onadit 0.0.1 → 0.1.0-beta.45

package/dist/src/ingestion/cursor/workspace-root.js

@@ -0,0 +1,187 @@
+/**
+ * Resolve a native Cursor session's `workspaceIdentifier` to a filesystem
+ * project root — the missing half of Cursor project-root attribution.
+ *
+ * DISCOVERY SLICE ONLY. Read-only, no LanceDB writes, no daemon, no capture.
+ * The pure resolver ({@link resolveProjectRoot}) takes an injected
+ * `readWorkspaceJson` so it is exhaustively unit-testable from synthetic
+ * metadata with no filesystem and no real Cursor install. The fs-backed wrapper
+ * ({@link resolveCursorProjectRoot}) only reads `workspace.json` (a tiny
+ * `{ "folder": "file://…" }` file) — never the workspace's `state.vscdb`,
+ * never tokens, never message content.
+ *
+ * ## The join contract (empirically verified, macOS Cursor "glass", 2026-06)
+ *
+ * On opening a folder window, Cursor writes the composer header's
+ * `workspaceIdentifier` with TWO independent links to the project root:
+ *
+ *   "workspaceIdentifier": {
+ *     "id": "<workspaceStorageHash>",          // e.g. "11ed9332b5bd79fd69ad93a7636abd28"
+ *     "uri": { "scheme": "file",
+ *              "fsPath": "/Users/me/projects/foo",
+ *              "path":   "/Users/me/projects/foo",
+ *              "external": "file:///Users/me/projects/foo" }
+ *   }
+ *
+ *   1. INLINE (primary, current builds): `uri.fsPath` IS the project root.
+ *      No disk lookup needed.
+ *   2. ON-DISK (authoritative fallback): `id` equals the directory name under
+ *      `~/Library/Application Support/Cursor/User/workspaceStorage/<id>/`,
+ *      whose `workspace.json` holds `{ "folder": "file://<root>" }`. Verified:
+ *      the same hash appears as both `workspaceIdentifier.id` and the storage
+ *      dir name, and its `workspace.json.folder` matches `uri.fsPath`.
+ *
+ * The storage hash is NOT a recomputable md5 of the folder URI (checked: no
+ * simple encoding reproduces it) — so resolution MUST read `workspace.json` (or
+ * trust the inline uri), never re-derive the hash.
+ *
+ * ## Failure modes (all first-class — never guessed)
+ *   - `id === "empty-window"` and no `uri`  → no folder open. NOT attributable.
+ *     This is the common case on a machine with only empty windows.
+ *   - `workspace.json` has `"workspace": "…code-workspace"` (multi-root) instead
+ *     of `"folder"` → project root is ambiguous. Do not pick one.
+ *   - `uri.scheme` / `folder` scheme is not `file` (ssh-remote, wsl, devcontainer)
+ *     → not a local path.
+ *   - `id` present but `workspace.json` missing/purged and no inline `uri`
+ *     → unresolved.
+ *
+ * The capture slice should scope to a project root ONLY on `resolved`; for every
+ * other status, capture WITHOUT project scoping (or behind a user-confirmed
+ * fallback). See docs/cursor-capture-discovery.md.
+ */
+import fs from "node:fs";
+import path from "node:path";
+/** Sentinel `workspaceIdentifier.id` Cursor uses when no folder is open. */
+export const EMPTY_WINDOW_ID = "empty-window";
+/** Parse a URI string into `{ scheme, path }`; `path` is null for non-file. */
+function parseUriString(uri) {
+    let parsed;
+    try {
+        parsed = new URL(uri);
+    }
+    catch {
+        return { scheme: null, path: null };
+    }
+    const scheme = parsed.protocol.replace(/:$/, "");
+    if (scheme !== "file")
+        return { scheme, path: null };
+    // file:///abs/path -> decode percent-encoding (spaces, accents in client dirs)
+    return { scheme, path: decodeURIComponent(parsed.pathname) };
+}
+/** Derive `{ scheme, root }` from an inline header `uri` object. */
+function rootFromInlineUri(uri) {
+    if (!uri || typeof uri !== "object")
+        return { scheme: null, root: null };
+    const scheme = typeof uri.scheme === "string" ? uri.scheme : null;
+    if (scheme && scheme !== "file")
+        return { scheme, root: null };
+    // fsPath / path are already decoded absolute paths in Cursor's serialization.
+    const direct = (typeof uri.fsPath === "string" && uri.fsPath) ||
+        (typeof uri.path === "string" && uri.path) ||
+        null;
+    if (direct)
+        return { scheme: scheme ?? "file", root: direct };
+    if (typeof uri.external === "string") {
+        const ext = parseUriString(uri.external);
