pdf-oxide-wasm 0.3.50 → 0.3.51

package/bundler/pdf_oxide.d.ts

@@ -815,6 +815,15 @@ export class WasmPdfDocument {
      * @returns true if authentication succeeded
      */
     authenticate(password: string): boolean;
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     */
+    classifyDocument(): string;
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     */
+    classifyPage(page_index: number): string;
     /**
      * Clear all pending erase operations for a page.
      */
@@ -934,6 +943,12 @@ export class WasmPdfDocument {
      * @returns Array of path objects
      */
     extractLines(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     */
+    extractPageAuto(page_index: number, options_json?: string | null): string;
     /**
      * Extract complete page text data in a single call.
      *
@@ -990,6 +1005,11 @@ export class WasmPdfDocument {
      * @param region - Optional [x, y, width, height] to filter by
      */
     extractText(page_index: number, region: any): string;
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     */
+    extractTextAuto(page_index: number): string;
     /**
      * Extract text lines from a page.
      *
@@ -1600,6 +1620,20 @@ export function generateQrSvg(data: string, error_correction: number, size: numb
  */
 export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ */
+export function modelManifest(): string;
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -1607,6 +1641,13 @@ export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
  */
 export function planSplitByBookmarks(src_bytes: Uint8Array, title_prefix: string | null | undefined, ignore_case: boolean, level: number, include_front_matter: boolean): any;
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ */
+export function prefetchAvailable(): boolean;
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:

package/bundler/pdf_oxide.js

@@ -5,5 +5,5 @@ import { __wbg_set_wasm } from "./pdf_oxide_bg.js";
 __wbg_set_wasm(wasm);
 
 export {
-    Align, ArtifactStyle, Dss, PadesLevel, RevocationMaterial, StreamingTable, WasmArtifact, WasmCertificate, WasmDocumentBuilder, WasmEmbeddedFont, WasmFluentPageBuilder, WasmFooter, WasmHeader, WasmOcrConfig, WasmOcrEngine, WasmPageTemplate, WasmPdf, WasmPdfDocument, WasmPdfPageRegion, WasmSignature, WasmTimestamp, cryptoCbom, cryptoInventory, cryptoPolicy, disableLogging, generateBarcodeSvg, generateQrSvg, hasDocumentTimestamp, planSplitByBookmarks, setCryptoPolicy, setLogLevel, signPdfBytes, signPdfBytesPades, splitByBookmarks
+    Align, ArtifactStyle, Dss, PadesLevel, RevocationMaterial, StreamingTable, WasmArtifact, WasmCertificate, WasmDocumentBuilder, WasmEmbeddedFont, WasmFluentPageBuilder, WasmFooter, WasmHeader, WasmOcrConfig, WasmOcrEngine, WasmPageTemplate, WasmPdf, WasmPdfDocument, WasmPdfPageRegion, WasmSignature, WasmTimestamp, cryptoCbom, cryptoInventory, cryptoPolicy, disableLogging, generateBarcodeSvg, generateQrSvg, hasDocumentTimestamp, modelManifest, planSplitByBookmarks, prefetchAvailable, setCryptoPolicy, setLogLevel, signPdfBytes, signPdfBytesPades, splitByBookmarks
 } from "./pdf_oxide_bg.js";

package/bundler/pdf_oxide_bg.js

@@ -3125,6 +3125,64 @@ export class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     * @returns {string}
+     */
+    classifyDocument() {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyDocument(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_export4(deferred2_0, deferred2_1, 1);
+        }
+    }
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     * @param {number} page_index
+     * @returns {string}
+     */
+    classifyPage(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyPage(retptr, this.__wbg_ptr, page_index);
+            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);
+        }
+    }
     /**
      * Clear all pending erase operations for a page.
      * @param {number} page_index
@@ -3553,6 +3611,40 @@ export class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     * @param {number} page_index
+     * @param {string | null} [options_json]
+     * @returns {string}
+     */
+    extractPageAuto(page_index, options_json) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(options_json) ? 0 : passStringToWasm0(options_json, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractPageAuto(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 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);
+        }
+    }
     /**
      * Extract complete page text data in a single call.
      *
@@ -3754,6 +3846,36 @@ export class WasmPdfDocument {
             wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     * @param {number} page_index
+     * @returns {string}
+     */
+    extractTextAuto(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_extractTextAuto(retptr, this.__wbg_ptr, page_index);
+            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 text lines from a page.
      *
@@ -6153,6 +6275,36 @@ export function hasDocumentTimestamp(pdf_data) {
     return ret !== 0;
 }
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ * @returns {string}
+ */
+export function modelManifest() {
+    let deferred1_0;
+    let deferred1_1;
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        wasm.modelManifest(retptr);
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        deferred1_0 = r0;
+        deferred1_1 = r1;
+        return getStringFromWasm0(r0, r1);
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+        wasm.__wbindgen_export4(deferred1_0, deferred1_1, 1);
+    }
+}
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -6184,6 +6336,17 @@ export function planSplitByBookmarks(src_bytes, title_prefix, ignore_case, level
     }
 }
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ * @returns {boolean}
+ */
+export function prefetchAvailable() {
+    const ret = wasm.prefetchAvailable();
+    return ret !== 0;
+}
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:

