pdf-oxide-wasm 0.3.13 → 0.3.14

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pdf-oxide-wasm",
-  "version": "0.3.13",
+  "version": "0.3.14",
   "description": "Fast, zero-dependency PDF toolkit for Node.js, browsers, and edge runtimes — text extraction, markdown/HTML conversion, search, form filling, creation, and editing. Rust core compiled to WebAssembly.",
   "license": "MIT OR Apache-2.0",
   "repository": {

package/pdf_oxide.d.ts

@@ -1,6 +1,24 @@
 /* tslint:disable */
 /* eslint-disable */
 
+/**
+ * OCR configuration for WebAssembly.
+ */
+export class WasmOcrConfig {
+    free(): void;
+    [Symbol.dispose](): void;
+    constructor();
+}
+
+/**
+ * OCR engine for WebAssembly.
+ */
+export class WasmOcrEngine {
+    free(): void;
+    [Symbol.dispose](): void;
+    constructor(_det_model_path: string, _rec_model_path: string, _dict_path: string, _config?: WasmOcrConfig | null);
+}
+
 /**
  * Create new PDF documents from Markdown, HTML, or plain text.
  *
@@ -128,8 +146,11 @@ export class WasmPdfDocument {
      *
      * Returns an array of objects with: char, bbox {x, y, width, height},
      * font_name, font_size, font_weight, is_italic, color {r, g, b}, etc.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      */
-    extractChars(page_index: number): any;
+    extractChars(page_index: number, region?: Float32Array | null): any;
     /**
      * Extract image bytes from a page as PNG data.
      *
@@ -141,28 +162,80 @@ export class WasmPdfDocument {
      *
      * Returns an array of objects with: width, height, color_space,
      * bits_per_component, bbox (if available). Does NOT return raw image bytes.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     */
+    extractImages(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Extract only straight lines from a page (v0.3.14).
+     *
+     * Identifies paths that form a single straight line segment.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     * @returns Array of path objects
      */
-    extractImages(page_index: number): any;
+    extractLines(page_index: number, region?: Float32Array | null): any;
     /**
      * Extract vector paths (lines, curves, shapes) from a page.
      *
      * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      * @returns Array of path objects with bbox, stroke_color, fill_color, etc.
      */
-    extractPaths(page_index: number): any;
+    extractPaths(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Extract only rectangles from a page (v0.3.14).
+     *
+     * Identifies paths that form axis-aligned rectangles.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     * @returns Array of path objects
+     */
+    extractRects(page_index: number, region?: Float32Array | null): any;
     /**
      * Extract span-level data from a page.
      *
      * Returns an array of objects with: text, bbox, font_name, font_size,
      * font_weight, is_italic, color, etc.
      */
-    extractSpans(page_index: number): any;
+    extractSpans(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Extract tables from a page (v0.3.14).
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     */
+    extractTables(page_index: number, region?: Float32Array | null): any;
     /**
      * Extract plain text from a single page.
      *
      * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     */
+    extractText(page_index: number, region?: Float32Array | null): string;
+    /**
+     * Extract text lines from a page.
+     *
+     * Returns an array of objects with: text, bbox, words (array of Word objects).
+     */
+    extractTextLines(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Extract text using OCR (optical character recognition).
+     *
+     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
+     * ONNX Runtime support for the web backend in the current implementation.
      */
-    extractText(page_index: number): string;
+    extractTextOcr(_page_index: number, _engine?: WasmOcrEngine | null): string;
+    /**
+     * Extract word-level data from a page.
+     *
+     * Returns an array of objects with: text, bbox, font_name, font_size,
+     * font_weight, is_italic, is_bold.
+     */
+    extractWords(page_index: number, region?: Float32Array | null): any;
     /**
      * Flatten all annotations in the document into page content.
      */
@@ -289,13 +362,6 @@ export class WasmPdfDocument {
     rotatePage(page_index: number, degrees: number): void;
     /**
      * Save with encryption and return the resulting PDF as bytes.
-     *
-     * @param user_password - Password required to open the document
-     * @param owner_password - Password for full access (defaults to user_password)
-     * @param allow_print - Allow printing (default: true)
-     * @param allow_copy - Allow copying text (default: true)
-     * @param allow_modify - Allow modifying (default: true)
-     * @param allow_annotate - Allow annotations (default: true)
      */
     saveEncryptedToBytes(user_password: string, owner_password?: string | null, allow_print?: boolean | null, allow_copy?: boolean | null, allow_modify?: boolean | null, allow_annotate?: boolean | null): Uint8Array;
     /**
@@ -395,6 +461,13 @@ export class WasmPdfDocument {
      * Get the PDF version as [major, minor].
      */
     version(): Uint8Array;
+    /**
+     * Focus extraction on a specific rectangular region of a page (v0.3.14).
+     *
+     * @param page_index - Zero-based page number
+     * @param region - [x, y, width, height] in points
+     */
+    within(page_index: number, region: Float32Array): WasmPdfPageRegion;
     /**
      * Get XMP metadata from the document.
      *
@@ -402,3 +475,52 @@ export class WasmPdfDocument {
      */
     xmpMetadata(): any;
 }
+
+/**
+ * A focused view of a PDF page region for scoped extraction (v0.3.14).
+ */
+export class WasmPdfPageRegion {
+    private constructor();
+    free(): void;
+    [Symbol.dispose](): void;
+    /**
+     * Extract character-level data from this region.
+     */
+    extractChars(): any;
+    /**
+     * Extract images from this region.
+     */
+    extractImages(): any;
+    /**
+     * Extract straight lines from this region.
+     */
+    extractLines(): any;
+    /**
+     * Extract vector paths from this region.
+     */
+    extractPaths(): any;
+    /**
+     * Extract rectangles from this region.
+     */
+    extractRects(): any;
+    /**
+     * Extract tables from this region.
+     */
+    extractTables(): any;
+    /**
+     * Extract text from this region.
+     */
+    extractText(): string;
+    /**
+     * Extract text lines from this region.
+     */
+    extractTextLines(): any;
+    /**
+     * Extract text using OCR from this region.
+     */
+    extractTextOcr(_engine?: WasmOcrEngine | null): string;
+    /**
+     * Extract words from this region.
+     */
+    extractWords(): any;
+}