+        return { scheme: ext.scheme ?? scheme, root: ext.path };
+    }
+    return { scheme, root: null };
+}
+/**
+ * Parse a `workspace.json` file's text. Returns the folder root (single-folder
+ * window), a multi-root marker (`.code-workspace`), or unknown. Pure.
+ */
+export function parseWorkspaceJson(text) {
+    let obj;
+    try {
+        obj = JSON.parse(text);
+    }
+    catch {
+        return { kind: "unknown" };
+    }
+    if (!obj || typeof obj !== "object")
+        return { kind: "unknown" };
+    const rec = obj;
+    if (typeof rec.folder === "string") {
+        const { scheme, path: p } = parseUriString(rec.folder);
+        return { kind: "folder", scheme, root: p };
+    }
+    if (typeof rec.workspace === "string") {
+        const { path: p } = parseUriString(rec.workspace);
+        return { kind: "multi_root", configPath: p };
+    }
+    return { kind: "unknown" };
+}
+/**
+ * Resolve a composer header's `workspaceIdentifier` to a project root. Pure:
+ * the only I/O is the injected {@link ResolveOptions.readWorkspaceJson}. Tries
+ * the inline header uri first (current builds), then the on-disk
+ * `workspace.json` join (legacy builds / uri stripped). Never guesses — every
+ * non-resolvable case maps to an explicit status.
+ */
+export function resolveProjectRoot(identifier, options = {}) {
+    const id = typeof identifier?.id === "string" ? identifier.id : null;
+    // 1. Sentinel: no folder open. Common, expected, not attributable.
+    if (id === EMPTY_WINDOW_ID) {
+        return { status: "empty_window", workspaceId: EMPTY_WINDOW_ID };
+    }
+    // 2. Inline uri (Cursor embeds the folder directly on the header).
+    const inline = rootFromInlineUri(identifier?.uri);
+    if (inline.scheme && inline.scheme !== "file") {
+        return { status: "remote", workspaceId: id, scheme: inline.scheme };
+    }
+    if (inline.root) {
+        return { status: "resolved", root: inline.root, source: "header_uri", workspaceId: id };
+    }
+    // 3. On-disk fallback: join id -> workspaceStorage/<id>/workspace.json.
+    if (id && options.readWorkspaceJson) {
+        const text = options.readWorkspaceJson(id);
+        if (text) {
+            const parsed = parseWorkspaceJson(text);
+            if (parsed.kind === "folder") {
+                if (parsed.scheme && parsed.scheme !== "file") {
+                    return { status: "remote", workspaceId: id, scheme: parsed.scheme };
+                }
+                if (parsed.root) {
+                    return { status: "resolved", root: parsed.root, source: "workspace_json", workspaceId: id };
+                }
+            }
+            else if (parsed.kind === "multi_root") {
+                return { status: "multi_root", workspaceId: id, configPath: parsed.configPath };
+            }
+        }
+    }
+    // 4. Nothing reliable — do not guess.
+    return { status: "unresolved", workspaceId: id };
+}
+/** Path to a workspace's `workspace.json` under a Cursor app-support dir. */
+export function workspaceJsonPath(cursorDir, workspaceId) {
+    return path.join(cursorDir, "User", "workspaceStorage", workspaceId, "workspace.json");
+}
+/** Read `workspace.json` text read-only; null if it does not exist / unreadable. */
+export function readWorkspaceJsonFromDisk(cursorDir, workspaceId) {
+    try {
+        return fs.readFileSync(workspaceJsonPath(cursorDir, workspaceId), "utf-8");
+    }
+    catch {
+        return null;
+    }
+}
+/**
+ * Filesystem-backed resolution: resolve a header's identifier against a real
+ * Cursor app-support dir, reading `workspace.json` read-only when needed.
+ */
+export function resolveCursorProjectRoot(identifier, cursorDir) {
+    return resolveProjectRoot(identifier, {
+        readWorkspaceJson: (workspaceId) => readWorkspaceJsonFromDisk(cursorDir, workspaceId),
+    });
+}
+/**
+ * Resolve a project root from a workspace id alone (no inline uri available).
+ * Used by the probe to upgrade `unresolved` discovery candidates on legacy
+ * builds via the on-disk `workspace.json` join.
+ */
+export function resolveProjectRootByWorkspaceId(cursorDir, workspaceId) {
+    return resolveCursorProjectRoot({ id: workspaceId }, cursorDir);
+}
+//# sourceMappingURL=workspace-root.js.map