package/bundler/pdf_oxide_bg.wasm.d.ts

@@ -23,7 +23,9 @@ export const cryptoPolicy: (a: number) => void;
 export const generateBarcodeSvg: (a: number, b: number, c: number, d: number) => void;
 export const generateQrSvg: (a: number, b: number, c: number, d: number, e: number) => void;
 export const hasDocumentTimestamp: (a: number, b: number) => number;
+export const modelManifest: (a: number) => void;
 export const planSplitByBookmarks: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+export const prefetchAvailable: () => number;
 export const setCryptoPolicy: (a: number, b: number, c: number) => void;
 export const setLogLevel: (a: number, b: number, c: number) => void;
 export const signPdfBytes: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
@@ -158,6 +160,8 @@ export const wasmpdfdocument_applyAllRedactions: (a: number, b: number) => void;
 export const wasmpdfdocument_applyPageRedactions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_applyRedactionsDestructive: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_authenticate: (a: number, b: number, c: number, d: number) => void;
+export const wasmpdfdocument_classifyDocument: (a: number, b: number) => void;
+export const wasmpdfdocument_classifyPage: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_clearEraseRegions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_convertToPdfA: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_cropMargins: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
@@ -177,6 +181,7 @@ export const wasmpdfdocument_extractChars: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractImageBytes: (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_extractPageAuto: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPageText: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPages: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_extractPaths: (a: number, b: number, c: number, d: number, e: number) => void;
@@ -184,6 +189,7 @@ export const wasmpdfdocument_extractRects: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractSpans: (a: number, b: number, c: number, d: number, e: number, f: number, g: 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) => void;
+export const wasmpdfdocument_extractTextAuto: (a: number, b: number, c: 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;

package/nodejs/pdf_oxide.d.ts

@@ -815,6 +815,15 @@ export class WasmPdfDocument {
      * @returns true if authentication succeeded
      */
     authenticate(password: string): boolean;
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     */
+    classifyDocument(): string;
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     */
+    classifyPage(page_index: number): string;
     /**
      * Clear all pending erase operations for a page.
      */
@@ -934,6 +943,12 @@ export class WasmPdfDocument {
      * @returns Array of path objects
      */
     extractLines(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     */
+    extractPageAuto(page_index: number, options_json?: string | null): string;
     /**
      * Extract complete page text data in a single call.
      *
@@ -990,6 +1005,11 @@ export class WasmPdfDocument {
      * @param region - Optional [x, y, width, height] to filter by
      */
     extractText(page_index: number, region: any): string;
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     */
+    extractTextAuto(page_index: number): string;
     /**
      * Extract text lines from a page.
      *
@@ -1600,6 +1620,20 @@ export function generateQrSvg(data: string, error_correction: number, size: numb
  */
 export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ */
+export function modelManifest(): string;
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -1607,6 +1641,13 @@ export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
  */
 export function planSplitByBookmarks(src_bytes: Uint8Array, title_prefix: string | null | undefined, ignore_case: boolean, level: number, include_front_matter: boolean): any;
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ */
+export function prefetchAvailable(): boolean;
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:

