@pdfluent/sdk-wasm 1.0.0 → 1.1.0

package/LICENSE-COMMERCIAL

@@ -98,6 +98,14 @@ to relicense and need no licence from PDFluent. `NOTICE` records which they are;
 
 This licence covers the components on PDFluent's side of that boundary.
 
+Optional model files and runtime libraries that the software downloads on
+your explicit request (for example with `pdfluent models install`) are
+licensed as follows. Model files trained by PDFluent are licensed under
+the same terms as the software. Third-party model files and runtime
+libraries are licensed to you under their own terms, listed in the notice
+that accompanies them, and are provided as-is; PDFluent gives no warranty
+or indemnity for them.
+
 
 4. Attribution
 --------------

package/NOTICE

@@ -71,6 +71,9 @@ that. See https://pdfluent.com/sdk/pricing.
 
   pdfluent             — the unified facade SDK
   pdf-engine           — orchestration and document model
+                         (the TATR post-processing behind `ort-tables` is a port of
+                          Microsoft's table-transformer, © Microsoft Corporation,
+                          MIT; that notice ships inside the crate as THIRD_PARTY.md)
   pdf-manip            — page-level manipulation (merge, split, rotate, etc.)
                          (six unmodified Adobe CMap resources, compiled in
                           to read the character codes of a predefined CMap;
@@ -81,11 +84,14 @@ that. See https://pdfluent.com/sdk/pricing.
   pdfluent-sign        — digital signatures (PAdES, S/MIME)
   pdf-compliance       — PDF/A, PDF/UA, PDF/X validation and conversion
   pdfluent-extract     — text and image extraction
+  epub-extract         — EPUB content extraction
   pdfluent-forms       — AcroForm support
   pdf-annot            — annotations
   pdf-docx             — DOCX export
   pdf-xlsx             — XLSX export and table extraction
+  xlsx-extract         — XLSX workbook extraction to engine blocks
   pdf-pptx             — PowerPoint PPTX export
+  pptx-extract         — PowerPoint PPTX extraction to blocks
   pdf-invoice          — FDF/XFDF form data exchange and ZUGFeRD/Factur-X
   pdf-text-format      — text run formatting
   pdf-standard-fonts   — the Standard 14 font programs, compiled in
@@ -107,6 +113,7 @@ that. See https://pdfluent.com/sdk/pricing.
   xfa-js-sandboxed     — sandboxed JavaScript runtime for XFA scripts
   formcalc-interpreter — FormCalc evaluation
   xfa-json             — XFA JSON conversion
+  pdfluent-cli         — command-line interface over the facade
 
 The Software produces the same output whether or not a commercial licence is
 held: there is no licence key, nothing is withheld and no marker is added

package/THIRD_PARTY_LICENSES.txt

@@ -731,6 +731,114 @@ CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
 OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
 SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
+================================================================================
+tract — ONNX inference for the optional vision layer (#636)
+https://github.com/sonos/tract
+================================================================================
+
+The `vision` cargo feature of pdf-engine (off by default; passthrough in
+pdfluent) runs ONNX models through tract-onnx 0.23 (default-features off, so
+without the `transformers` op-set). No trained model ships; the only ONNX
+bytes in the tree are the 152-byte identity fixture the tests run.
+
+Copyright (c) Mathieu Poumeyrol and the tract contributors.
+
+Dual-licensed under MIT and Apache-2.0, at our option. The crates as resolved
+(tract-onnx, tract-onnx-opl, tract-hir, tract-core, tract-data,
+tract-extra, tract-nnef, tract-linalg and their transitive dependencies
+bytes, derive-new, log, memmap2, nom, nom-language, num-integer,
+num-rational, num-traits, prost, smallvec and the tract-*-sys/metal
+helpers) each declare MIT OR Apache-2.0 and are covered by the texts below;
+both are permissive and impose no copyleft obligation on the product. This
+entry stands outside the `cargo license 2026-05-26` list above because that
+list predates the dependency; the next regeneration absorbs these crates.
+
+--- MIT License ---
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
+
+--- Apache License, Version 2.0 ---
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+
+================================================================================
+table-transformer — Table Transformer structure post-processing (MIT)
+https://github.com/microsoft/table-transformer
+commit 16d124f616109746b7785f03085100f1f6247575
+================================================================================
+
+Copyright (c) Microsoft Corporation.
+
+What is used, in two forms:
+
+  crates/pdf-engine/src/learned/tables_tatr.rs        a Rust port of the
+             publisher's row/column/header/spanning-cell reconciliation and
+             grid-cell construction (`objects_to_cells` and everything it
+             calls), behind the `ort-tables` feature of pdf-engine. Same
+             thresholds, same refinement, same cells; the parity test feeds
+             raw ONNX outputs and requires the Python-decoded cells, value
+             for value.
+  scripts/extraction/tables_b/tatr_postprocess.py     the Python original,
+             `src/postprocess.py` at the commit above, vendored verbatim
+             (see the header in that file). The zero-shot eval runs TATR
+             through it rather than through a reimplementation, so a TATR
+             number is the publisher's reading of its own boxes.
+
+MIT is permissive, no copyleft: the licence text below is that commit's
+LICENSE, word for word. The same copyright line and text ship inside the
+pdf-engine crate as THIRD_PARTY.md, because a crates.io reader has this
+file's section in front of them only through the repository.
+
+--- MIT License ---
+
+MIT License
+
+Copyright (c) Microsoft Corporation.
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE
+
 ================================================================================
 END OF THIRD PARTY LICENSES
 ================================================================================