package/dist/src/ingestion/cursor/workspace-root.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"workspace-root.js","sourceRoot":"","sources":["../../../../src/ingestion/cursor/workspace-root.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GAkDG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAE7B,4EAA4E;AAC5E,MAAM,CAAC,MAAM,eAAe,GAAG,cAAc,CAAC;AAuC9C,+EAA+E;AAC/E,SAAS,cAAc,CAAC,GAAW;IACjC,IAAI,MAAW,CAAC;IAChB,IAAI,CAAC;QACH,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;IACxB,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IACtC,CAAC;IACD,MAAM,MAAM,GAAG,MAAM,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACjD,IAAI,MAAM,KAAK,MAAM;QAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IACrD,+EAA+E;IAC/E,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,kBAAkB,CAAC,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC;AAC/D,CAAC;AAED,oEAAoE;AACpE,SAAS,iBAAiB,CACxB,GAA0C;IAE1C,IAAI,CAAC,GAAG,IAAI,OAAO,GAAG,KAAK,QAAQ;QAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IACzE,MAAM,MAAM,GAAG,OAAO,GAAG,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,IAAI,CAAC;IAClE,IAAI,MAAM,IAAI,MAAM,KAAK,MAAM;QAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IAC/D,8EAA8E;IAC9E,MAAM,MAAM,GACV,CAAC,OAAO,GAAG,CAAC,MAAM,KAAK,QAAQ,IAAI,GAAG,CAAC,MAAM,CAAC;QAC9C,CAAC,OAAO,GAAG,CAAC,IAAI,KAAK,QAAQ,IAAI,GAAG,CAAC,IAAI,CAAC;QAC1C,IAAI,CAAC;IACP,IAAI,MAAM;QAAE,OAAO,EAAE,MAAM,EAAE,MAAM,IAAI,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC;IAC9D,IAAI,OAAO,GAAG,CAAC,QAAQ,KAAK,QAAQ,EAAE,CAAC;QACrC,MAAM,GAAG,GAAG,cAAc,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC;QACzC,OAAO,EAAE,MAAM,EAAE,GAAG,CAAC,MAAM,IAAI,MAAM,EAAE,IAAI,EAAE,GAAG,CAAC,IAAI,EAAE,CAAC;IAC1D,CAAC;IACD,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;AAChC,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,kBAAkB,CAChC,IAAY;IAKZ,IAAI,GAAY,CAAC;IACjB,IAAI,CAAC;QACH,GAAG,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IACzB,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC;IAC7B,CAAC;IACD,IAAI,CAAC,GAAG,IAAI,OAAO,GAAG,KAAK,QAAQ;QAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC;IAChE,MAAM,GAAG,GAAG,GAA8B,CAAC;IAC3C,IAAI,OAAO,GAAG,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;QACnC,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,cAAc,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC;QACvD,OAAO,EAAE,IAAI,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC;IAC7C,CAAC;IACD,IAAI,OAAO,GAAG,CAAC,SAAS,KAAK,QAAQ,EAAE,CAAC;QACtC,MAAM,EAAE,IAAI,EAAE,CAAC,EAAE,GAAG,cAAc,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAClD,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,UAAU,EAAE,CAAC,EAAE,CAAC;IAC/C,CAAC;IACD,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC;AAC7B,CAAC;AAYD;;;;;;GAMG;AACH,MAAM,UAAU,kBAAkB,CAChC,UAAwD,EACxD,UAA0B,EAAE;IAE5B,MAAM,EAAE,GAAG,OAAO,UAAU,EAAE,EAAE,KAAK,QAAQ,CAAC,CAAC,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC;IAErE,mEAAmE;IACnE,IAAI,EAAE,KAAK,eAAe,EAAE,CAAC;QAC3B,OAAO,EAAE,MAAM,EAAE,cAAc,EAAE,WAAW,EAAE,eAAe,EAAE,CAAC;IAClE,CAAC;IAED,mEAAmE;IACnE,MAAM,MAAM,GAAG,iBAAiB,CAAC,UAAU,EAAE,GAAG,CAAC,CAAC;IAClD,IAAI,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,MAAM,KAAK,MAAM,EAAE,CAAC;QAC9C,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,CAAC;IACtE,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;QAChB,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,YAAY,EAAE,WAAW,EAAE,EAAE,EAAE,CAAC;IAC1F,CAAC;IAED,wEAAwE;IACxE,IAAI,EAAE,IAAI,OAAO,CAAC,iBAAiB,EAAE,CAAC;QACpC,MAAM,IAAI,GAAG,OAAO,CAAC,iBAAiB,CAAC,EAAE,CAAC,CAAC;QAC3C,IAAI,IAAI,EAAE,CAAC;YACT,MAAM,MAAM,GAAG,kBAAkB,CAAC,IAAI,CAAC,CAAC;YACxC,IAAI,MAAM,CAAC,IAAI,KAAK,QAAQ,EAAE,CAAC;gBAC7B,IAAI,MAAM,CAAC,MAAM,IAAI,MAAM,CAAC,MAAM,KAAK,MAAM,EAAE,CAAC;oBAC9C,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,CAAC;gBACtE,CAAC;gBACD,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;oBAChB,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,gBAAgB,EAAE,WAAW,EAAE,EAAE,EAAE,CAAC;gBAC9F,CAAC;YACH,CAAC;iBAAM,IAAI,MAAM,CAAC,IAAI,KAAK,YAAY,EAAE,CAAC;gBACxC,OAAO,EAAE,MAAM,EAAE,YAAY,EAAE,WAAW,EAAE,EAAE,EAAE,UAAU,EAAE,MAAM,CAAC,UAAU,EAAE,CAAC;YAClF,CAAC;QACH,CAAC;IACH,CAAC;IAED,sCAAsC;IACtC,OAAO,EAAE,MAAM,EAAE,YAAY,EAAE,WAAW,EAAE,EAAE,EAAE,CAAC;AACnD,CAAC;AAED,6EAA6E;AAC7E,MAAM,UAAU,iBAAiB,CAAC,SAAiB,EAAE,WAAmB;IACtE,OAAO,IAAI,CAAC,IAAI,CAAC,SAAS,EAAE,MAAM,EAAE,kBAAkB,EAAE,WAAW,EAAE,gBAAgB,CAAC,CAAC;AACzF,CAAC;AAED,oFAAoF;AACpF,MAAM,UAAU,yBAAyB,CACvC,SAAiB,EACjB,WAAmB;IAEnB,IAAI,CAAC;QACH,OAAO,EAAE,CAAC,YAAY,CAAC,iBAAiB,CAAC,SAAS,EAAE,WAAW,CAAC,EAAE,OAAO,CAAC,CAAC;IAC7E,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,wBAAwB,CACtC,UAAwD,EACxD,SAAiB;IAEjB,OAAO,kBAAkB,CAAC,UAAU,EAAE;QACpC,iBAAiB,EAAE,CAAC,WAAW,EAAE,EAAE,CAAC,yBAAyB,CAAC,SAAS,EAAE,WAAW,CAAC;KACtF,CAAC,CAAC;AACL,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,+BAA+B,CAC7C,SAAiB,EACjB,WAAmB;IAEnB,OAAO,wBAAwB,CAAC,EAAE,EAAE,EAAE,WAAW,EAAE,EAAE,SAAS,CAAC,CAAC;AAClE,CAAC"}

