flavor-code 1.2.13 → 1.2.15

package/dist/astgraph/indexer.mjs

@@ -1,175 +1,175 @@
-// astgraph indexer — full and incremental project indexing.
-// Walks the workspace, extracts symbols per file (content-hash incremental),
-// then resolves cross-file references into graph edges.
-
-import { createHash } from "node:crypto";
-import { readdirSync, readFileSync, statSync } from "node:fs";
-import { relative, sep } from "node:path";
-
-import {
-  deleteEdgesForFile,
-  deleteFileRecord,
-  deleteUnresolvedRefsForFile,
-  getFileRecord,
-  insertEdge,
-  insertUnresolvedRef,
-  replaceNodes,
-  setMetadata,
-  upsertFileRecord,
-} from "./db.mjs";
-import { extract } from "./extract.mjs";
-import { languageForFile, SUPPORTED_LANGUAGES } from "./grammars.mjs";
-import { resolveImportPath, resolveRefs } from "./resolve.mjs";
-
-/** Directory names never indexed, regardless of depth. */
-const EXCLUDED_DIRS = new Set([
-  "node_modules", "dist", "build", "out", "coverage", "release", ".git",
-  ".flavor", ".claude", ".worktrees", ".cache", ".idea", ".vscode", "vendor",
-]);
-
-function toPosix(path) {
-  return path.split(sep).join("/");
-}
-
-/** Walk the workspace and return code files as workspace-relative posix paths. */
-export function listCodeFiles(workspace) {
-  const files = [];
-  const walk = (dir) => {
-    let entries;
-    try { entries = readdirSync(dir, { withFileTypes: true }); }
-    catch { return; }
-    for (const entry of entries) {
-      const full = `${dir}/${entry.name}`;
-      if (entry.isDirectory()) {
-        if (!EXCLUDED_DIRS.has(entry.name)) walk(full);
-      } else if (entry.isFile() && languageForFile(entry.name) !== undefined) {
-        files.push(toPosix(relative(workspace, full)));
-      }
-    }
-  };
-  walk(workspace);
-  return files.sort();
-}
-
-/** SHA-256 content hash used for incremental detection. */
-export function contentHash(content) {
-  return createHash("sha256").update(content).digest("hex");
-}
-
-async function indexOneFile(db, workspace, relativePath) {
-  const content = readFileSync(`${workspace}/${relativePath}`, "utf8");
-  const language = languageForFile(relativePath);
-  const { nodes, edges, refs } = await extract(content, relativePath, language);
-
-  deleteEdgesForFile(db, relativePath);
-  deleteUnresolvedRefsForFile(db, relativePath);
-  replaceNodes(db, relativePath, nodes);
-  for (const edge of edges) insertEdge(db, edge);
-  for (const ref of refs) {
-    insertUnresolvedRef(db, {
-      fromNodeId: ref.fromNodeId, referenceName: ref.referenceName,
-      referenceKind: ref.referenceKind, line: ref.line, col: ref.col,
-      filePath: ref.filePath, moduleSpecifier: ref.moduleSpecifier,
-    });
-  }
-  let size;
-  try { size = statSync(`${workspace}/${relativePath}`).size; } catch { size = content.length; }
-  upsertFileRecord(db, {
-    path: relativePath, contentHash: contentHash(content), language, size, nodeCount: nodes.length,
-  });
-  return { nodeCount: nodes.length };
-}
-
-/** Resolve every pending unresolved_ref whose module specifier points at an indexed file. */
-function resolvePass(db, workspace) {
-  const rows = db.prepare(`
-    SELECT id, from_node_id, reference_name, reference_kind, line, col, file_path, module_specifier
-    FROM unresolved_refs WHERE status = 'pending' AND module_specifier IS NOT NULL
-  `).all();
-  if (rows.length === 0) return 0;
-
-  const nodesByFile = new Map();
-  for (const file of db.prepare("SELECT DISTINCT file_path FROM nodes").all()) {
-    nodesByFile.set(file.file_path, db.prepare(
-      "SELECT id, name, is_exported FROM nodes WHERE file_path = ?",
-    ).all(file.file_path).map((row) => ({ id: row.id, name: row.name, isExported: row.is_exported === 1 })));
-  }
-  const exists = (path) => {
-    try { statSync(`${workspace}/${path}`); return true; } catch { return false; }
-  };
-
-  const refs = rows.map((row) => ({
-    id: row.id, fromNodeId: row.from_node_id, referenceName: row.reference_name,
-    referenceKind: row.reference_kind, line: row.line, col: row.col,
-    filePath: row.file_path, moduleSpecifier: row.module_specifier,
-  }));
-  // Import statements live outside any function, so their refs carry no from-node.