package/pdf_oxide.js

@@ -1,5 +1,81 @@
 /* @ts-self-types="./pdf_oxide.d.ts" */
 
+/**
+ * OCR configuration for WebAssembly.
+ */
+class WasmOcrConfig {
+    __destroy_into_raw() {
+        const ptr = this.__wbg_ptr;
+        this.__wbg_ptr = 0;
+        WasmOcrConfigFinalization.unregister(this);
+        return ptr;
+    }
+    free() {
+        const ptr = this.__destroy_into_raw();
+        wasm.__wbg_wasmocrconfig_free(ptr, 0);
+    }
+    constructor() {
+        const ret = wasm.wasmocrconfig_new();
+        this.__wbg_ptr = ret >>> 0;
+        WasmOcrConfigFinalization.register(this, this.__wbg_ptr, this);
+        return this;
+    }
+}
+if (Symbol.dispose) WasmOcrConfig.prototype[Symbol.dispose] = WasmOcrConfig.prototype.free;
+exports.WasmOcrConfig = WasmOcrConfig;
+
+/**
+ * OCR engine for WebAssembly.
+ */
+class WasmOcrEngine {
+    __destroy_into_raw() {
+        const ptr = this.__wbg_ptr;
+        this.__wbg_ptr = 0;
+        WasmOcrEngineFinalization.unregister(this);
+        return ptr;
+    }
+    free() {
+        const ptr = this.__destroy_into_raw();
+        wasm.__wbg_wasmocrengine_free(ptr, 0);
+    }
+    /**
+     * @param {string} _det_model_path
+     * @param {string} _rec_model_path
+     * @param {string} _dict_path
+     * @param {WasmOcrConfig | null} [_config]
+     */
+    constructor(_det_model_path, _rec_model_path, _dict_path, _config) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            const ptr0 = passStringToWasm0(_det_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const len0 = WASM_VECTOR_LEN;
+            const ptr1 = passStringToWasm0(_rec_model_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const len1 = WASM_VECTOR_LEN;
+            const ptr2 = passStringToWasm0(_dict_path, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            const len2 = WASM_VECTOR_LEN;
+            let ptr3 = 0;
+            if (!isLikeNone(_config)) {
+                _assertClass(_config, WasmOcrConfig);
+                ptr3 = _config.__destroy_into_raw();
+            }
+            wasm.wasmocrengine_new(retptr, ptr0, len0, ptr1, len1, ptr2, len2, ptr3);
+            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);
+            }
+            this.__wbg_ptr = r0 >>> 0;
+            WasmOcrEngineFinalization.register(this, this.__wbg_ptr, this);
+            return this;
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+        }
+    }
+}
+if (Symbol.dispose) WasmOcrEngine.prototype[Symbol.dispose] = WasmOcrEngine.prototype.free;
+exports.WasmOcrEngine = WasmOcrEngine;
+
 /**
  * Create new PDF documents from Markdown, HTML, or plain text.
  *
@@ -442,13 +518,19 @@ class WasmPdfDocument {
      *
      * Returns an array of objects with: char, bbox {x, y, width, height},
      * font_name, font_size, font_weight, is_italic, color {r, g, b}, etc.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      * @param {number} page_index
+     * @param {Float32Array | null} [region]
      * @returns {any}
      */
-    extractChars(page_index) {
+    extractChars(page_index, region) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            wasm.wasmpdfdocument_extractChars(retptr, this.__wbg_ptr, page_index);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractChars(retptr, this.__wbg_ptr, page_index, 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);
@@ -487,13 +569,48 @@ class WasmPdfDocument {
      *
      * Returns an array of objects with: width, height, color_space,
      * bits_per_component, bbox (if available). Does NOT return raw image bytes.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      * @param {number} page_index
+     * @param {Float32Array | null} [region]
      * @returns {any}
      */
-    extractImages(page_index) {
+    extractImages(page_index, region) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            wasm.wasmpdfdocument_extractImages(retptr, this.__wbg_ptr, page_index);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractImages(retptr, this.__wbg_ptr, page_index, 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);
+        }
+    }
+    /**
+     * Extract only straight lines from a page (v0.3.14).
+     *
+     * Identifies paths that form a single straight line segment.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     * @returns Array of path objects
+     * @param {number} page_index
+     * @param {Float32Array | null} [region]
+     * @returns {any}
+     */
+    extractLines(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractLines(retptr, this.__wbg_ptr, page_index, 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);
@@ -509,14 +626,47 @@ class WasmPdfDocument {
      * Extract vector paths (lines, curves, shapes) from a page.
      *
      * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      * @returns Array of path objects with bbox, stroke_color, fill_color, etc.
      * @param {number} page_index
+     * @param {Float32Array | null} [region]
      * @returns {any}
      */
-    extractPaths(page_index) {
+    extractPaths(page_index, region) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            wasm.wasmpdfdocument_extractPaths(retptr, this.__wbg_ptr, page_index);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractPaths(retptr, this.__wbg_ptr, page_index, 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);
+        }
+    }
+    /**
+     * Extract only rectangles from a page (v0.3.14).
+     *
+     * Identifies paths that form axis-aligned rectangles.
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     * @returns Array of path objects
+     * @param {number} page_index
+     * @param {Float32Array | null} [region]
+     * @returns {any}
+     */
+    extractRects(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractRects(retptr, this.__wbg_ptr, page_index, 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);
@@ -534,12 +684,41 @@ class WasmPdfDocument {
      * Returns an array of objects with: text, bbox, font_name, font_size,
      * font_weight, is_italic, color, etc.
      * @param {number} page_index
+     * @param {Float32Array | null} [region]
      * @returns {any}
      */
-    extractSpans(page_index) {
+    extractSpans(page_index, region) {
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            wasm.wasmpdfdocument_extractSpans(retptr, this.__wbg_ptr, page_index);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractSpans(retptr, this.__wbg_ptr, page_index, 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);