package/dist/src/ingestion/extractor.d.ts

@@ -0,0 +1,16 @@
+export interface ExtractedDoc {
+    content: string;
+    metadata: Record<string, unknown>;
+}
+/**
+ * Whether the file at the given path is a supported format.
+ */
+export declare function isSupported(filePath: string): boolean;
+/**
+ * Extract text content and metadata from a file.
+ * Dispatches to format-specific extractors for PDF/DOCX,
+ * falls through to UTF-8 text read for everything else.
+ * Caller should verify isSupported() first.
+ */
+export declare function extract(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=extractor.d.ts.map

package/dist/src/ingestion/extractor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractor.d.ts","sourceRoot":"","sources":["../../../src/ingestion/extractor.ts"],"names":[],"mappings":"AAWA,MAAM,WAAW,YAAY;IAC3B,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACnC;AAsCD;;GAEG;AACH,wBAAgB,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAOrD;AAED;;;;;GAKG;AACH,wBAAsB,OAAO,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAuBrE"}

package/dist/src/ingestion/extractor.js

@@ -0,0 +1,85 @@
+/**
+ * File content extraction with extension-based dispatch.
+ *
+ * Supports text/code/markdown (read as UTF-8), PDF (via pdfjs-dist),
+ * and DOCX (via mammoth). Returns file content and basic metadata.
+ */
+import fs from "node:fs";
+import path from "node:path";
+import { extractPdf } from "./extractors/pdf.js";
+import { extractDocx } from "./extractors/docx.js";
+const TEXT_EXTENSIONS = new Set([
+    // Markdown
+    ".md", ".markdown", ".mdx",
+    // Plain text
+    ".txt", ".text", ".log",
+    // JavaScript / TypeScript
+    ".ts", ".tsx", ".js", ".jsx", ".mjs", ".cjs",
+    // Python
+    ".py",
+    // Other languages
+    ".rb", ".go", ".rs", ".java", ".kt", ".scala",
+    ".c", ".cpp", ".h", ".hpp", ".cs", ".swift",
+    // Shell
+    ".sh", ".bash", ".zsh", ".fish",
+    // Config / data
+    ".json", ".yaml", ".yml", ".toml", ".ini", ".cfg",
+    ".xml", ".svg",
+    // Web
+    ".html", ".htm", ".css", ".scss", ".less",
+    // SQL
+    ".sql",
+    // Other
+    ".graphql", ".proto", ".r", ".R",
+    ".dockerfile", ".makefile",
+]);
+/** Names (no extension) that are extractable. */
+const TEXT_NAMES = new Set([
+    "Makefile", "Dockerfile", "Rakefile", "Gemfile",
+    ".gitignore", ".env", ".env.example",
+    "LICENSE", "CHANGELOG",
+]);
+const PDF_EXTENSIONS = new Set([".pdf"]);
+const DOCX_EXTENSIONS = new Set([".docx"]);
+/**
+ * Whether the file at the given path is a supported format.
+ */
+export function isSupported(filePath) {
+    const ext = path.extname(filePath).toLowerCase();
+    if (TEXT_EXTENSIONS.has(ext))
+        return true;
+    if (PDF_EXTENSIONS.has(ext))
+        return true;
+    if (DOCX_EXTENSIONS.has(ext))
+        return true;
+    const base = path.basename(filePath);
+    return TEXT_NAMES.has(base);
+}
+/**
+ * Extract text content and metadata from a file.
+ * Dispatches to format-specific extractors for PDF/DOCX,
+ * falls through to UTF-8 text read for everything else.
+ * Caller should verify isSupported() first.
+ */
+export async function extract(filePath) {
+    const ext = path.extname(filePath).toLowerCase();
+    if (PDF_EXTENSIONS.has(ext)) {
+        return extractPdf(filePath);
+    }
+    if (DOCX_EXTENSIONS.has(ext)) {
+        return extractDocx(filePath);
+    }
+    const [content, stats] = await Promise.all([
+        fs.promises.readFile(filePath, "utf-8"),
+        fs.promises.stat(filePath),
+    ]);
+    return {
+        content,
+        metadata: {
+            extension: ext,
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+        },
+    };
+}
+//# sourceMappingURL=extractor.js.map

package/dist/src/ingestion/extractor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractor.js","sourceRoot":"","sources":["../../../src/ingestion/extractor.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AACjD,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAOnD,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC;IAC9B,WAAW;IACX,KAAK,EAAE,WAAW,EAAE,MAAM;IAC1B,aAAa;IACb,MAAM,EAAE,OAAO,EAAE,MAAM;IACvB,0BAA0B;IAC1B,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM;IAC5C,SAAS;IACT,KAAK;IACL,kBAAkB;IAClB,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ;IAC7C,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ;IAC3C,QAAQ;IACR,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO;IAC/B,gBAAgB;IAChB,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM;IACjD,MAAM,EAAE,MAAM;IACd,MAAM;IACN,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IACzC,MAAM;IACN,MAAM;IACN,QAAQ;IACR,UAAU,EAAE,QAAQ,EAAE,IAAI,EAAE,IAAI;IAChC,aAAa,EAAE,WAAW;CAC3B,CAAC,CAAC;AAEH,iDAAiD;AACjD,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IACzB,UAAU,EAAE,YAAY,EAAE,UAAU,EAAE,SAAS;IAC/C,YAAY,EAAE,MAAM,EAAE,cAAc;IACpC,SAAS,EAAE,WAAW;CACvB,CAAC,CAAC;AAEH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;AACzC,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;AAE3C;;GAEG;AACH,MAAM,UAAU,WAAW,CAAC,QAAgB;IAC1C,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IACjD,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACzC,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACrC,OAAO,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;AAC9B,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,QAAgB;IAC5C,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IAEjD,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC;IAC9B,CAAC;IACD,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QAC7B,OAAO,WAAW,CAAC,QAAQ,CAAC,CAAC;IAC/B,CAAC;IAED,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACzC,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC;QACvC,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC;KAC3B,CAAC,CAAC;IAEH,OAAO;QACL,OAAO;QACP,QAAQ,EAAE;YACR,SAAS,EAAE,GAAG;YACd,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;SACvC;KACF,CAAC;AACJ,CAAC"}