-  // Attribute them to a node that actually uses the binding (same file + name);
-  // unused imports resolve silently without an edge.
-  for (const ref of refs) {
-    if (ref.fromNodeId !== null || ref.referenceKind !== "import") continue;
-    const user = refs.find((other) => other.referenceKind !== "import"
-      && other.referenceName === ref.referenceName
-      && other.filePath === ref.filePath
-      && other.fromNodeId !== null);
-    ref.fromNodeId = user?.fromNodeId ?? null;
-  }
-  const { edges, resolvedIds } = resolveRefs(refs.filter((ref) => ref.fromNodeId !== null), nodesByFile, exists);
-  for (const edge of edges) insertEdge(db, edge);
-  if (resolvedIds.length > 0) {
-    db.exec("BEGIN");
-    try {
-      const mark = db.prepare("DELETE FROM unresolved_refs WHERE id = ?");
-      for (const id of resolvedIds) mark.run(id);
-      db.exec("COMMIT");
-    } catch (error) { db.exec("ROLLBACK"); throw error; }
-  }
-  return resolvedIds.length;
-}
-
-/**
- * Index (or re-index) a workspace into the graph database.
- *
- * Options:
- *   onlyPaths — restrict the pass to these workspace-relative paths (hook-driven sync).
- *               Disappeared files are still pruned; everything else is left untouched.
- * Returns { files, indexed, skipped, removed, resolved }.
- */
-export async function indexProject(workspace, { db, onlyPaths } = {}) {
-  const restrict = onlyPaths === undefined ? undefined : new Set(onlyPaths.map(toPosix));
-  const diskFiles = listCodeFiles(workspace).filter((path) => restrict === undefined || restrict.has(path));
-
-  let indexed = 0;
-  let skipped = 0;
-  for (const path of diskFiles) {
-    const current = readFileSync(`${workspace}/${path}`, "utf8");
-    const hash = contentHash(current);
-    const record = getFileRecord(db, path);
-    if (record !== undefined && record.contentHash === hash) { skipped += 1; continue; }
-    await indexOneFile(db, workspace, path);
-    indexed += 1;
-  }
-
-  // Prune files that no longer exist on disk.
-  let removed = 0;
-  const known = db.prepare("SELECT path FROM files").all();
-  const diskSet = new Set(diskFiles);
-  for (const row of known) {
-    if (restrict !== undefined && !restrict.has(row.path)) continue;
-    if (!diskSet.has(row.path)) {
-      deleteEdgesForFile(db, row.path);
-      deleteUnresolvedRefsForFile(db, row.path);
-      replaceNodes(db, row.path, []);
-      deleteFileRecord(db, row.path);
-      removed += 1;
-    }
-  }
-
-  const resolved = resolvePass(db, workspace);
-  setMetadata(db, "last_index", JSON.stringify({ at: Date.now(), files: diskFiles.length, indexed, removed }));
-
-  return { files: diskFiles.length, indexed, skipped, removed, resolved };
-}
-
-/** Exposed for diagnostics. */
-export const INDEXER_LANGUAGES = SUPPORTED_LANGUAGES;
+// astgraph indexer — full and incremental project indexing.
