, SymbolKind } from "../enums.js";
+
+type SyntaxNode = import("tree-sitter").SyntaxNode;
+
+/**
+ * Def forms we recognize and how they map to symbol kinds.
+ */
+const DEF_FORMS: Record<
+  string,
+  { kind: SymbolKind; isPrivate?: boolean; hasChildren?: boolean }
+> = {
+  defn: { kind: SymbolKind.Function },
+  "defn-": { kind: SymbolKind.Function, isPrivate: true },
+  def: { kind: SymbolKind.Variable },
+  defonce: { kind: SymbolKind.Variable },
+  defmacro: { kind: SymbolKind.Function },
+  defmulti: { kind: SymbolKind.Function },
+  defmethod: { kind: SymbolKind.Method },
+  defprotocol: { kind: SymbolKind.Interface, hasChildren: true },
+  defrecord: { kind: SymbolKind.Class },
+  deftype: { kind: SymbolKind.Class },
+};
+
+// Lazy-loaded parser
+let parser: import("tree-sitter") | null = null;
+let parserInitialized = false;
+
+function ensureWritableTypeProperty(parserCtor: unknown): void {
+  const syntaxNode = (parserCtor as { SyntaxNode?: { prototype?: object } })
+    .SyntaxNode;
+  const proto = syntaxNode?.prototype;
+  if (!proto) {
+    return;
+  }
+  const desc = Object.getOwnPropertyDescriptor(proto, "type");
+  if (!desc || desc.set) {
+    return;
+  }
+  Object.defineProperty(proto, "type", { ...desc, set: () => {} });
+}
+
+function getParser(): import("tree-sitter") | null {
+  if (parserInitialized) {
+    return parser;
+  }
+
+  parserInitialized = true;
+
+  const isBun = typeof (globalThis as { Bun?: unknown }).Bun !== "undefined";
+  if (isBun) {
+    return null;
+  }
+
+  try {
+    const require = createRequire(import.meta.url);
+    const ParserCtor = require("tree-sitter") as typeof import("tree-sitter");
+    const Clojure =
+      require("tree-sitter-clojure") as import("tree-sitter").Language;
+    ensureWritableTypeProperty(ParserCtor);
+    parser = new ParserCtor();
+    parser.setLanguage(Clojure);
+    return parser;
+  } catch {
+    return null;
+  }
+}
+
+function getNodeText(node: SyntaxNode, source: string): string {
+  return source.slice(node.startIndex, node.endIndex);
+}
+
+/**
+ * Get the text of a sym_lit's sym_name child.
+ */
+function getSymName(node: SyntaxNode, source: string): string | null {
+  if (node.type !== "sym_lit") {
+    return null;
+  }
+  const nameNode = node.namedChildren.find((c) => c.type === "sym_name");
+  if (!nameNode) {
+    return null;
+  }
+  return getNodeText(nameNode, source);
+}
+
+/**
+ * Check if a sym_lit has ^:private metadata.
+ */
+function hasPrivateMeta(node: SyntaxNode): boolean {
+  return node.namedChildren.some(
+    (c) =>
+      c.type === "meta_lit" &&
+      c.namedChildren.some((v) => v.type === "kwd_lit" && v.text === ":private")
+  );
+}
+
+/**
+ * Extract the string content from a str_lit node (strips surrounding quotes).
+ */
+function extractString(node: SyntaxNode, source: string): string {
+  const text = getNodeText(node, source);
+  return text.slice(1, -1);
+}
+
+/**
+ * Extract the named values from a list_lit (skipping gaps/whitespace).
+ */
+function getValueChildren(node: SyntaxNode): SyntaxNode[] {
+  return node.namedChildren.filter(
+    (c) => c.type !== "comment" && c.type !== "dis_expr"
+  );
+}
+
+/**
+ * Extract protocol method signatures from a defprotocol body.
+ */
+function extractProtocolMethods(
+  children: SyntaxNode[],
+  source: string
+): FileSymbol[] {
+  const methods: FileSymbol[] = [];
+  for (const child of children) {
+    if (child.type !== "list_lit") {
+      continue;
+    }
+    const values = getValueChildren(child);
+    const [firstValue] = values;
+    if (!firstValue || firstValue.type !== "sym_lit") {
+      continue;
+    }
+    const name = getSymName(firstValue, source);
+    if (!name) {
+      continue;
+    }
+
+    const paramsNode = values.find((v) => v.type === "vec_lit");
+    const params = paramsNode ? getNodeText(paramsNode, source) : "";
+    const signature = params ? `(${name} ${params})` : `(${name})`;
+
+    const docNode = values.find(
+      (v, i) =>
+        v.type === "str_lit" && i > values.indexOf(paramsNode ?? firstValue)
+    );
+
+    methods.push({
+      name,