package/package.json

@@ -2,12 +2,8 @@
   "name": "@pdfluent/sdk-wasm",
   "type": "module",
   "description": "PDFluent browser SDK — find and replace PDF text programmatically while preserving fonts and layout, plus edit, annotate, redact, sign and validate PDFs (including XFA), entirely client-side via WASM.",
-  "version": "1.0.0",
+  "version": "1.1.0",
   "license": "SEE LICENSE IN LICENSE-OFFER",
-  "repository": {
-    "type": "git",
-    "url": "https://github.com/pdfluent/pdfluent-sdk"
-  },
   "files": [
     "xfa_wasm_bg.wasm",
     "xfa_wasm.js",
@@ -44,10 +40,8 @@
   ],
   "homepage": "https://pdfluent.com",
   "author": "Innovation Trigger BV <team@pdfluent.com>",
-  "bugs": {
-    "url": "https://github.com/pdfluent/pdfluent-sdk/issues"
-  },
-  "publishConfig": {
-    "access": "public"
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/pdfluent/pdfluent-sdk"
   }
 }

package/xfa_wasm.d.ts

@@ -1,6 +1,85 @@
 /* tslint:disable */
 /* eslint-disable */
 
+/** A stamp with preset text, size, rotation, opacity and colour (red, 60pt, 45°, 25%). */
+export type WatermarkPreset = "draft" | "confidential" | "approved" | "final" | "copy" | "notForDistribution";
+
+/** Options for `PdfDoc.addWatermark`. Give `text`, `preset`, or both (then `text` wins). */
+export interface WatermarkOptions {
+    text?: string;
+    preset?: WatermarkPreset;
+    /** Points. Default 48 (60 with a preset). */
+    fontSize?: number;
+    /** Degrees counter-clockwise. Default 45. */
+    rotation?: number;
+    /** 0..1. Default 0.3 (0.25 with a preset). */
+    opacity?: number;
+    /** `"#rrggbb"`. Default mid grey (red with a preset). */
+    color?: string;
+    /** `"center"` (default), or a corner / exact point in PDF points. */
+    position?: "center" | { anchor: "topLeft" | "topRight" | "bottomLeft" | "bottomRight" | "exact"; x: number; y: number };
+    /** Default `"foreground"`. */
+    layer?: "foreground" | "background";
+    /** Repeat across the page. `true` = 200pt spacing. */
+    tile?: boolean | { hSpacing?: number; vSpacing?: number };
+    /** `"all"` (default), `"even"`, `"odd"`, 0-based page indices, or an inclusive 0-based range. */
+    pages?: "all" | "even" | "odd" | number[] | { from: number; to: number };
+}
+
+
+
+/** An image exported faithfully. `page` is 0-based: the first page it is used on. */
+export interface ExtractedImage {
+    page: number;
+    width: number;
+    height: number;
+    mime: "image/png" | "image/jpeg";
+    data: Uint8Array;
+}
+
+/** An image left out because it could not be exported without changing how it looks. */
+export interface SkippedImage {
+    page: number;
+    width: number;
+    height: number;
+    reason: "unsupported_filter" | "transparency" | "decode_array" | "cmyk_jpeg" | "color_space" | "bit_depth" | "image_mask" | "undecodable";
+    detail: string;
+}
+
+export interface ImageExtraction {
+    images: ExtractedImage[];
+    skipped: SkippedImage[];
+}
+
+
+
+/** Options for `signPdf`. The signature is a cryptographic (PAdES) signature, not a drawn one. */
+export interface SignPdfOptions {
+    reason?: string;
+    location?: string;
+    contactInfo?: string;
+    /** Default "Signature1". */
+    fieldName?: string;
+    /** Show the signature on a page (0-based) at a rectangle in PDF points. Invisible when absent. */
+    visibleRect?: { page: number; x: number; y: number; width: number; height: number };
+}
+
+
+
+/** What a user who opens the file with the user password may do. Every flag defaults to `true`. */
+export interface EncryptPermissions {
+    print?: boolean;
+    printHighQuality?: boolean;
+    modify?: boolean;
+    copy?: boolean;
+    annotate?: boolean;
+    fillForms?: boolean;
+    accessibility?: boolean;
+    assemble?: boolean;
