@xberg-io/xberg-wasm 1.3.8 → 1.3.9

package/package.json

@@ -10,7 +10,7 @@
     }
   ],
   "name": "@xberg-io/xberg-wasm",
-  "version": "1.3.8",
+  "version": "1.3.9",
   "private": false,
   "description": "High-performance document intelligence library",
   "license": "MIT",

package/pkg/web/package.json

@@ -6,7 +6,7 @@
     "Tobias Perlstein <tobias@xberg.io>"
   ],
   "description": "High-performance document intelligence library",
-  "version": "1.3.8",
+  "version": "1.3.9",
   "license": "MIT",
   "repository": {
     "type": "git",

package/pkg/web/xberg_wasm.d.ts

@@ -3908,7 +3908,7 @@ export class WasmExtractionConfig {
      * keep a handle you already hold, build with `default()` and assign the property instead:
      * the generated setter borrows and leaves your handle alive.
      */
-    constructor(mimeDetectionPolicy?: WasmMimeDetectionPolicy | null, useCache?: boolean | null, enableQualityProcessing?: boolean | null, forceOcr?: boolean | null, ocrStrategy?: any | null, disableOcr?: boolean | null, resultFormat?: WasmResultFormat | null, outputFormat?: any | null, escapeMarkdown?: boolean | null, tableAnchors?: boolean | null, jupyterCellRendering?: WasmJupyterCellRendering | null, applyNotebookCellTags?: boolean | null, useLayoutForMarkdown?: boolean | null, includeDocumentStructure?: boolean | null, url?: WasmUrlExtractionConfig | null, maxArchiveDepth?: number | null, ocr?: WasmOcrConfig | null, forceOcrPages?: Uint32Array | null, ocrNearEmptyFallback?: boolean | null, ocrScannedPageQualityGate?: boolean | null, ocrEmbeddedImages?: boolean | null, chunking?: WasmChunkingConfig | null, contentFilter?: WasmContentFilterConfig | null, images?: WasmImageExtractionConfig | null, pdfOptions?: WasmPdfConfig | null, tokenReduction?: WasmTokenReductionOptions | null, languageDetection?: WasmLanguageDetectionConfig | null, pages?: WasmPageConfig | null, keywords?: WasmKeywordConfig | null, postprocessor?: WasmPostProcessorConfig | null, htmlOptions?: WasmConversionOptions | null, htmlOutput?: WasmHtmlOutputConfig | null, extractionTimeoutSecs?: bigint | null, maxConcurrentExtractions?: number | null, securityLimits?: WasmSecurityLimits | null, maxEmbeddedFileBytes?: bigint | null, layout?: WasmLayoutDetectionConfig | null, transcription?: WasmTranscriptionConfig | null, acceleration?: WasmAccelerationConfig | null, cacheNamespace?: string | null, cacheTtlSecs?: bigint | null, email?: WasmEmailConfig | null, csv?: WasmCsvConfig | null, geojson?: WasmGeoJsonExtractionConfig | null, concurrency?: WasmConcurrencyConfig | null, structuredExtraction?: WasmStructuredExtractionConfig | null, ner?: WasmNerConfig | null, redaction?: WasmRedactionConfig | null, summarization?: WasmSummarizationConfig | null, translation?: WasmTranslationConfig | null, pageClassification?: WasmPageClassificationConfig | null, chunkClassification?: WasmChunkClassificationConfig | null, captioning?: WasmCaptioningConfig | null, qrCodes?: boolean | null);
+    constructor(mimeDetectionPolicy?: WasmMimeDetectionPolicy | null, useCache?: boolean | null, enableQualityProcessing?: boolean | null, forceOcr?: boolean | null, ocrStrategy?: any | null, disableOcr?: boolean | null, resultFormat?: WasmResultFormat | null, outputFormat?: any | null, escapeMarkdown?: boolean | null, tableAnchors?: boolean | null, jupyterCellRendering?: WasmJupyterCellRendering | null, applyNotebookCellTags?: boolean | null, useLayoutForMarkdown?: boolean | null, includeDocumentStructure?: boolean | null, strictParse?: boolean | null, url?: WasmUrlExtractionConfig | null, maxArchiveDepth?: number | null, ocr?: WasmOcrConfig | null, forceOcrPages?: Uint32Array | null, ocrNearEmptyFallback?: boolean | null, ocrScannedPageQualityGate?: boolean | null, ocrEmbeddedImages?: boolean | null, chunking?: WasmChunkingConfig | null, contentFilter?: WasmContentFilterConfig | null, images?: WasmImageExtractionConfig | null, pdfOptions?: WasmPdfConfig | null, tokenReduction?: WasmTokenReductionOptions | null, languageDetection?: WasmLanguageDetectionConfig | null, pages?: WasmPageConfig | null, keywords?: WasmKeywordConfig | null, postprocessor?: WasmPostProcessorConfig | null, htmlOptions?: WasmConversionOptions | null, htmlOutput?: WasmHtmlOutputConfig | null, extractionTimeoutSecs?: bigint | null, maxConcurrentExtractions?: number | null, securityLimits?: WasmSecurityLimits | null, maxEmbeddedFileBytes?: bigint | null, layout?: WasmLayoutDetectionConfig | null, transcription?: WasmTranscriptionConfig | null, acceleration?: WasmAccelerationConfig | null, cacheNamespace?: string | null, cacheTtlSecs?: bigint | null, email?: WasmEmailConfig | null, csv?: WasmCsvConfig | null, geojson?: WasmGeoJsonExtractionConfig | null, concurrency?: WasmConcurrencyConfig | null, structuredExtraction?: WasmStructuredExtractionConfig | null, ner?: WasmNerConfig | null, redaction?: WasmRedactionConfig | null, summarization?: WasmSummarizationConfig | null, translation?: WasmTranslationConfig | null, pageClassification?: WasmPageClassificationConfig | null, chunkClassification?: WasmChunkClassificationConfig | null, captioning?: WasmCaptioningConfig | null, qrCodes?: boolean | null);