+      kind: SymbolKind.Method,
+      startLine: child.startPosition.row + 1,
+      endLine: child.endPosition.row + 1,
+      signature,
+      ...(docNode ? { docstring: extractString(docNode, source) } : {}),
+    });
+  }
+  return methods;
+}
+
+/**
+ * Extract a defmethod form into an ExtractedDef-like result.
+ */
+function extractDefmethod(
+  node: SyntaxNode,
+  values: SyntaxNode[],
+  source: string
+): FileSymbol | null {
+  const [, nameNode, dispatchNode] = values;
+  if (!nameNode || nameNode.type !== "sym_lit") {
+    return null;
+  }
+  const multiName = getSymName(nameNode, source);
+  if (!multiName) {
+    return null;
+  }
+
+  const dispatchVal = dispatchNode
+    ? getNodeText(dispatchNode, source)
+    : "unknown";
+
+  const paramsNode = values.find((v) => v.type === "vec_lit");
+  const params = paramsNode ? getNodeText(paramsNode, source) : "";
+  const signature = `(defmethod ${multiName} ${dispatchVal} ${params})`;
+
+  return {
+    name: `${multiName} ${dispatchVal}`,
+    kind: SymbolKind.Method,
+    startLine: node.startPosition.row + 1,
+    endLine: node.endPosition.row + 1,
+    signature,
+    isExported: true,
+  };
+}
+
+/**
+ * Extract docstring from values after the name.
+ *
+ * For function-like forms (defn, defmacro, defprotocol, defmulti):
+ *   docstring is the first str_lit that appears before any vec_lit or list_lit.
+ *
+ * For value forms (def, defonce):
+ *   a str_lit is a docstring only if another value follows it.
+ *   e.g. (def x "doc" 42) → docstring="doc", but (def x "val") → no docstring.
+ */
+function extractDocstring(
+  restValues: SyntaxNode[],
+  source: string,
+  isValueForm: boolean
+): string | undefined {
+  const firstStr = restValues.find((v) => v.type === "str_lit");
+  if (!firstStr) {
+    return undefined;
+  }
+  const firstVec = restValues.find((v) => v.type === "vec_lit");
+  const firstList = restValues.find((v) => v.type === "list_lit");
+
+  const strIdx = restValues.indexOf(firstStr);
+  const vecIdx = firstVec ? restValues.indexOf(firstVec) : Infinity;
+  const listIdx = firstList ? restValues.indexOf(firstList) : Infinity;
+
+  if (strIdx >= vecIdx || strIdx >= listIdx) {
+    return undefined;
+  }
+
+  // For def/defonce: the string is a docstring only if a value follows it
+  if (isValueForm) {
+    const hasValueAfter = restValues.some(
+      (v, i) => i > strIdx && v.type !== "comment" && v.type !== "dis_expr"
+    );
+    if (!hasValueAfter) {
+      return undefined;
+    }
+  }
+
+  return extractString(firstStr, source);
+}
+
+/**
+ * Build signature for function-like forms (defn, defn-, defmacro).
+ */
+function buildFnSignature(
+  formName: string,
+  name: string,
+  restValues: SyntaxNode[],
+  source: string
+): string {
+  const firstVec = restValues.find((v) => v.type === "vec_lit");
+  if (firstVec) {
+    return `(${formName} ${name} ${getNodeText(firstVec, source)})`;
+  }
+
+  // Multi-arity: look for list_lit children starting with vec_lit
+  const arities = restValues.filter(
+    (v) =>
+      v.type === "list_lit" &&
+      getValueChildren(v).some((c) => c.type === "vec_lit")
+  );
+  if (arities.length > 0) {
+    const arityStrs = arities.map((a) => {
+      const vec = getValueChildren(a).find((c) => c.type === "vec_lit");
+      return vec ? getNodeText(vec, source) : "[]";
+    });
+    return `(${formName} ${name} ${arityStrs.join(" ")})`;
+  }
+
+  return `(${formName} ${name})`;
+}
+
+/**
+ * Build signature for non-function def forms.
+ */
+function buildDefSignature(
+  formName: string,
+  name: string,
+  restValues: SyntaxNode[],
+  source: string
+): string {
+  if (formName === "defmulti") {
+    const dispatchNode = restValues.find((v) => v.type !== "str_lit");
+    const dispatch = dispatchNode ? getNodeText(dispatchNode, source) : "";
+    return dispatch ? `(defmulti ${name} ${dispatch})` : `(defmulti ${name})`;
+  }
+
+  if (formName === "defprotocol") {
+    return `(defprotocol ${name})`;
+  }
+
+  if (formName === "defrecord" || formName === "deftype") {
+    const firstVec = restValues.find((v) => v.type === "vec_lit");