+}
+
+
+
 /**
  * Chroma subsampling format
  */
@@ -77,8 +156,16 @@ export class PdfDoc {
     /**
      * Apply a diagonal text watermark to every page.
      * `opacity` in 0.0..=1.0. Default rotation 45 deg, default colour gray.
+     *
+     * Kept for compatibility; `addWatermark` takes the full set of options.
      */
     addTextWatermark(text: string, opacity: number): Uint8Array;
+    /**
+     * Apply a text watermark with full control over text or preset, size,
+     * rotation, opacity, colour, position, layer, tiling and pages. Returns
+     * the new PDF bytes. Text only: there is no image watermark.
+     */
+    addWatermark(options: WatermarkOptions): Uint8Array;
     /**
      * Re-encode all content streams with deflate compression. Removes
      * no-op overhead. Returns the optimised PDF as a `Uint8Array`.
@@ -99,6 +186,58 @@ export class PdfDoc {
      * DSS (Document Security Store) info as JSON, or null if absent.
      */
     dssInfo(): string | undefined;
+    /**
+     * The document's blocks as JSON, in reading order (#573).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * `{"blocks": [...]}`, one entry per paragraph, heading, list item, table
+     * cell and running head, each with its kind, page, box, text, spans and
+     * place in the reading order. `docs/STRUCTURED_EXTRACTION.md` is the key
+     * to every field.
+     *
+     * A JSON string and not a structured value, which is what every other
+     * structured reader in this binding hands back: one shape crosses the
+     * boundary, `JSON.parse` is what the caller already does for
+     * `getTextPositions` and the annotations, and a `serde` mirror of the
+     * block model would be a second definition of it to keep in step.
+     *
+     * An XFA form is read from its flattened form, for the reason
+     * `extractMarkdown` above gives.
+     */
+    extractBlocks(first_page?: number | null, page_count?: number | null): string;
+    /**
+     * The document as a Docling `DoclingDocument`, as JSON (#629).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * The schema version the engine pins, which docling-core's
+     * `DoclingDocument.model_validate_json` loads as it is. `name` is
+     * `DoclingDocument.name`, which Docling fills with the file's stem;
+     * absent means `"document"`. An XFA form is read from its flattened form,
+     * for the reason `extractMarkdown` above gives.
+     */
+    extractDocling(first_page?: number | null, page_count?: number | null, name?: string | null): string;
+    /**
+     * The document as markdown, in reading order (#573).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * Headings become `#` to `###`, list items `-` and `1.`, a table a pipe
+     * table or an HTML one where markdown cannot hold its shape, and a change
+     * of page an HTML comment. The running heads a document repeats are not in
+     * it; they are in `extractBlocks`.
+     *
+     * `firstPage` and `pageCount` choose a window of pages; leave both out for
+     * the whole document. The document is read whole whatever the window is,
+     * because what a block IS is decided against the rest of the document.
+     *
+     * AN XFA FORM IS READ FROM ITS FLATTENED FORM, the way `text` and
+     * `getTextPositions` read one: a dynamic XFA document draws its fields
+     * from a template rather than from page content, so the blocks of the
+     * unflattened file are the blocks of an empty page.
+     */
+    extractMarkdown(first_page?: number | null, page_count?: number | null): string;
     /**
      * Extract the listed pages (0-based) into a new document.
      * Used to implement "split" workflows in the browser.
@@ -661,6 +800,52 @@ export class XfaEngine {
     static version(): string;
 }
 