+// Walks the workspace, extracts symbols per file (content-hash incremental),
+// then resolves cross-file references into graph edges.
+
+import { createHash } from "node:crypto";
+import { readdirSync, readFileSync, statSync } from "node:fs";
+import { relative, sep } from "node:path";
+
+import {
+  deleteEdgesForFile,
+  deleteFileRecord,
+  deleteUnresolvedRefsForFile,
+  getFileRecord,
+  insertEdge,
+  insertUnresolvedRef,
+  replaceNodes,
+  setMetadata,
+  upsertFileRecord,
+} from "./db.mjs";
+import { extract } from "./extract.mjs";
+import { languageForFile, SUPPORTED_LANGUAGES } from "./grammars.mjs";
+import { resolveImportPath, resolveRefs } from "./resolve.mjs";
+
+/** Directory names never indexed, regardless of depth. */
+const EXCLUDED_DIRS = new Set([
+  "node_modules", "dist", "build", "out", "coverage", "release", ".git",
+  ".flavor", ".claude", ".worktrees", ".cache", ".idea", ".vscode", "vendor",
+]);
+
+function toPosix(path) {
+  return path.split(sep).join("/");
+}
+
+/** Walk the workspace and return code files as workspace-relative posix paths. */
+export function listCodeFiles(workspace) {
+  const files = [];
+  const walk = (dir) => {
+    let entries;
+    try { entries = readdirSync(dir, { withFileTypes: true }); }
+    catch { return; }
+    for (const entry of entries) {
+      const full = `${dir}/${entry.name}`;
+      if (entry.isDirectory()) {
+        if (!EXCLUDED_DIRS.has(entry.name)) walk(full);
+      } else if (entry.isFile() && languageForFile(entry.name) !== undefined) {
+        files.push(toPosix(relative(workspace, full)));
+      }
+    }
+  };
+  walk(workspace);
+  return files.sort();
+}
+
+/** SHA-256 content hash used for incremental detection. */
+export function contentHash(content) {
+  return createHash("sha256").update(content).digest("hex");
+}
+
+async function indexOneFile(db, workspace, relativePath) {
+  const content = readFileSync(`${workspace}/${relativePath}`, "utf8");
+  const language = languageForFile(relativePath);
+  const { nodes, edges, refs } = await extract(content, relativePath, language);
+
+  deleteEdgesForFile(db, relativePath);
+  deleteUnresolvedRefsForFile(db, relativePath);
+  replaceNodes(db, relativePath, nodes);
+  for (const edge of edges) insertEdge(db, edge);
+  for (const ref of refs) {
+    insertUnresolvedRef(db, {
+      fromNodeId: ref.fromNodeId, referenceName: ref.referenceName,
+      referenceKind: ref.referenceKind, line: ref.line, col: ref.col,
+      filePath: ref.filePath, moduleSpecifier: ref.moduleSpecifier,
+    });
+  }
+  let size;
+  try { size = statSync(`${workspace}/${relativePath}`).size; } catch { size = content.length; }
+  upsertFileRecord(db, {
+    path: relativePath, contentHash: contentHash(content), language, size, nodeCount: nodes.length,
+  });
+  return { nodeCount: nodes.length };
+}
+
+/** Resolve every pending unresolved_ref whose module specifier points at an indexed file. */
+function resolvePass(db, workspace) {
+  const rows = db.prepare(`
+    SELECT id, from_node_id, reference_name, reference_kind, line, col, file_path, module_specifier
+    FROM unresolved_refs WHERE status = 'pending' AND module_specifier IS NOT NULL
+  `).all();
+  if (rows.length === 0) return 0;
+
+  const nodesByFile = new Map();
+  for (const file of db.prepare("SELECT DISTINCT file_path FROM nodes").all()) {
+    nodesByFile.set(file.file_path, db.prepare(
+      "SELECT id, name, is_exported FROM nodes WHERE file_path = ?",
+    ).all(file.file_path).map((row) => ({ id: row.id, name: row.name, isExported: row.is_exported === 1 })));
+  }
+  const exists = (path) => {
+    try { statSync(`${workspace}/${path}`); return true; } catch { return false; }
+  };
+
+  const refs = rows.map((row) => ({
+    id: row.id, fromNodeId: row.from_node_id, referenceName: row.reference_name,
+    referenceKind: row.reference_kind, line: row.line, col: row.col,
+    filePath: row.file_path, moduleSpecifier: row.module_specifier,
+  }));
+  // Import statements live outside any function, so their refs carry no from-node.