+    return firstVec
+      ? `(${formName} ${name} ${getNodeText(firstVec, source)})`
+      : `(${formName} ${name})`;
+  }
+
+  // def, defonce
+  return `(${formName} ${name})`;
+}
+
+const FN_FORMS = new Set(["defn", "defn-", "defmacro"]);
+
+/**
+ * Try to extract a def form from a list_lit node.
+ */
+function extractDef(node: SyntaxNode, source: string): FileSymbol | null {
+  if (node.type !== "list_lit") {
+    return null;
+  }
+
+  const values = getValueChildren(node);
+  if (values.length < 2) {
+    return null;
+  }
+
+  const [formNode, nameNode] = values;
+  if (!formNode || formNode.type !== "sym_lit") {
+    return null;
+  }
+  const formName = getSymName(formNode, source);
+  if (!formName) {
+    return null;
+  }
+
+  const defInfo = DEF_FORMS[formName];
+  if (!defInfo) {
+    return null;
+  }
+
+  // defmethod has unique structure
+  if (formName === "defmethod") {
+    return extractDefmethod(node, values, source);
+  }
+
+  if (!nameNode || nameNode.type !== "sym_lit") {
+    return null;
+  }
+  const name = getSymName(nameNode, source);
+  if (!name) {
+    return null;
+  }
+
+  const isPrivate = defInfo.isPrivate === true || hasPrivateMeta(nameNode);
+  const modifiers: string[] = [];
+  if (isPrivate) {
+    modifiers.push("private");
+  }
+  if (formName === "defmacro") {
+    modifiers.push("macro");
+  }
+
+  const restValues = values.slice(2);
+  const isValueForm = formName === "def" || formName === "defonce";
+  const docstring = extractDocstring(restValues, source, isValueForm);
+
+  const signature = FN_FORMS.has(formName)
+    ? buildFnSignature(formName, name, restValues, source)
+    : buildDefSignature(formName, name, restValues, source);
+
+  // Only defprotocol has extractable method children.
+  // defrecord/deftype inline protocol methods are not yet extracted.
+  let children: FileSymbol[] | undefined;
+  if (defInfo.hasChildren && formName === "defprotocol") {
+    const extracted = extractProtocolMethods(restValues, source);
+    children = extracted.length > 0 ? extracted : undefined;
+  }
+
+  const symbol: FileSymbol = {
+    name,
+    kind: defInfo.kind,
+    startLine: node.startPosition.row + 1,
+    endLine: node.endPosition.row + 1,
+    isExported: !isPrivate,
+  };
+
+  if (signature) {
+    symbol.signature = signature;
+  }
+  if (docstring) {
+    symbol.docstring = docstring;
+  }
+  if (modifiers.length > 0) {
+    symbol.modifiers = modifiers;
+  }
+  if (children) {
+    symbol.children = children;
+  }
+
+  return symbol;
+}
+
+/**
+ * Extract ns form for namespace and imports.
+ */
+function extractNs(
+  node: SyntaxNode,
+  source: string
+): { namespace: string; imports: string[]; docstring?: string } | null {
+  if (node.type !== "list_lit") {
+    return null;
+  }
+
+  const values = getValueChildren(node);
+  if (values.length < 2) {
+    return null;
+  }
+
+  const [formNode, nameNode, docNode] = values;
+  if (!formNode || formNode.type !== "sym_lit") {
+    return null;
+  }
+  if (getSymName(formNode, source) !== "ns") {
+    return null;
+  }
+
+  if (!nameNode || nameNode.type !== "sym_lit") {
+    return null;
+  }
+  const namespace = getSymName(nameNode, source);
+  if (!namespace) {
+    return null;
+  }
+
+  // Docstring at index 2
+  let docstring: string | undefined;
+  if (docNode?.type === "str_lit") {
+    docstring = extractString(docNode, source);
+  }
+
+  // Extract :require and :import clauses
+  const imports: string[] = [];
+  for (const child of values) {
+    if (child.type !== "list_lit") {
+      continue;
+    }
+    const listValues = getValueChildren(child);
+    const [kwd] = listValues;
+    if (!kwd || kwd.type !== "kwd_lit") {
+      continue;
+    }
+    const kwdText = getNodeText(kwd, source);
+    if (kwdText === ":require" || kwdText === ":import") {
+      for (const spec of listValues.slice(1)) {
+        // Unwrap reader conditionals inside require/import:
+        // #?(:clj [clojure.java.io] :cljs [cljs.reader])
+        // → each platform's specs get added with a platform tag
+        if (
+          spec.type === "read_cond_lit" ||