package/nodejs/pdf_oxide.js

@@ -3144,6 +3144,64 @@ class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     * @returns {string}
+     */
+    classifyDocument() {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyDocument(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_export4(deferred2_0, deferred2_1, 1);
+        }
+    }
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     * @param {number} page_index
+     * @returns {string}
+     */
+    classifyPage(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyPage(retptr, this.__wbg_ptr, page_index);
+            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);
+        }
+    }
     /**
      * Clear all pending erase operations for a page.
      * @param {number} page_index
@@ -3572,6 +3630,40 @@ class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     * @param {number} page_index
+     * @param {string | null} [options_json]
+     * @returns {string}
+     */
+    extractPageAuto(page_index, options_json) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(options_json) ? 0 : passStringToWasm0(options_json, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractPageAuto(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 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);
+        }
+    }
     /**
      * Extract complete page text data in a single call.
      *
@@ -3773,6 +3865,36 @@ class WasmPdfDocument {
             wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     * @param {number} page_index
+     * @returns {string}
+     */
+    extractTextAuto(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_extractTextAuto(retptr, this.__wbg_ptr, page_index);
+            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 text lines from a page.
      *
@@ -6183,6 +6305,37 @@ function hasDocumentTimestamp(pdf_data) {
 }
 exports.hasDocumentTimestamp = hasDocumentTimestamp;
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ * @returns {string}
+ */
+function modelManifest() {
+    let deferred1_0;
+    let deferred1_1;
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        wasm.modelManifest(retptr);
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        deferred1_0 = r0;
+        deferred1_1 = r1;
+        return getStringFromWasm0(r0, r1);
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+        wasm.__wbindgen_export4(deferred1_0, deferred1_1, 1);
+    }
+}
+exports.modelManifest = modelManifest;
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -6215,6 +6368,18 @@ function planSplitByBookmarks(src_bytes, title_prefix, ignore_case, level, inclu
 }
 exports.planSplitByBookmarks = planSplitByBookmarks;
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ * @returns {boolean}
+ */
+function prefetchAvailable() {
+    const ret = wasm.prefetchAvailable();
+    return ret !== 0;
+}
+exports.prefetchAvailable = prefetchAvailable;
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:

package/nodejs/pdf_oxide_bg.wasm.d.ts

@@ -23,7 +23,9 @@ export const cryptoPolicy: (a: number) => void;
 export const generateBarcodeSvg: (a: number, b: number, c: number, d: number) => void;
 export const generateQrSvg: (a: number, b: number, c: number, d: number, e: number) => void;
 export const hasDocumentTimestamp: (a: number, b: number) => number;
+export const modelManifest: (a: number) => void;
 export const planSplitByBookmarks: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+export const prefetchAvailable: () => number;
 export const setCryptoPolicy: (a: number, b: number, c: number) => void;
 export const setLogLevel: (a: number, b: number, c: number) => void;
 export const signPdfBytes: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
@@ -158,6 +160,8 @@ export const wasmpdfdocument_applyAllRedactions: (a: number, b: number) => void;
 export const wasmpdfdocument_applyPageRedactions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_applyRedactionsDestructive: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_authenticate: (a: number, b: number, c: number, d: number) => void;
+export const wasmpdfdocument_classifyDocument: (a: number, b: number) => void;
+export const wasmpdfdocument_classifyPage: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_clearEraseRegions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_convertToPdfA: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_cropMargins: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
@@ -177,6 +181,7 @@ export const wasmpdfdocument_extractChars: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractImageBytes: (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_extractPageAuto: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPageText: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPages: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_extractPaths: (a: number, b: number, c: number, d: number, e: number) => void;
@@ -184,6 +189,7 @@ export const wasmpdfdocument_extractRects: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractSpans: (a: number, b: number, c: number, d: number, e: number, f: number, g: 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) => void;
+export const wasmpdfdocument_extractTextAuto: (a: number, b: number, c: 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;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pdf-oxide-wasm",
-  "version": "0.3.50",
+  "version": "0.3.51",
   "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/web/pdf_oxide.d.ts