     /**
      * Whether embedded images get OCR'd.
      *
@@ -4195,6 +4195,7 @@ export class WasmExtractionConfig {
      */
     get securityLimits(): WasmSecurityLimits | undefined;
     set securityLimits(value: WasmSecurityLimits);
+    strictParse: boolean;
     /**
      * Returns a detached copy of `structuredExtraction`.
      *
@@ -4436,7 +4437,7 @@ export class WasmFileExtractionConfig {
      * keep a handle you already hold, build with `default()` and assign the property instead:
      * the generated setter borrows and leaves your handle alive.
      */
-    constructor(mimeDetectionPolicy?: WasmMimeDetectionPolicy | null, enableQualityProcessing?: boolean | null, ocr?: WasmOcrConfig | null, forceOcr?: boolean | null, ocrStrategy?: any | null, forceOcrPages?: Uint32Array | null, disableOcr?: boolean | null, chunking?: WasmChunkingConfig | null, contentFilter?: WasmContentFilterConfig | null, images?: WasmImageExtractionConfig | null, pdfOptions?: WasmPdfConfig | null, tokenReduction?: WasmTokenReductionOptions | null, languageDetection?: WasmLanguageDetectionConfig | null, pages?: WasmPageConfig | null, keywords?: WasmKeywordConfig | null, postprocessor?: WasmPostProcessorConfig | null, htmlOptions?: WasmConversionOptions | null, htmlOutput?: WasmHtmlOutputConfig | null, resultFormat?: WasmResultFormat | null, outputFormat?: any | null, includeDocumentStructure?: boolean | null, layout?: WasmLayoutDetectionConfig | null, transcription?: WasmTranscriptionConfig | null, timeoutSecs?: bigint | null, structuredExtraction?: WasmStructuredExtractionConfig | null, url?: WasmUrlExtractionConfig | null, ner?: WasmNerConfig | null, redaction?: WasmRedactionConfig | null, summarization?: WasmSummarizationConfig | null, translation?: WasmTranslationConfig | null, pageClassification?: WasmPageClassificationConfig | null, chunkClassification?: WasmChunkClassificationConfig | null, captioning?: WasmCaptioningConfig | null, qrCodes?: boolean | null);
+    constructor(mimeDetectionPolicy?: WasmMimeDetectionPolicy | null, enableQualityProcessing?: boolean | null, ocr?: WasmOcrConfig | null, forceOcr?: boolean | null, ocrStrategy?: any | null, forceOcrPages?: Uint32Array | null, disableOcr?: boolean | null, chunking?: WasmChunkingConfig | null, contentFilter?: WasmContentFilterConfig | null, images?: WasmImageExtractionConfig | null, pdfOptions?: WasmPdfConfig | null, tokenReduction?: WasmTokenReductionOptions | null, languageDetection?: WasmLanguageDetectionConfig | null, pages?: WasmPageConfig | null, keywords?: WasmKeywordConfig | null, postprocessor?: WasmPostProcessorConfig | null, htmlOptions?: WasmConversionOptions | null, htmlOutput?: WasmHtmlOutputConfig | null, resultFormat?: WasmResultFormat | null, outputFormat?: any | null, includeDocumentStructure?: boolean | null, strictParse?: boolean | null, layout?: WasmLayoutDetectionConfig | null, transcription?: WasmTranscriptionConfig | null, timeoutSecs?: bigint | null, structuredExtraction?: WasmStructuredExtractionConfig | null, url?: WasmUrlExtractionConfig | null, ner?: WasmNerConfig | null, redaction?: WasmRedactionConfig | null, summarization?: WasmSummarizationConfig | null, translation?: WasmTranslationConfig | null, pageClassification?: WasmPageClassificationConfig | null, chunkClassification?: WasmChunkClassificationConfig | null, captioning?: WasmCaptioningConfig | null, qrCodes?: boolean | null);
     /**
      * Returns a detached copy of `captioning`.
      *
@@ -4593,6 +4594,8 @@ export class WasmFileExtractionConfig {
     set redaction(value: WasmRedactionConfig);
     get resultFormat(): string | undefined;
     set resultFormat(value: WasmResultFormat | null | undefined);
+    get strictParse(): boolean | undefined;
+    set strictParse(value: boolean | null | undefined);
     /**
      * Returns a detached copy of `structuredExtraction`.
      *
@@ -7361,9 +7364,9 @@ export class WasmOcrMetadata {
  *
  * The warning in `ExtractedDocument.processing_warnings` stays available. One warning
  * can cover several pages: when an automatic OCR run of selected pages fails as a whole,
- * it adds one warning and one record for each selected page. A failure of the OCR fallback
- * for the whole document adds a warning and no record. Repeated failures on one page make
- * one record.
+ * it adds one warning and one record for each selected page. When the OCR fallback for the
+ * whole document fails, it adds one warning and one record for each page that has no native
+ * text. Repeated failures on one page make one record.
  */
 export class WasmOcrPageFailure {
     free(): void;
@@ -13281,7 +13284,7 @@ export interface InitOutput {
     readonly wasmextractionconfig_needsImageData: (a: number) => number;
     readonly wasmextractionconfig_needsImageProcessing: (a: number) => number;
     readonly wasmextractionconfig_ner: (a: number) => number;