+          spec.type === "splicing_read_cond_lit"
+        ) {
+          const rcChildren = getValueChildren(spec);
+          let platform: string | undefined;
+          for (const rc of rcChildren) {
+            if (rc.type === "kwd_lit") {
+              platform = getNodeText(rc, source);
+              continue;
+            }
+            if (platform) {
+              imports.push(`${getNodeText(rc, source)} ${platform}`);
+              platform = undefined;
+            }
+          }
+        } else {
+          imports.push(getNodeText(spec, source));
+        }
+      }
+    }
+  }
+
+  return { namespace, imports, docstring };
+}
+
+/**
+ * Extract def forms from a reader conditional (#? or #?@).
+ *
+ * Reader conditionals contain kwd_lit/form pairs like:
+ *   #?(:clj (defn foo [x] ...) :cljs (defn foo [x] ...))
+ *
+ * We extract defs from all platform branches, annotating each with
+ * its platform keyword as a modifier. When the same name appears in
+ * multiple branches, all variants are included — the map consumer
+ * sees the full picture.
+ */
+function extractReaderConditionalDefs(
+  node: SyntaxNode,
+  source: string
+): FileSymbol[] {
+  const results: FileSymbol[] = [];
+  const children = getValueChildren(node);
+
+  // Children alternate: kwd_lit, form, kwd_lit, form, ...
+  let currentPlatform: string | undefined;
+  for (const child of children) {
+    if (child.type === "kwd_lit") {
+      currentPlatform = getNodeText(child, source);
+      continue;
+    }
+
+    if (child.type === "list_lit") {
+      const def = extractDef(child, source);
+      if (def && currentPlatform) {
+        const platformName = currentPlatform.replace(/^:/, "");
+        const platformMod = `platform-${platformName}`;
+        def.modifiers = def.modifiers
+          ? [...def.modifiers, platformMod]
+          : [platformMod];
+        results.push(def);
+      }
+    }
+  }
+
+  return results;
+}
+
+/**
+ * Generate a file map for Clojure files using tree-sitter.
+ */
+export async function clojureMapper(
+  filePath: string,
+  signal?: AbortSignal
+): Promise<FileMap | null> {
+  try {
+    const p = getParser();
+    if (!p) {
+      return null;
+    }
+
+    const stats = await stat(filePath);
+    const totalBytes = stats.size;
+
+    const content = await readFile(filePath, "utf8");
+
+    if (signal?.aborted) {
+      return null;
+    }
+
+    let tree: import("tree-sitter").Tree;
+    try {
+      tree = p.parse(content);
+    } catch {
+      return null;
+    }
+
+    const symbols: FileSymbol[] = [];
+    const imports: string[] = [];
+
+    for (const child of tree.rootNode.namedChildren) {
+      // Reader conditionals: extract defs from all platform branches
+      if (
+        child.type === "read_cond_lit" ||
+        child.type === "splicing_read_cond_lit"
+      ) {
+        const condDefs = extractReaderConditionalDefs(child, content);
+        symbols.push(...condDefs);
+        continue;
+      }
+
+      if (child.type !== "list_lit") {
+        continue;
+      }
+
+      const ns = extractNs(child, content);
+      if (ns) {
+        imports.push(...ns.imports);
+
+        symbols.push({
+          name: ns.namespace,
+          kind: SymbolKind.Namespace,
+          startLine: child.startPosition.row + 1,
+          endLine: child.endPosition.row + 1,
+          signature: `(ns ${ns.namespace})`,
+          ...(ns.docstring ? { docstring: ns.docstring } : {}),
+          isExported: true,
+        });
+        continue;
+      }
+
+      const def = extractDef(child, content);
+      if (def) {
+        symbols.push(def);
+      }
+    }
+
+    if (symbols.length === 0) {
+      return null;
+    }
+
+    const totalLines = content.split("\n").length;
+
+    return {
+      path: filePath,
+      totalLines,
+      totalBytes,
+      language: "Clojure",
+      symbols,
+      imports,
+      detailLevel: DetailLevel.Full,
+    };
+  } catch (error) {
+    if (signal?.aborted) {
+      return null;
+    }
+    console.error(`Clojure mapper failed: ${error}`);
+    return null;
+  }
+}