@@ -815,6 +815,15 @@ export class WasmPdfDocument {
      * @returns true if authentication succeeded
      */
     authenticate(password: string): boolean;
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     */
+    classifyDocument(): string;
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     */
+    classifyPage(page_index: number): string;
     /**
      * Clear all pending erase operations for a page.
      */
@@ -934,6 +943,12 @@ export class WasmPdfDocument {
      * @returns Array of path objects
      */
     extractLines(page_index: number, region?: Float32Array | null): any;
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     */
+    extractPageAuto(page_index: number, options_json?: string | null): string;
     /**
      * Extract complete page text data in a single call.
      *
@@ -990,6 +1005,11 @@ export class WasmPdfDocument {
      * @param region - Optional [x, y, width, height] to filter by
      */
     extractText(page_index: number, region: any): string;
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     */
+    extractTextAuto(page_index: number): string;
     /**
      * Extract text lines from a page.
      *
@@ -1600,6 +1620,20 @@ export function generateQrSvg(data: string, error_correction: number, size: numb
  */
 export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ */
+export function modelManifest(): string;
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -1607,6 +1641,13 @@ export function hasDocumentTimestamp(pdf_data: Uint8Array): boolean;
  */
 export function planSplitByBookmarks(src_bytes: Uint8Array, title_prefix: string | null | undefined, ignore_case: boolean, level: number, include_front_matter: boolean): any;
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ */
+export function prefetchAvailable(): boolean;
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:
@@ -1687,7 +1728,9 @@ export interface InitOutput {
     readonly generateBarcodeSvg: (a: number, b: number, c: number, d: number) => void;
     readonly generateQrSvg: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly hasDocumentTimestamp: (a: number, b: number) => number;
+    readonly modelManifest: (a: number) => void;
     readonly planSplitByBookmarks: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+    readonly prefetchAvailable: () => number;
     readonly setCryptoPolicy: (a: number, b: number, c: number) => void;
     readonly setLogLevel: (a: number, b: number, c: number) => void;
     readonly signPdfBytes: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
@@ -1822,6 +1865,8 @@ export interface InitOutput {
     readonly wasmpdfdocument_applyPageRedactions: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_applyRedactionsDestructive: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_authenticate: (a: number, b: number, c: number, d: number) => void;
+    readonly wasmpdfdocument_classifyDocument: (a: number, b: number) => void;
+    readonly wasmpdfdocument_classifyPage: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_clearEraseRegions: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_convertToPdfA: (a: number, b: number, c: number, d: number) => void;
     readonly wasmpdfdocument_cropMargins: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
@@ -1841,6 +1886,7 @@ export interface InitOutput {
     readonly wasmpdfdocument_extractImageBytes: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_extractImages: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly wasmpdfdocument_extractLines: (a: number, b: number, c: number, d: number, e: number) => void;
+    readonly wasmpdfdocument_extractPageAuto: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly wasmpdfdocument_extractPageText: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly wasmpdfdocument_extractPages: (a: number, b: number, c: number, d: number) => void;
     readonly wasmpdfdocument_extractPaths: (a: number, b: number, c: number, d: number, e: number) => void;
@@ -1848,6 +1894,7 @@ export interface InitOutput {
     readonly wasmpdfdocument_extractSpans: (a: number, b: number, c: number, d: number, e: number, f: number, g: number) => void;
     readonly wasmpdfdocument_extractTables: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly wasmpdfdocument_extractText: (a: number, b: number, c: number, d: number) => void;
+    readonly wasmpdfdocument_extractTextAuto: (a: number, b: number, c: number) => void;
     readonly wasmpdfdocument_extractTextLines: (a: number, b: number, c: number, d: number, e: number) => void;
     readonly wasmpdfdocument_extractTextOcr: (a: number, b: number, c: number, d: number) => void;
     readonly wasmpdfdocument_extractWords: (a: number, b: number, c: number, d: number, e: number) => void;

package/web/pdf_oxide.js

@@ -3127,6 +3127,64 @@ export class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Cheap per-page text-vs-OCR classification → JSON
+     * `DocumentClassification`.
+     * @returns {string}
+     */
+    classifyDocument() {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyDocument(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_export4(deferred2_0, deferred2_1, 1);
+        }
+    }
+    /**
+     * Cheap per-page classification → JSON `PageClassification`.
+     * @param {number} page_index
+     * @returns {string}
+     */
+    classifyPage(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_classifyPage(retptr, this.__wbg_ptr, page_index);
+            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);
+        }
+    }
     /**
      * Clear all pending erase operations for a page.
      * @param {number} page_index
@@ -3555,6 +3613,40 @@ export class WasmPdfDocument {
             wasm.__wbindgen_add_to_stack_pointer(16);
         }
     }
+    /**
+     * Rich per-page extraction → JSON `PageExtraction` (per-region
+     * bbox + typed reason). `optionsJson` is `{}`-tolerant
+     * `AutoExtractOptions`; undefined/empty → defaults.
+     * @param {number} page_index
+     * @param {string | null} [options_json]
+     * @returns {string}
+     */
+    extractPageAuto(page_index, options_json) {
+        let deferred3_0;
+        let deferred3_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            var ptr0 = isLikeNone(options_json) ? 0 : passStringToWasm0(options_json, wasm.__wbindgen_export, wasm.__wbindgen_export2);
+            var len0 = WASM_VECTOR_LEN;
+            wasm.wasmpdfdocument_extractPageAuto(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 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);
+        }
+    }
     /**
      * Extract complete page text data in a single call.
      *
@@ -3756,6 +3848,36 @@ export class WasmPdfDocument {
             wasm.__wbindgen_export4(deferred2_0, deferred2_1, 1);
         }
     }
+    /**
+     * One-shot auto text extraction — graceful native fallback (never
+     * the opaque OCR error #513).
+     * @param {number} page_index
+     * @returns {string}
+     */
+    extractTextAuto(page_index) {
+        let deferred2_0;
+        let deferred2_1;
+        try {
+            const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+            wasm.wasmpdfdocument_extractTextAuto(retptr, this.__wbg_ptr, page_index);
+            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 text lines from a page.
      *
@@ -6155,6 +6277,36 @@ export function hasDocumentTimestamp(pdf_data) {
     return ret !== 0;
 }
 