+/**
+ * Remove the password from a PDF whose password you know, and return the
+ * unprotected bytes; open them with `PdfDoc.open`.
+ *
+ * The owner password always works. The user password works only when the
+ * file has no permission restrictions. Throws `PdfluentError` with
+ * `E-SECURITY-DECRYPTION-FAILED` for a password that does not entitle the
+ * caller, and `E-WASM-INVALID-ARGUMENT` for a PDF without a password.
+ */
+export function decryptPdf(bytes: Uint8Array, password: string): Uint8Array;
+
+/**
+ * Encrypt a PDF with AES-256 and return the encrypted bytes.
+ *
+ * `userPassword` is needed to open the file (may be empty: then anyone can
+ * open it, and the permissions apply); `ownerPassword` lifts the permissions
+ * (empty: the user password serves as both). Refuses a PDF that is already
+ * encrypted. Throws `PdfluentError` with `E-WASM-INVALID-ARGUMENT` or
+ * `E-PARSE-INVALID-PDF`.
+ */
+export function encryptPdf(bytes: Uint8Array, user_password: string, owner_password: string, permissions?: EncryptPermissions | null): Uint8Array;
+
+/**
+ * Extract the images of a PDF as PNG or JPEG files.
+ *
+ * Returns `{ images, skipped }`. An image that could not be exported exactly
+ * as it looks -- JPEG 2000, JBIG2 or fax compression, a transparency mask, a
+ * Decode array, CMYK, an ICC profile other than sRGB -- is listed in `skipped`
+ * with the reason, never emitted wrong. Each image is listed once, on the
+ * first page (0-based) that uses it.
+ */
+export function extractImages(bytes: Uint8Array): ImageExtraction;
+
+/**
+ * Sign a PDF with your own certificate (a .p12/.pfx file and its password)
+ * and return the signed bytes. The private key never leaves the page.
+ *
+ * This is a cryptographic PAdES signature (ETSI.CAdES.detached), appended as
+ * an incremental update so earlier signatures stay valid. Without a
+ * timestamp authority: no TSA or LTV in the browser. A self-signed
+ * certificate verifies, but readers such as Acrobat show its validity as
+ * unknown because they do not trust the issuer. Throws `PdfluentError` with
+ * `E-SECURITY-DECRYPTION-FAILED` for a wrong certificate password.
+ */
+export function signPdf(bytes: Uint8Array, pkcs12: Uint8Array, password: string, options?: SignPdfOptions | null): Uint8Array;
+
 export type InitInput = RequestInfo | URL | Response | BufferSource | WebAssembly.Module;
 
 export interface InitOutput {
@@ -669,14 +854,21 @@ export interface InitOutput {
     readonly __wbg_pdfdocmut_free: (a: number, b: number) => void;
     readonly __wbg_texteditor_free: (a: number, b: number) => void;
     readonly __wbg_xfaengine_free: (a: number, b: number) => void;
+    readonly decryptPdf: (a: number, b: number, c: number, d: number, e: number) => void;
+    readonly encryptPdf: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+    readonly extractImages: (a: number, b: number, c: number) => void;
     readonly pdfdoc_addFreeText: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => void;
     readonly pdfdoc_addHighlight: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => void;
     readonly pdfdoc_addStickyNote: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => void;
     readonly pdfdoc_addTextWatermark: (a: number, b: number, c: number, d: number, e: number) => void;
+    readonly pdfdoc_addWatermark: (a: number, b: number, c: number) => void;
     readonly pdfdoc_compress: (a: number, b: number) => void;
     readonly pdfdoc_convertToPdfa: (a: number, b: number, c: number, d: number) => void;
     readonly pdfdoc_deletePages: (a: number, b: number, c: number, d: number) => void;
     readonly pdfdoc_dssInfo: (a: number, b: number) => void;
+    readonly pdfdoc_extractBlocks: (a: number, b: number, c: number, d: number) => void;
+    readonly pdfdoc_extractDocling: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
+    readonly pdfdoc_extractMarkdown: (a: number, b: number, c: number, d: number) => void;
     readonly pdfdoc_extractPages: (a: number, b: number, c: number, d: number) => void;
     readonly pdfdoc_flattenXfa: (a: number, b: number) => void;
     readonly pdfdoc_getAnnotations: (a: number, b: number, c: number) => void;
@@ -728,6 +920,7 @@ export interface InitOutput {
     readonly pdfdocmut_setFormFields: (a: number, b: number, c: number, d: number) => void;
     readonly pdfdocmut_setMultiSelect: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
     readonly pdfdocmut_setTextRunStyle: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+    readonly signPdf: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
     readonly texteditor_findText: (a: number, b: number, c: number, d: number) => void;
     readonly texteditor_open: (a: number, b: number, c: number) => void;
     readonly texteditor_pageCount: (a: number) => number;