+        }
+    }
+    /**
+     * Extract tables from a page (v0.3.14).
+     *
+     * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
+     * @param {number} page_index
+     * @param {Float32Array | null} [region]
+     * @returns {any}
+     */
+    extractTables(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractTables(retptr, this.__wbg_ptr, page_index, 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);
@@ -555,31 +734,124 @@ class WasmPdfDocument {
      * Extract plain text from a single page.
      *
      * @param page_index - Zero-based page number
+     * @param region - Optional [x, y, width, height] to filter by
      * @param {number} page_index
+     * @param {Float32Array | null} [region]
      * @returns {string}
      */
-    extractText(page_index) {
-        let deferred2_0;
-        let deferred2_1;
+    extractText(page_index, region) {
+        let deferred3_0;
+        let deferred3_1;
         try {
             const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
-            wasm.wasmpdfdocument_extractText(retptr, this.__wbg_ptr, page_index);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractText(retptr, this.__wbg_ptr, page_index, 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 ptr1 = r0;
-            var len1 = r1;
+            var ptr2 = r0;
+            var len2 = r1;
             if (r3) {
-                ptr1 = 0; len1 = 0;
+                ptr2 = 0; len2 = 0;
                 throw takeObject(r2);
             }
-            deferred2_0 = ptr1;
-            deferred2_1 = len1;
-            return getStringFromWasm0(ptr1, len1);
+            deferred3_0 = ptr2;
+            deferred3_1 = len2;
+            return getStringFromWasm0(ptr2, len2);
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+            wasm.__wbindgen_export3(deferred3_0, deferred3_1, 1);
+        }
+    }
+    /**
+     * Extract text lines from a page.
+     *
+     * Returns an array of objects with: text, bbox, words (array of Word objects).
+     * @param {number} page_index
+     * @param {Float32Array | null} [region]
+     * @returns {any}
+     */
+    extractTextLines(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractTextLines(retptr, this.__wbg_ptr, page_index, 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);
+        }
+    }
+    /**
+     * Extract text using OCR (optical character recognition).
+     *
+     * NOTE: OCR is not yet supported in the WebAssembly build due to missing
+     * ONNX Runtime support for the web backend in the current implementation.
+     * @param {number} _page_index
+     * @param {WasmOcrEngine | null} [_engine]
+     * @returns {string}
+     */
+    extractTextOcr(_page_index, _engine) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            let ptr0 = 0;
+            if (!isLikeNone(_engine)) {
+                _assertClass(_engine, WasmOcrEngine);
+                ptr0 = _engine.__destroy_into_raw();
+            }
+            wasm.wasmpdfdocument_extractTextOcr(retptr, this.__wbg_ptr, _page_index, ptr0);
+            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_export3(deferred3_0, deferred3_1, 1);
+        }
+    }
+    /**
+     * Extract word-level data from a page.
+     *
+     * Returns an array of objects with: text, bbox, font_name, font_size,
+     * font_weight, is_italic, is_bold.
+     * @param {number} page_index
+     * @param {Float32Array | null} [region]
+     * @returns {any}
+     */
+    extractWords(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(region) ? 0 : passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractWords(retptr, this.__wbg_ptr, page_index, 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);
-            wasm.__wbindgen_export3(deferred2_0, deferred2_1, 1);
         }
     }
     /**
@@ -757,8 +1029,19 @@ class WasmPdfDocument {
      * @returns {boolean}
      */
     hasStructureTree() {
-        const ret = wasm.wasmpdfdocument_hasStructureTree(this.__wbg_ptr);
-        return ret !== 0;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_hasStructureTree(retptr, this.__wbg_ptr);
+            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 r0 !== 0;
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+        }
     }
     /**
      * Check if the document contains XFA form data.
@@ -1038,13 +1321,6 @@ class WasmPdfDocument {
     }
     /**
      * Save with encryption and return the resulting PDF as bytes.
-     *
-     * @param user_password - Password required to open the document
-     * @param owner_password - Password for full access (defaults to user_password)
-     * @param allow_print - Allow printing (default: true)
-     * @param allow_copy - Allow copying text (default: true)
-     * @param allow_modify - Allow modifying (default: true)
-     * @param allow_annotate - Allow annotations (default: true)
      * @param {string} user_password
      * @param {string | null} [owner_password]
      * @param {boolean | null} [allow_print]
@@ -1544,6 +1820,11 @@ class WasmPdfDocument {
             wasm.wasmpdfdocument_version(retptr, this.__wbg_ptr);
             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_export3(r0, r1 * 1, 1);
             return v1;
@@ -1551,6 +1832,32 @@ class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Focus extraction on a specific rectangular region of a page (v0.3.14).
+     *
+     * @param page_index - Zero-based page number
+     * @param region - [x, y, width, height] in points
+     * @param {number} page_index
+     * @param {Float32Array} region
+     * @returns {WasmPdfPageRegion}
+     */
+    within(page_index, region) {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            const ptr0 = passArrayF32ToWasm0(region, wasm.__wbindgen_export);
+            const len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_within(retptr, this.__wbg_ptr, page_index, 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 WasmPdfPageRegion.__wrap(r0);
+        } finally {
+            wasm.__wbindgen_add_to_stack_pointer(16);
+        }
+    }
     /**
      * Get XMP metadata from the document.
      *
@@ -1576,6 +1883,245 @@ class WasmPdfDocument {
 if (Symbol.dispose) WasmPdfDocument.prototype[Symbol.dispose] = WasmPdfDocument.prototype.free;
 exports.WasmPdfDocument = WasmPdfDocument;
 