+/**
+ * #519: Air-gapped OCR model manifest — JSON (detector + every
+ * supported language's cache filenames and source URLs).
+ *
+ * WASM provisioning is **host-side**: browser/WASM has no filesystem
+ * or network-to-disk, so a download-to-cache prefetch cannot run
+ * here. This manifest is informational — it lets the JS host learn
+ * which model files/URLs to fetch and bundle (or ship out of band)
+ * before driving OCR. There is intentionally no `prefetchModels` in
+ * the WASM surface (see `prefetchAvailable`, which always returns
+ * `false`).
+ * @returns {string}
+ */
+export function modelManifest() {
+    let deferred1_0;
+    let deferred1_1;
+    try {
+        const retptr = wasm.__wbindgen_add_to_stack_pointer(-16);
+        wasm.modelManifest(retptr);
+        var r0 = getDataViewMemory0().getInt32(retptr + 4 * 0, true);
+        var r1 = getDataViewMemory0().getInt32(retptr + 4 * 1, true);
+        deferred1_0 = r0;
+        deferred1_1 = r1;
+        return getStringFromWasm0(r0, r1);
+    } finally {
+        wasm.__wbindgen_add_to_stack_pointer(16);
+        wasm.__wbindgen_export4(deferred1_0, deferred1_1, 1);
+    }
+}
+
 /**
  * Plan a bookmark split without producing PDFs. Returns a JSON array
  * of segment objects (`index, startPage…` shape from
@@ -6186,6 +6338,17 @@ export function planSplitByBookmarks(src_bytes, title_prefix, ignore_case, level
     }
 }
 
+/**
+ * #519: Whether this build can download OCR models to a local cache.
+ * Always `false` in WASM — provisioning is host-side (see
+ * `modelManifest`).
+ * @returns {boolean}
+ */
+export function prefetchAvailable() {
+    const ret = wasm.prefetchAvailable();
+    return ret !== 0;
+}
+
 /**
  * Install the process-wide runtime crypto policy from its grammar
  * string (`"compat"|"strict"|"fips-strict"[;…]`). Fail-closed:

package/web/pdf_oxide_bg.wasm.d.ts

@@ -23,7 +23,9 @@ export const cryptoPolicy: (a: number) => void;
 export const generateBarcodeSvg: (a: number, b: number, c: number, d: number) => void;
 export const generateQrSvg: (a: number, b: number, c: number, d: number, e: number) => void;
 export const hasDocumentTimestamp: (a: number, b: number) => number;
+export const modelManifest: (a: number) => void;
 export const planSplitByBookmarks: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
+export const prefetchAvailable: () => number;
 export const setCryptoPolicy: (a: number, b: number, c: number) => void;
 export const setLogLevel: (a: number, b: number, c: number) => void;
 export const signPdfBytes: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number) => void;
@@ -158,6 +160,8 @@ export const wasmpdfdocument_applyAllRedactions: (a: number, b: number) => void;
 export const wasmpdfdocument_applyPageRedactions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_applyRedactionsDestructive: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_authenticate: (a: number, b: number, c: number, d: number) => void;
+export const wasmpdfdocument_classifyDocument: (a: number, b: number) => void;
+export const wasmpdfdocument_classifyPage: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_clearEraseRegions: (a: number, b: number, c: number) => void;
 export const wasmpdfdocument_convertToPdfA: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_cropMargins: (a: number, b: number, c: number, d: number, e: number, f: number) => void;
@@ -177,6 +181,7 @@ export const wasmpdfdocument_extractChars: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractImageBytes: (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_extractPageAuto: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPageText: (a: number, b: number, c: number, d: number, e: number) => void;
 export const wasmpdfdocument_extractPages: (a: number, b: number, c: number, d: number) => void;
 export const wasmpdfdocument_extractPaths: (a: number, b: number, c: number, d: number, e: number) => void;
@@ -184,6 +189,7 @@ export const wasmpdfdocument_extractRects: (a: number, b: number, c: number, d:
 export const wasmpdfdocument_extractSpans: (a: number, b: number, c: number, d: number, e: number, f: number, g: 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) => void;
+export const wasmpdfdocument_extractTextAuto: (a: number, b: number, c: 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;