AGENTS.md

 # AGENTS.md
 
-> Last updated: 2026-02-10
+> Last updated: 2026-02-14
 
 Pi extension that augments the built-in `read` tool with structural file maps for large files (>2,000 lines or >50 KB). Intercepts `read` calls, generates symbol maps via language-specific parsers, and sends them as separate `file-map` messages after the tool result.
 
-## Commands (verified 2026-02-10)
+## Commands (verified 2026-02-14)
 
@@ -33,6 +33,7 @@ | Command | Purpose | ~Time |
 ├── constants.ts          → THRESHOLDS: lines, bytes, budget tiers
-└── mappers/              → One mapper per language (16 total)
+└── mappers/              → One mapper per language (17 total)
     ├── typescript.ts     → ts-morph (handles TS + JS)
     ├── rust.ts           → tree-sitter-rust
     ├── cpp.ts            → tree-sitter-cpp (C++ and .h files)
+    ├── clojure.ts        → tree-sitter-clojure (.clj, .cljs, .cljc, .edn)
     ├── python.ts         → subprocess: scripts/python_outline.py
@@ -62,3 +63,3 @@ ├── go.ts             → subprocess: scripts/go_outline.go
 ├── benchmarks/           → Mapper performance benchmarks
-└── helpers/              → Test utilities (pi-runner, constants, models)
+└── helpers/              → Test utilities (pi-runner, constants, tree-sitter)
 