package/xfa_wasm.js

@@ -1,5 +1,5 @@
 /* @ts-self-types="./xfa_wasm.d.ts" */
-import { __pdfluent_error_ctor } from './snippets/xfa-wasm-df2db187f9a0e994/inline0.js';
+import { __pdfluent_error_ctor } from './snippets/xfa-wasm-d1985887593da269/inline0.js';
 
 
 /**
@@ -168,6 +168,8 @@ export class PdfDoc {
     /**
      * Apply a diagonal text watermark to every page.
      * `opacity` in 0.0..=1.0. Default rotation 45 deg, default colour gray.
+     *
+     * Kept for compatibility; `addWatermark` takes the full set of options.
      * @param {string} text
      * @param {number} opacity
      * @returns {Uint8Array}
@@ -192,6 +194,31 @@ export class PdfDoc {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Apply a text watermark with full control over text or preset, size,
+     * rotation, opacity, colour, position, layer, tiling and pages. Returns
+     * the new PDF bytes. Text only: there is no image watermark.
+     * @param {WatermarkOptions} options
+     * @returns {Uint8Array}
+     */
+    addWatermark(options) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.pdfdoc_addWatermark(retptr, this.__wbg_ptr, addHeapObject(options));
+            var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+            var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+            var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+            var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+            if (r3) {
+                throw takeObject(r2);
+            }
+            var v1 = getArrayU8FromWasm0(r0, r1).slice();
+            wasm.__wbindgen_export4(r0, r1 * 1, 1);
+            return v1;
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+        }
+    }
     /**
      * Re-encode all content streams with deflate compression. Removes
      * no-op overhead. Returns the optimised PDF as a `Uint8Array`.
@@ -288,6 +315,139 @@ export class PdfDoc {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * The document's blocks as JSON, in reading order (#573).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * `{"blocks": [...]}`, one entry per paragraph, heading, list item, table
+     * cell and running head, each with its kind, page, box, text, spans and
+     * place in the reading order. `docs/STRUCTURED_EXTRACTION.md` is the key
+     * to every field.
+     *
+     * A JSON string and not a structured value, which is what every other
+     * structured reader in this binding hands back: one shape crosses the
+     * boundary, `JSON.parse` is what the caller already does for
+     * `getTextPositions` and the annotations, and a `serde` mirror of the
+     * block model would be a second definition of it to keep in step.
+     *
+     * An XFA form is read from its flattened form, for the reason
+     * `extractMarkdown` above gives.
+     * @param {number | null} [first_page]
+     * @param {number | null} [page_count]
+     * @returns {string}
+     */
+    extractBlocks(first_page, page_count) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.pdfdoc_extractBlocks(retptr, this.__wbg_ptr, isLikeNone(first_page) ? Number.MAX_SAFE_INTEGER : (first_page) >>> 0, isLikeNone(page_count) ? Number.MAX_SAFE_INTEGER : (page_count) >>> 0);
+            var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+            var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+            var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+            var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+            var ptr1 = r0;
+            var len1 = r1;
+            if (r3) {
+                ptr1 = 0; len1 = 0;
+                throw takeObject(r2);
+            }
+            deferred2_0 = ptr1;
+            deferred2_1 = len1;
+            return getStringFromWasm0(ptr1, len1);
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+            wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
+        }
+    }
+    /**
+     * The document as a Docling `DoclingDocument`, as JSON (#629).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * The schema version the engine pins, which docling-core's
+     * `DoclingDocument.model_validate_json` loads as it is. `name` is
+     * `DoclingDocument.name`, which Docling fills with the file's stem;
+     * absent means `"document"`. An XFA form is read from its flattened form,
+     * for the reason `extractMarkdown` above gives.
+     * @param {number | null} [first_page]
+     * @param {number | null} [page_count]
+     * @param {string | null} [name]
+     * @returns {string}
+     */
+    extractDocling(first_page, page_count, name) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(name) ? 0 : passStringToWasm0(name, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.pdfdoc_extractDocling(retptr, this.__wbg_ptr, isLikeNone(first_page) ? Number.MAX_SAFE_INTEGER : (first_page) >>> 0, isLikeNone(page_count) ? Number.MAX_SAFE_INTEGER : (page_count) >>> 0, ptr0, len0);
+            var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+            var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+            var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+            var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+            var ptr2 = r0;
+            var len2 = r1;
+            if (r3) {
+                ptr2 = 0; len2 = 0;
+                throw takeObject(r2);
+            }
+            deferred3_0 = ptr2;
+            deferred3_1 = len2;