+  // Attribute them to a node that actually uses the binding (same file + name);
+  // unused imports resolve silently without an edge.
+  for (const ref of refs) {
+    if (ref.fromNodeId !== null || ref.referenceKind !== "import") continue;
+    const user = refs.find((other) => other.referenceKind !== "import"
+      && other.referenceName === ref.referenceName
+      && other.filePath === ref.filePath
+      && other.fromNodeId !== null);
+    ref.fromNodeId = user?.fromNodeId ?? null;
+  }
+  const { edges, resolvedIds } = resolveRefs(refs.filter((ref) => ref.fromNodeId !== null), nodesByFile, exists);
+  for (const edge of edges) insertEdge(db, edge);
+  if (resolvedIds.length > 0) {
+    db.exec("BEGIN");
+    try {
+      const mark = db.prepare("DELETE FROM unresolved_refs WHERE id = ?");
+      for (const id of resolvedIds) mark.run(id);
+      db.exec("COMMIT");
+    } catch (error) { db.exec("ROLLBACK"); throw error; }
+  }
+  return resolvedIds.length;
+}
+
+/**
+ * Index (or re-index) a workspace into the graph database.
+ *
+ * Options:
+ *   onlyPaths — restrict the pass to these workspace-relative paths (hook-driven sync).
+ *               Disappeared files are still pruned; everything else is left untouched.
+ * Returns { files, indexed, skipped, removed, resolved }.
+ */
+export async function indexProject(workspace, { db, onlyPaths } = {}) {
+  const restrict = onlyPaths === undefined ? undefined : new Set(onlyPaths.map(toPosix));
+  const diskFiles = listCodeFiles(workspace).filter((path) => restrict === undefined || restrict.has(path));
+
+  let indexed = 0;
+  let skipped = 0;
+  for (const path of diskFiles) {
+    const current = readFileSync(`${workspace}/${path}`, "utf8");
+    const hash = contentHash(current);
+    const record = getFileRecord(db, path);
+    if (record !== undefined && record.contentHash === hash) { skipped += 1; continue; }
+    await indexOneFile(db, workspace, path);
+    indexed += 1;
+  }
+
+  // Prune files that no longer exist on disk.
+  let removed = 0;
+  const known = db.prepare("SELECT path FROM files").all();
+  const diskSet = new Set(diskFiles);
+  for (const row of known) {
+    if (restrict !== undefined && !restrict.has(row.path)) continue;
+    if (!diskSet.has(row.path)) {
+      deleteEdgesForFile(db, row.path);
+      deleteUnresolvedRefsForFile(db, row.path);
+      replaceNodes(db, row.path, []);
+      deleteFileRecord(db, row.path);
+      removed += 1;
+    }
+  }
+
+  const resolved = resolvePass(db, workspace);
+  setMetadata(db, "last_index", JSON.stringify({ at: Date.now(), files: diskFiles.length, indexed, removed }));
+
+  return { files: diskFiles.length, indexed, skipped, removed, resolved };
+}
+
+/** Exposed for diagnostics. */
+export const INDEXER_LANGUAGES = SUPPORTED_LANGUAGES;