+/**
+ * A focused view of a PDF page region for scoped extraction (v0.3.14).
+ */
+class WasmPdfPageRegion {
+    static __wrap(ptr) {
+        ptr = ptr >>> 0;
+        const obj = Object.create(WasmPdfPageRegion.prototype);
+        obj.__wbg_ptr = ptr;
+        WasmPdfPageRegionFinalization.register(obj, obj.__wbg_ptr, obj);
+        return obj;
+    }
+    __destroy_into_raw() {
+        const ptr = this.__wbg_ptr;
+        this.__wbg_ptr = 0;
+        WasmPdfPageRegionFinalization.unregister(this);
+        return ptr;
+    }
+    free() {
+        const ptr = this.__destroy_into_raw();
+        wasm.__wbg_wasmpdfpageregion_free(ptr, 0);
+    }
+    /**
+     * Extract character-level data from this region.
+     * @returns {any}
+     */
+    extractChars() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractChars(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract images from this region.
+     * @returns {any}
+     */
+    extractImages() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractImages(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract straight lines from this region.
+     * @returns {any}
+     */
+    extractLines() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractLines(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract vector paths from this region.
+     * @returns {any}
+     */
+    extractPaths() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractPaths(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract rectangles from this region.
+     * @returns {any}
+     */
+    extractRects() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractRects(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract tables from this region.
+     * @returns {any}
+     */
+    extractTables() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractTables(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract text from this region.
+     * @returns {string}
+     */
+    extractText() {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractText(retptr, this.__wbg_ptr);
+            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_export3(deferred2_0, deferred2_1, 1);
+        }
+    }
+    /**
+     * Extract text lines from this region.
+     * @returns {any}
+     */
+    extractTextLines() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractTextLines(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+    /**
+     * Extract text using OCR from this region.
+     * @param {WasmOcrEngine | null} [_engine]
+     * @returns {string}
+     */
+    extractTextOcr(_engine) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            let ptr0 = 0;
+            if (!isLikeNone(_engine)) {
+                _assertClass(_engine, WasmOcrEngine);
+                ptr0 = _engine.__destroy_into_raw();
+            }
+            wasm.wasmpdfpageregion_extractTextOcr(retptr, this.__wbg_ptr, ptr0);
+            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_export3(deferred3_0, deferred3_1, 1);
+        }
+    }
+    /**
+     * Extract words from this region.
+     * @returns {any}
+     */
+    extractWords() {
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfpageregion_extractWords(retptr, this.__wbg_ptr);
+            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);
+        }
+    }
+}
+if (Symbol.dispose) WasmPdfPageRegion.prototype[Symbol.dispose] = WasmPdfPageRegion.prototype.free;
+exports.WasmPdfPageRegion = WasmPdfPageRegion;
+
 function __wbg_get_imports() {
     const import0 = {
         __proto__: null,
@@ -1742,12 +2288,21 @@ function __wbg_get_imports() {
     };
 }
 