+            return getStringFromWasm0(ptr2, len2);
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+            wasm.__wbindgen_export4(deferred3_0, deferred3_1, 1);
+        }
+    }
+    /**
+     * The document as markdown, in reading order (#573).
+     *
+     * Preview: output quality is still improving and the API may change in a minor release.
+     *
+     * Headings become `#` to `###`, list items `-` and `1.`, a table a pipe
+     * table or an HTML one where markdown cannot hold its shape, and a change
+     * of page an HTML comment. The running heads a document repeats are not in
+     * it; they are in `extractBlocks`.
+     *
+     * `firstPage` and `pageCount` choose a window of pages; leave both out for
+     * the whole document. The document is read whole whatever the window is,
+     * because what a block IS is decided against the rest of the document.
+     *
+     * AN XFA FORM IS READ FROM ITS FLATTENED FORM, the way `text` and
+     * `getTextPositions` read one: a dynamic XFA document draws its fields
+     * from a template rather than from page content, so the blocks of the
+     * unflattened file are the blocks of an empty page.
+     * @param {number | null} [first_page]
+     * @param {number | null} [page_count]
+     * @returns {string}
+     */
+    extractMarkdown(first_page, page_count) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.pdfdoc_extractMarkdown(retptr, this.__wbg_ptr, isLikeNone(first_page) ? Number.MAX_SAFE_INTEGER : (first_page) >>> 0, isLikeNone(page_count) ? Number.MAX_SAFE_INTEGER : (page_count) >>> 0);
+            var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+            var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+            var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+            var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+            var ptr1 = r0;
+            var len1 = r1;
+            if (r3) {
+                ptr1 = 0; len1 = 0;
+                throw takeObject(r2);
+            }
+            deferred2_0 = ptr1;
+            deferred2_1 = len1;
+            return getStringFromWasm0(ptr1, len1);
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+            wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
+        }
+    }
     /**
      * Extract the listed pages (0-based) into a new document.
      * Used to implement "split" workflows in the browser.
@@ -2109,10 +2269,154 @@ export class XfaEngine {
     }
 }
 if (Symbol.dispose) XfaEngine.prototype[Symbol.dispose] = XfaEngine.prototype.free;
+
+/**
+ * Remove the password from a PDF whose password you know, and return the
+ * unprotected bytes; open them with `PdfDoc.open`.
+ *
+ * The owner password always works. The user password works only when the
+ * file has no permission restrictions. Throws `PdfluentError` with
+ * `E-SECURITY-DECRYPTION-FAILED` for a password that does not entitle the
+ * caller, and `E-WASM-INVALID-ARGUMENT` for a PDF without a password.
+ * @param {Uint8Array} bytes
+ * @param {string} password
+ * @returns {Uint8Array}
+ */
+export function decryptPdf(bytes, password) {
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        const ptr0 = passArray8ToWasm0(bytes, wasm.__wbindgen_export);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passStringToWasm0(password, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+        const len1 = WASM_VECTOR_LEN;
+        wasm.decryptPdf(retptr, ptr0, len0, ptr1, len1);
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+        var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+        if (r3) {
+            throw takeObject(r2);
+        }
+        var v3 = getArrayU8FromWasm0(r0, r1).slice();
+        wasm.__wbindgen_export4(r0, r1 * 1, 1);
+        return v3;
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+    }
+}
+
+/**
+ * Encrypt a PDF with AES-256 and return the encrypted bytes.
+ *
+ * `userPassword` is needed to open the file (may be empty: then anyone can
+ * open it, and the permissions apply); `ownerPassword` lifts the permissions
+ * (empty: the user password serves as both). Refuses a PDF that is already
+ * encrypted. Throws `PdfluentError` with `E-WASM-INVALID-ARGUMENT` or
+ * `E-PARSE-INVALID-PDF`.
+ * @param {Uint8Array} bytes
+ * @param {string} user_password
+ * @param {string} owner_password
+ * @param {EncryptPermissions | null} [permissions]
+ * @returns {Uint8Array}
+ */
+export function encryptPdf(bytes, user_password, owner_password, permissions) {
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        const ptr0 = passArray8ToWasm0(bytes, wasm.__wbindgen_export);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passStringToWasm0(user_password, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+        const len1 = WASM_VECTOR_LEN;
+        const ptr2 = passStringToWasm0(owner_password, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+        const len2 = WASM_VECTOR_LEN;
+        wasm.encryptPdf(retptr, ptr0, len0, ptr1, len1, ptr2, len2, isLikeNone(permissions) ? 0 : addHeapObject(permissions));
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+        var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+        if (r3) {
+            throw takeObject(r2);
+        }
+        var v4 = getArrayU8FromWasm0(r0, r1).slice();
+        wasm.__wbindgen_export4(r0, r1 * 1, 1);
+        return v4;