-    readonly wasmextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: bigint, j1: number, k1: number, l1: number, m1: bigint, n1: number, o1: number, p1: number, q1: number, r1: number, s1: number, t1: bigint, u1: number, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: number, g2: number) => number;
+    readonly wasmextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: bigint, k1: number, l1: number, m1: number, n1: bigint, o1: number, p1: number, q1: number, r1: number, s1: number, t1: number, u1: bigint, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: number, g2: number, h2: number) => number;
     readonly wasmextractionconfig_ocr: (a: number) => number;
     readonly wasmextractionconfig_ocrEmbeddedImages: (a: number) => number;
     readonly wasmextractionconfig_ocrNearEmptyFallback: (a: number) => number;
@@ -13342,6 +13345,7 @@ export interface InitOutput {
     readonly wasmextractionconfig_set_redaction: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_resultFormat: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_securityLimits: (a: number, b: number) => void;
+    readonly wasmextractionconfig_set_strictParse: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_structuredExtraction: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_summarization: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_tableAnchors: (a: number, b: number) => void;
@@ -13351,6 +13355,7 @@ export interface InitOutput {
     readonly wasmextractionconfig_set_url: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_useCache: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_useLayoutForMarkdown: (a: number, b: number) => void;
+    readonly wasmextractionconfig_strictParse: (a: number) => number;
     readonly wasmextractionconfig_structuredExtraction: (a: number) => number;
     readonly wasmextractionconfig_summarization: (a: number) => number;
     readonly wasmextractionconfig_tableAnchors: (a: number) => number;
@@ -13457,7 +13462,7 @@ export interface InitOutput {
     readonly wasmfileextractionconfig_layout: (a: number) => number;
     readonly wasmfileextractionconfig_mimeDetectionPolicy: (a: number) => [number, number];
     readonly wasmfileextractionconfig_ner: (a: number) => number;
-    readonly wasmfileextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: bigint, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number) => number;
+    readonly wasmfileextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: bigint, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number, k1: number) => number;
     readonly wasmfileextractionconfig_ocr: (a: number) => number;
     readonly wasmfileextractionconfig_ocrStrategy: (a: number) => any;
     readonly wasmfileextractionconfig_outputFormat: (a: number) => any;
@@ -13495,6 +13500,7 @@ export interface InitOutput {
     readonly wasmfileextractionconfig_set_qrCodes: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_redaction: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_resultFormat: (a: number, b: number) => void;
+    readonly wasmfileextractionconfig_set_strictParse: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_structuredExtraction: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_summarization: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_timeoutSecs: (a: number, b: number, c: bigint) => void;
@@ -13502,6 +13508,7 @@ export interface InitOutput {
     readonly wasmfileextractionconfig_set_transcription: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_translation: (a: number, b: number) => void;
     readonly wasmfileextractionconfig_set_url: (a: number, b: number) => void;
+    readonly wasmfileextractionconfig_strictParse: (a: number) => number;
     readonly wasmfileextractionconfig_structuredExtraction: (a: number) => number;
     readonly wasmfileextractionconfig_summarization: (a: number) => number;
     readonly wasmfileextractionconfig_timeoutSecs: (a: number) => [number, bigint];
@@ -15859,9 +15866,9 @@ export interface InitOutput {
     readonly xbergengine_new: (a: any, b: any) => [number, number, number];
     readonly qcms_enable_iccv4: () => void;
     readonly xbergengine_ocr: (a: number, b: number, c: number, d: any) => any;
-    readonly wasm_bindgen__convert__closures_____invoke__h0577a323e04f4984: (a: number, b: number, c: any, d: any) => void;
-    readonly wasm_bindgen__convert__closures_____invoke__h4bf1310a73163971: (a: number, b: number, c: any) => [number, number];
-    readonly wasm_bindgen__convert__closures_____invoke__hc34ab747265d79ec: (a: number, b: number) => void;
+    readonly wasm_bindgen__convert__closures_____invoke__h1697dacc3af5e769: (a: number, b: number, c: any, d: any) => void;
+    readonly wasm_bindgen__convert__closures_____invoke__hcbfea0afcfcd2f5c: (a: number, b: number, c: any) => [number, number];
+    readonly wasm_bindgen__convert__closures_____invoke__h697ad9e8f3481d94: (a: number, b: number) => void;
     readonly __wbindgen_malloc_command_export: (a: number, b: number) => number;
     readonly __wbindgen_realloc_command_export: (a: number, b: number, c: number, d: number) => number;
     readonly __wbindgen_exn_store_command_export: (a: number) => void;