+const WasmOcrConfigFinalization = (typeof FinalizationRegistry === 'undefined')
+    ? { register: () => {}, unregister: () => {} }
+    : new FinalizationRegistry(ptr => wasm.__wbg_wasmocrconfig_free(ptr >>> 0, 1));
+const WasmOcrEngineFinalization = (typeof FinalizationRegistry === 'undefined')
+    ? { register: () => {}, unregister: () => {} }
+    : new FinalizationRegistry(ptr => wasm.__wbg_wasmocrengine_free(ptr >>> 0, 1));
 const WasmPdfFinalization = (typeof FinalizationRegistry === 'undefined')
     ? { register: () => {}, unregister: () => {} }
     : new FinalizationRegistry(ptr => wasm.__wbg_wasmpdf_free(ptr >>> 0, 1));
 const WasmPdfDocumentFinalization = (typeof FinalizationRegistry === 'undefined')
     ? { register: () => {}, unregister: () => {} }
     : new FinalizationRegistry(ptr => wasm.__wbg_wasmpdfdocument_free(ptr >>> 0, 1));
+const WasmPdfPageRegionFinalization = (typeof FinalizationRegistry === 'undefined')
+    ? { register: () => {}, unregister: () => {} }
+    : new FinalizationRegistry(ptr => wasm.__wbg_wasmpdfpageregion_free(ptr >>> 0, 1));
 
 function addHeapObject(obj) {
     if (heap_next === heap.length) heap.push(heap.length + 1);
@@ -1758,6 +2313,12 @@ function addHeapObject(obj) {
     return idx;
 }
 