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+    }
+}
+
+/**
+ * Extract the images of a PDF as PNG or JPEG files.
+ *
+ * Returns `{ images, skipped }`. An image that could not be exported exactly
+ * as it looks -- JPEG 2000, JBIG2 or fax compression, a transparency mask, a
+ * Decode array, CMYK, an ICC profile other than sRGB -- is listed in `skipped`
+ * with the reason, never emitted wrong. Each image is listed once, on the
+ * first page (0-based) that uses it.
+ * @param {Uint8Array} bytes
+ * @returns {ImageExtraction}
+ */
+export function extractImages(bytes) {
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        const ptr0 = passArray8ToWasm0(bytes, wasm.__wbindgen_export);
+        const len0 = WASM_VECTOR_LEN;
+        wasm.extractImages(retptr, ptr0, len0);
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+        if (r2) {
+            throw takeObject(r1);
+        }
+        return takeObject(r0);
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+    }
+}
+
+/**
+ * Sign a PDF with your own certificate (a .p12/.pfx file and its password)
+ * and return the signed bytes. The private key never leaves the page.
+ *
+ * This is a cryptographic PAdES signature (ETSI.CAdES.detached), appended as
+ * an incremental update so earlier signatures stay valid. Without a
+ * timestamp authority: no TSA or LTV in the browser. A self-signed
+ * certificate verifies, but readers such as Acrobat show its validity as
+ * unknown because they do not trust the issuer. Throws `PdfluentError` with
+ * `E-SECURITY-DECRYPTION-FAILED` for a wrong certificate password.
+ * @param {Uint8Array} bytes
+ * @param {Uint8Array} pkcs12
+ * @param {string} password
+ * @param {SignPdfOptions | null} [options]
+ * @returns {Uint8Array}
+ */
+export function signPdf(bytes, pkcs12, password, options) {
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        const ptr0 = passArray8ToWasm0(bytes, wasm.__wbindgen_export);
+        const len0 = WASM_VECTOR_LEN;
+        const ptr1 = passArray8ToWasm0(pkcs12, wasm.__wbindgen_export);
+        const len1 = WASM_VECTOR_LEN;
+        const ptr2 = passStringToWasm0(password, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+        const len2 = WASM_VECTOR_LEN;
+        wasm.signPdf(retptr, ptr0, len0, ptr1, len1, ptr2, len2, isLikeNone(options) ? 0 : addHeapObject(options));
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        var r2 = getDataViewMemory0().getInt32(retptr + 4 * 2, true);
+        var r3 = getDataViewMemory0().getInt32(retptr + 4 * 3, true);
+        if (r3) {
+            throw takeObject(r2);
+        }
+        var v4 = getArrayU8FromWasm0(r0, r1).slice();
+        wasm.__wbindgen_export4(r0, r1 * 1, 1);
+        return v4;
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+    }
+}
 function __wbg_get_imports() {
     const import0 = {
         __proto__: null,
-        __wbg___pdfluent_error_ctor_e55fae67914d7975: function() {
+        __wbg___pdfluent_error_ctor_17e86bb485874967: function() {
             const ret = __pdfluent_error_ctor();
             return addHeapObject(ret);
         },
@@ -2127,6 +2431,14 @@ function __wbg_get_imports() {
             const ret = typeof(getObject(arg0)) === 'function';
             return ret;
         },
+        __wbg___wbindgen_is_null_2042690d351e14f0: function(arg0) {
+            const ret = getObject(arg0) === null;
+            return ret;
+        },
+        __wbg___wbindgen_is_undefined_35bb9f4c7fd651d5: function(arg0) {
+            const ret = getObject(arg0) === undefined;
+            return ret;
+        },
         __wbg___wbindgen_string_get_d109740c0d18f4d7: function(arg0, arg1) {
             const obj = getObject(arg1);
             const ret = typeof(obj) === 'string' ? obj : undefined;
@@ -2192,6 +2504,10 @@ function __wbg_get_imports() {
         __wbg_moveTo_11bf5a977e6b8610: function(arg0, arg1, arg2) {
             getObject(arg0).moveTo(arg1, arg2);
         },
+        __wbg_new_02d162bc6cf02f60: function() {
+            const ret = new Object();
+            return addHeapObject(ret);
+        },
         __wbg_new_1f236d63ba0c4784: function(arg0, arg1) {
             const ret = new Error(getStringFromWasm0(arg0, arg1));
             return addHeapObject(ret);
@@ -2200,6 +2516,10 @@ function __wbg_get_imports() {
             const ret = new Array();
             return addHeapObject(ret);
         },
+        __wbg_new_from_slice_269e35316ed2d061: function(arg0, arg1) {
+            const ret = new Uint8Array(getArrayU8FromWasm0(arg0, arg1));
+            return addHeapObject(ret);
+        },
         __wbg_new_with_u8_clamped_array_and_sh_13504c3c5394c7c9: function() { return handleError(function (arg0, arg1, arg2, arg3) {
             const ret = new ImageData(getClampedArrayU8FromWasm0(arg0, arg1), arg2 >>> 0, arg3 >>> 0);
             return addHeapObject(ret);
@@ -2273,6 +2593,10 @@ function __wbg_get_imports() {
         __wbg_set_width_25112eb6bf1148df: function(arg0, arg1) {
             getObject(arg0).width = arg1 >>> 0;
         },
+        __wbg_stringify_ef0c105b1ccc3849: function() { return handleError(function (arg0) {
+            const ret = JSON.stringify(getObject(arg0));
+            return addHeapObject(ret);
+        }, arguments); },
         __wbg_strokeRect_3430a0df4158ba67: function(arg0, arg1, arg2, arg3, arg4) {
             getObject(arg0).strokeRect(arg1, arg2, arg3, arg4);
         },

