@reportwright/pdf 0.1.0-beta.2 → 0.1.0-beta.4

package/examples/bulk.mjs

@@ -0,0 +1,19 @@
+// Many files at once: extract text and encrypt in worker threads; a broken file fails alone.
+//
+//   node examples/bulk.mjs [outDir]   writes locked-1.pdf … locked-N.pdf
+import fs from 'node:fs';
+import path from 'node:path';
+import { toBytes, bulk } from '../dist/index.js';
+
+const dir = process.argv[2] ?? '.';
+const make = (label) => toBytes(async (pdf) => {
+  const p = pdf.addPage(); p.text(`${label}: total due 1,250.00`, { x: 40, y: 40, font: pdf.standardFont('Helvetica'), size: 14 }); await p.end();
+});
+const files = [await make('Invoice 1'), new Uint8Array([0x25, 0x50]), await make('Invoice 3')]; // the second is not a PDF
+
+const texts = await bulk(files, 'extractText', { workers: 2 });
+texts.forEach((r, i) => console.log(i + 1, r.ok ? r.value : `failed: ${r.error.code} ${r.error.message}`));
+
+const locked = await bulk(files, 'encrypt', { workers: 2, encrypt: { userPassword: 'open me', ownerPassword: 'full access' } });
+locked.forEach((r, i) => { if (r.ok) fs.writeFileSync(path.join(dir, `locked-${i + 1}.pdf`), r.value); });
+console.log(`encrypted ${locked.filter((r) => r.ok).length} of ${files.length}`);

package/index.d.ts

@@ -650,7 +650,7 @@ export interface Pdf {
    * textMapping: 'carrier' (default: Firefox, pdf.js, search) or 'actualText' (a shaped cluster's text only in its ActualText: MuPDF, AI pipelines) */
   embedFont(bytes: Uint8Array | ArrayBuffer, o?: { subset?: boolean | Subsetter; shaper?: Shaper; textMapping?: 'carrier' | 'actualText' }): Promise<Font>;
   /** a JPEG or PNG, written now as an image XObject. colorSpace: an ICC colour space with the image's component count */
-  embedImage(bytes: Uint8Array | ArrayBuffer, o?: { colorSpace?: IccColorSpace; /** the decoded size a PNG with alpha or interlacing may have; default 64 MB */ maxDecodedBytes?: number }): Promise<Image>;
+  embedImage(bytes: Uint8Array | ArrayBuffer, o?: { colorSpace?: IccColorSpace; /** the decoded size a PNG with alpha or interlacing may have; default 64 MB */ maxDecodedBytes?: number; /** zlib level for a re-compressed PNG (alpha, interlacing): 'default' = 6, or 3 above 2 megapixels; 'fast' = 1; or 1–9 */ compression?: 'default' | 'fast' | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 }): Promise<Image>;
   /** a spot colour: its name and the device colour (gray, RGB, CMYK) shown for it at full tint. One definition per name */
   spotColor(name: string, alternate: number | string | readonly number[]): SpotColor;
   /** an ICC-based colour space from a gray, RGB or CMYK profile (its header is checked) */
@@ -846,3 +846,32 @@ export interface LoadedPdf {
 export function loadPdf(bytes: Uint8Array | ArrayBuffer, options?: LoadOptions): Promise<LoadedPdf>;
 /** every page of each document, in order, in a new file (the first one's catalog, with what the others bring) */
 export function mergePdfs(docs: LoadedPdf[], options?: SaveOptions): Promise<SaveResult & { bytes: Uint8Array }>;
+/** the tasks bulk() runs, by name (no functions cross to a worker) */
+export type BulkTask = 'extractText' | 'merge' | 'split' | 'encrypt';
+export const BULK_TASKS: readonly BulkTask[];
+/** what each task takes per item, and gives back */
+export interface BulkTaskTypes {
+  extractText: { input: Uint8Array | ArrayBuffer; value: string };
+  merge: { input: (Uint8Array | ArrayBuffer)[]; value: Uint8Array };
+  split: { input: Uint8Array | ArrayBuffer; value: Uint8Array[] };
+  encrypt: { input: Uint8Array | ArrayBuffer; value: Uint8Array };
+}
+export type BulkResult<V> = { ok: true; value: V } | { ok: false; error: { name: string; code?: string; message: string } };
+export interface BulkOptions {
+  /** worker threads; default os.availableParallelism() - 1 (at least 1). 1, or no node:worker_threads (browsers): one at a time, in this thread */
+  workers?: number;
+  /** each worker's heap cap in MB (resourceLimits.maxOldGenerationSizeMb), default 512: an item past it fails with E_WORKER, alone */
+  memoryMb?: number;
+  /** loadPdf's options, for every input */
+  password?: string;
+  budget?: ReadBudget;
+  timeoutMs?: number;
+  /** the encrypt task's encryption */
+  encrypt?: EncryptOptions;
+  /** save()'s options for merge, split and encrypt */
+  save?: SaveOptions;
+  /** each item's result as it finishes (index in inputs) */
+  onResult?: (index: number, result: BulkResult<unknown>) => void;
+}
+/** One named task over many inputs, in parallel worker threads (Node); results in input order, a failing item never stops the rest. */
+export function bulk<T extends BulkTask>(inputs: Iterable<BulkTaskTypes[T]['input']>, task: T, options?: BulkOptions): Promise<BulkResult<BulkTaskTypes[T]['value']>[]>;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@reportwright/pdf",
-  "version": "0.1.0-beta.2",
+  "version": "0.1.0-beta.4",
   "description": "A streaming PDF writer for Node and browsers with no runtime dependencies: flat memory, PDF/A-1b to 3a, PDF/UA-1, Factur-X, embedded TrueType with Unicode text.",
   "type": "module",
   "license": "MIT",