+function _assertClass(instance, klass) {
+    if (!(instance instanceof klass)) {
+        throw new Error(`expected instance of ${klass.name}`);
+    }
+}
+
 function dropObject(idx) {
     if (idx < 1028) return;
     heap[idx] = heap_next;

package/pdf_oxide_bg.wasm.d.ts

@@ -1,8 +1,11 @@
 /* tslint:disable */
 /* eslint-disable */
 export const memory: WebAssembly.Memory;
+export const __wbg_wasmocrconfig_free: (a: number, b: number) => void;
 export const __wbg_wasmpdf_free: (a: number, b: number) => void;
 export const __wbg_wasmpdfdocument_free: (a: number, b: number) => void;
+export const __wbg_wasmpdfpageregion_free: (a: number, b: number) => void;
+export const wasmocrengine_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
 export const wasmpdf_fromHtml: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => void;
 export const wasmpdf_fromImageBytes: (a: number, b: number, c: number) => void;
 export const wasmpdf_fromMarkdown: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => void;
@@ -20,12 +23,18 @@ export const wasmpdfdocument_eraseRegion: (a: number, b: number, c: number, d: n
 export const wasmpdfdocument_eraseRegions: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_exportFormData: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_extractAllText: (a: number, b: number) => void;
-export const wasmpdfdocument_extractChars: (a: number, b: number, c: number) => void;
+export const wasmpdfdocument_extractChars: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractImageBytes: (a: number, b: number, c: number) => void;
-export const wasmpdfdocument_extractImages: (a: number, b: number, c: number) => void;
-export const wasmpdfdocument_extractPaths: (a: number, b: number, c: number) => void;
-export const wasmpdfdocument_extractSpans: (a: number, b: number, c: number) => void;
-export const wasmpdfdocument_extractText: (a: number, b: number, c: number) => void;
+export const wasmpdfdocument_extractImages: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractLines: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractPaths: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractRects: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractSpans: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractTables: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractText: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractTextLines: (a: number, b: number, c: number, d: number, e: number) => void;
+export const wasmpdfdocument_extractTextOcr: (a: number, b: number, c: number, d: number) => void;
+export const wasmpdfdocument_extractWords: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_flattenAllAnnotations: (a: number, b: number) => void;
 export const wasmpdfdocument_flattenForms: (a: number, b: number) => void;
 export const wasmpdfdocument_flattenFormsOnPage: (a: number, b: number, c: number) => void;
@@ -34,7 +43,7 @@ export const wasmpdfdocument_getAnnotations: (a: number, b: number, c: number) =
 export const wasmpdfdocument_getFormFieldValue: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_getFormFields: (a: number, b: number) => void;
 export const wasmpdfdocument_getOutline: (a: number, b: number) => void;
-export const wasmpdfdocument_hasStructureTree: (a: number) => number;
+export const wasmpdfdocument_hasStructureTree: (a: number, b: number) => void;
 export const wasmpdfdocument_hasXfa: (a: number, b: number) => void;
 export const wasmpdfdocument_mergeFrom: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_new: (a: number, b: number, c: number) => void;
@@ -68,7 +77,20 @@ export const wasmpdfdocument_toMarkdownAll: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_toPlainText: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_toPlainTextAll: (a: number, b: number) => void;
 export const wasmpdfdocument_version: (a: number, b: number) => void;
+export const wasmpdfdocument_within: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_xmpMetadata: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractChars: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractImages: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractLines: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractPaths: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractRects: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractTables: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractText: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractTextLines: (a: number, b: number) => void;
+export const wasmpdfpageregion_extractTextOcr: (a: number, b: number, c: number) => void;
+export const wasmpdfpageregion_extractWords: (a: number, b: number) => void;
+export const __wbg_wasmocrengine_free: (a: number, b: number) => void;
+export const wasmocrconfig_new: () => number;
 export const __wbindgen_export: (a: number, b: number) => number;
 export const __wbindgen_export2: (a: number, b: number, c: number, d: number) => number;
 export const __wbindgen_export3: (a: number, b: number, c: number) => void;