package/dist/src/ingestion/extractors/docx.d.ts

@@ -0,0 +1,3 @@
+import type { ExtractedDoc } from "../extractor.js";
+export declare function extractDocx(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=docx.d.ts.map

package/dist/src/ingestion/extractors/docx.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx.d.ts","sourceRoot":"","sources":["../../../../src/ingestion/extractors/docx.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAEpD,wBAAsB,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAYzE"}

package/dist/src/ingestion/extractors/docx.js

@@ -0,0 +1,20 @@
+/**
+ * DOCX text extraction using mammoth.
+ * Returns plain text from the document body.
+ */
+import fs from "node:fs";
+import path from "node:path";
+import mammoth from "mammoth";
+export async function extractDocx(filePath) {
+    const stats = await fs.promises.stat(filePath);
+    const result = await mammoth.extractRawText({ path: filePath });
+    return {
+        content: result.value.trim(),
+        metadata: {
+            extension: path.extname(filePath).toLowerCase(),
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+        },
+    };
+}
+//# sourceMappingURL=docx.js.map

package/dist/src/ingestion/extractors/docx.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx.js","sourceRoot":"","sources":["../../../../src/ingestion/extractors/docx.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,OAAO,MAAM,SAAS,CAAC;AAG9B,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC/C,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,cAAc,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;IAEhE,OAAO;QACL,OAAO,EAAE,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE;QAC5B,QAAQ,EAAE;YACR,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAC/C,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;SACvC;KACF,CAAC;AACJ,CAAC"}

package/dist/src/ingestion/extractors/pdf.d.ts

@@ -0,0 +1,3 @@
+import type { ExtractedDoc } from "../extractor.js";
+export declare function extractPdf(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=pdf.d.ts.map

package/dist/src/ingestion/extractors/pdf.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf.d.ts","sourceRoot":"","sources":["../../../../src/ingestion/extractors/pdf.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAEpD,wBAAsB,UAAU,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAyBxE"}

package/dist/src/ingestion/extractors/pdf.js

@@ -0,0 +1,32 @@
+/**
+ * PDF text extraction using pdfjs-dist (Mozilla PDF.js).
+ * Returns extracted text from all pages. Image-only PDFs
+ * produce empty content (not an error).
+ */
+import fs from "node:fs";
+import path from "node:path";
+import { getDocument } from "pdfjs-dist/legacy/build/pdf.mjs";
+export async function extractPdf(filePath) {
+    const data = new Uint8Array(await fs.promises.readFile(filePath));
+    const stats = await fs.promises.stat(filePath);
+    const doc = await getDocument({ data }).promise;
+    const pages = [];
+    for (let i = 1; i <= doc.numPages; i++) {
+        const page = await doc.getPage(i);
+        const textContent = await page.getTextContent();
+        const pageText = textContent.items
+            .map((item) => ("str" in item ? item.str : ""))
+            .join(" ");
+        pages.push(pageText);
+    }
+    return {
+        content: pages.join("\n").trim(),
+        metadata: {
+            extension: path.extname(filePath).toLowerCase(),
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+            page_count: doc.numPages,
+        },
+    };
+}
+//# sourceMappingURL=pdf.js.map

package/dist/src/ingestion/extractors/pdf.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf.js","sourceRoot":"","sources":["../../../../src/ingestion/extractors/pdf.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,WAAW,EAAE,MAAM,iCAAiC,CAAC;AAG9D,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,QAAgB;IAC/C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IAClE,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAE/C,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC;IAChD,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAClC,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;QAChD,MAAM,QAAQ,GAAG,WAAW,CAAC,KAAK;aAC/B,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,CAAC,CAAE,IAAwB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;aACnE,IAAI,CAAC,GAAG,CAAC,CAAC;QACb,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IACvB,CAAC;IAED,OAAO;QACL,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE;QAChC,QAAQ,EAAE;YACR,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAC/C,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;YACtC,UAAU,EAAE,GAAG,CAAC,QAAQ;SACzB;KACF,CAAC;AACJ,CAAC"}