@@ -91,2 +92,3 @@ docs/
 | Complex mapper | `src/mappers/typescript.ts` | ts-morph AST walk, nested symbols, modifiers |
+| Tree-sitter mapper | `src/mappers/clojure.ts` | tree-sitter AST walk, reader conditionals, platform modifiers |
 | Subprocess mapper | `src/mappers/python.ts` | Calls external script, parses JSON output |
@@ -131,4 +133,5 @@ | Unit test | `tests/unit/mappers/csv.test.ts` | Fixture-based, edge cases, null returns |
 - `tree-sitter` pinned to 0.22.4 due to peer dependency conflicts (see `docs/todo/upgrade-tree-sitter-0.26.md`)
+- `tree-sitter-clojure` pinned to commit SHA from `github:ghoseb/tree-sitter-clojure` (third-party fork)
 - Go outline script auto-compiles on first use; compiled binary checked in at `scripts/go_outline`
-- Phase 1-4 of implementation plan complete; remaining TODOs in `docs/todo/`
+- Phase 1-5 of implementation plan complete; remaining TODOs in `docs/todo/`
 
@@ -158,3 +161,3 @@ | Docstrings / JSDoc | `FileSymbol.docstring?: string` | First-line summary of doc comments |
 - **Linting:** oxlint + oxfmt
-- **Parsing:** ts-morph, tree-sitter, regex, subprocess (Python/Go/jq)
+- **Parsing:** ts-morph, tree-sitter (rust, cpp, clojure), regex, subprocess (Python/Go/jq)
 - **Framework:** pi extension API (`@mariozechner/pi-coding-agent`)

package.json

 {
   "name": "pi-read-map",
-  "version": "1.1.0",
+  "version": "1.2.0",
   "description": "Pi extension that adds structural file maps for large files",
@@ -47,2 +47,3 @@ "type": "module",
     "tree-sitter": "0.22.4",
+    "tree-sitter-clojure": "github:ghoseb/tree-sitter-clojure#78928e6",
     "tree-sitter-cpp": "0.23.4",
@@ -49,0 +50,0 @@ "tree-sitter-rust": "0.23.3",

README.md

@@ -26,3 +26,3 @@ # pi-read-map
 - **Generates structural maps** showing symbols, classes, functions, and their exact line ranges
-- **Supports 16 languages** through specialized parsers: TypeScript, JavaScript, Python, Go, Rust, C, C++, SQL, JSON, JSONL, YAML, TOML, CSV, Markdown
+- **Supports 17 languages** through specialized parsers: TypeScript, JavaScript, Python, Go, Rust, C, C++, Clojure, ClojureScript, SQL, JSON, JSONL, YAML, TOML, CSV, Markdown, EDN
 - **Extracts structural outlines** — functions, classes, and their line ranges — typically under 1% of file size
@@ -134,2 +134,3 @@ - **Enforces budgets** through progressive detail reduction (10 KB full → 15 KB compact → 20 KB minimal → 50 KB outline → 100 KB hard cap)
     ├── cpp.ts            # tree-sitter for C/C++
+    ├── clojure.ts        # tree-sitter for Clojure/ClojureScript/EDN
     ├── c.ts              # Regex patterns
@@ -178,2 +179,3 @@ ├── sql.ts            # Regex
 - `tree-sitter-rust` - Rust parsing
+- `tree-sitter-clojure` - Clojure parsing
 
@@ -190,4 +192,8 @@ **System tools (optional):**
 
+### Contributors
+
+- [Baishampayan Ghose](https://github.com/ghoseb) — Clojure tree-sitter mapper and [tree-sitter-clojure](https://github.com/ghoseb/tree-sitter-clojure) grammar
+
 ## License
 
 MIT

src/language-detect.ts

@@ -41,2 +41,8 @@ import { extname } from "node:path";
 