package/xfa_wasm_bg.wasm.d.ts

@@ -5,14 +5,21 @@ export const __wbg_pdfdoc_free: (a: number, b: number) => void;
 export const __wbg_pdfdocmut_free: (a: number, b: number) => void;
 export const __wbg_texteditor_free: (a: number, b: number) => void;
 export const __wbg_xfaengine_free: (a: number, b: number) => void;
+export const decryptPdf: (a: number, b: number, c: number, d: number, e: number) => void;
+export const encryptPdf: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+export const extractImages: (a: number, b: number, c: number) => void;
 export const pdfdoc_addFreeText: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => void;
 export const pdfdoc_addHighlight: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number) => void;
 export const pdfdoc_addStickyNote: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => void;
 export const pdfdoc_addTextWatermark: (a: number, b: number, c: number, d: number, e: number) => void;
+export const pdfdoc_addWatermark: (a: number, b: number, c: number) => void;
 export const pdfdoc_compress: (a: number, b: number) => void;
 export const pdfdoc_convertToPdfa: (a: number, b: number, c: number, d: number) => void;
 export const pdfdoc_deletePages: (a: number, b: number, c: number, d: number) => void;
 export const pdfdoc_dssInfo: (a: number, b: number) => void;
+export const pdfdoc_extractBlocks: (a: number, b: number, c: number, d: number) => void;
+export const pdfdoc_extractDocling: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
+export const pdfdoc_extractMarkdown: (a: number, b: number, c: number, d: number) => void;
 export const pdfdoc_extractPages: (a: number, b: number, c: number, d: number) => void;
 export const pdfdoc_flattenXfa: (a: number, b: number) => void;
 export const pdfdoc_getAnnotations: (a: number, b: number, c: number) => void;
@@ -64,6 +71,7 @@ export const pdfdocmut_setFormField: (a: number, b: number, c: number, d: number
 export const pdfdocmut_setFormFields: (a: number, b: number, c: number, d: number) => void;
 export const pdfdocmut_setMultiSelect: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
 export const pdfdocmut_setTextRunStyle: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+export const signPdf: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
 export const texteditor_findText: (a: number, b: number, c: number, d: number) => void;
 export const texteditor_open: (a: number, b: number, c: number) => void;
 export const texteditor_pageCount: (a: number) => number;