+  // Clojure
+  ".clj": { id: "clojure", name: "Clojure" },
+  ".cljs": { id: "clojure", name: "ClojureScript" },
+  ".cljc": { id: "clojure", name: "Clojure" },
+  ".edn": { id: "clojure", name: "EDN" },
+
   // SQL
@@ -43,0 +49,0 @@ ".sql": { id: "sql", name: "SQL" },

src/mapper.ts

@@ -6,2 +6,3 @@ import type { FileMap, MapOptions } from "./types.js";
 import { cMapper } from "./mappers/c.js";
+import { clojureMapper } from "./mappers/clojure.js";
 import { cppMapper } from "./mappers/cpp.js";
@@ -61,2 +62,5 @@ import { csvMapper } from "./mappers/csv.js";
   csv: csvMapper,
+
+  // Phase 5: Clojure tree-sitter
+  clojure: clojureMapper,
 };
@@ -63,0 +67,0 @@

src/mappers/ctags.ts

@@ -22,6 +22,8 @@ /**
 /**
- * Map ctags kind letters to SymbolKind.
+ * Map ctags kind identifiers to SymbolKind.
+ * Handles both single-letter kinds (legacy format) and full-word kinds (JSON format).
  * See: https://docs.ctags.io/en/latest/man/ctags.1.html
  */
 const CTAGS_KIND_MAP: Record<string, SymbolKind> = {
+  // Single-letter kinds (legacy/traditional format)
   c: SymbolKind.Class,
@@ -48,3 +50,3 @@ d: SymbolKind.Constant, // macro definition
   T: SymbolKind.Type, // type
-  // Language-specific mappings
+  // Language-specific single-letter mappings
   a: SymbolKind.Type, // alias
@@ -58,2 +60,26 @@ b: SymbolKind.Variable, // block (Ruby)
   z: SymbolKind.Property, // parameter
+
+  // Full-word kinds (JSON output format)
+  class: SymbolKind.Class,
+  enum: SymbolKind.Enum,
+  enumerator: SymbolKind.Enum,
+  function: SymbolKind.Function,
+  interface: SymbolKind.Interface,
+  macro: SymbolKind.Constant,
+  member: SymbolKind.Property,
+  method: SymbolKind.Method,
+  module: SymbolKind.Module,
+  namespace: SymbolKind.Namespace,
+  package: SymbolKind.Module,
+  property: SymbolKind.Property,
+  struct: SymbolKind.Struct,
+  type: SymbolKind.Type,
+  typedef: SymbolKind.Type,
+  union: SymbolKind.Type,
+  variable: SymbolKind.Variable,
+  field: SymbolKind.Property,
+  constant: SymbolKind.Constant,
+  prototype: SymbolKind.Function,
+  alias: SymbolKind.Type,
+  trait: SymbolKind.Interface,
 };
@@ -155,5 +181,6 @@
 
-    // Run ctags with JSON output
+    // Run ctags with JSON output and line numbers
     // --output-format=json requires Universal Ctags 5.9+
-    const cmd = `ctags --output-format=json -f - "${filePath}" 2>/dev/null`;
+    // --fields=+n ensures line numbers are included in JSON output
+    const cmd = `ctags --output-format=json --fields=+n -f - "${filePath}" 2>/dev/null`;
 
@@ -160,0 +187,0 @@ let stdout: string;

New alerts

GitHub dependency

Supply chain risk

Contains a dependency which resolves to a GitHub URL. Dependencies fetched from GitHub specifiers are not immutable can be used to inject untrusted code or reduce the likelihood of a reproducible install.

Filesystem access

Supply chain risk

Accesses the file system, and could potentially read sensitive data.

Improved metrics

Total package byte prevSize

1584307

1.3%

Number of package files

36

5.88%

Lines of code

6783

9.23%

Number of lines in readme file

196

3.16%

Worsened metrics

Dependency count

5

25%

Number of high supply chain risk alerts

1

Infinity%

Number of low supply chain risk alerts

23

9.52%

Dependency changes

• + Added

  tree-sitter-clojure@github:ghoseb/tree-sitter-clojure#78928e6