package/pkg/web/xberg_wasm.js

@@ -15853,6 +15853,7 @@ export class WasmExtractionConfig {
      * @param {boolean | null} [applyNotebookCellTags]
      * @param {boolean | null} [useLayoutForMarkdown]
      * @param {boolean | null} [includeDocumentStructure]
+     * @param {boolean | null} [strictParse]
      * @param {WasmUrlExtractionConfig | null} [url]
      * @param {number | null} [maxArchiveDepth]
      * @param {WasmOcrConfig | null} [ocr]
@@ -15894,7 +15895,7 @@ export class WasmExtractionConfig {
      * @param {WasmCaptioningConfig | null} [captioning]
      * @param {boolean | null} [qrCodes]
      */
-    constructor(mimeDetectionPolicy, useCache, enableQualityProcessing, forceOcr, ocrStrategy, disableOcr, resultFormat, outputFormat, escapeMarkdown, tableAnchors, jupyterCellRendering, applyNotebookCellTags, useLayoutForMarkdown, includeDocumentStructure, url, maxArchiveDepth, ocr, forceOcrPages, ocrNearEmptyFallback, ocrScannedPageQualityGate, ocrEmbeddedImages, chunking, contentFilter, images, pdfOptions, tokenReduction, languageDetection, pages, keywords, postprocessor, htmlOptions, htmlOutput, extractionTimeoutSecs, maxConcurrentExtractions, securityLimits, maxEmbeddedFileBytes, layout, transcription, acceleration, cacheNamespace, cacheTtlSecs, email, csv, geojson, concurrency, structuredExtraction, ner, redaction, summarization, translation, pageClassification, chunkClassification, captioning, qrCodes) {
+    constructor(mimeDetectionPolicy, useCache, enableQualityProcessing, forceOcr, ocrStrategy, disableOcr, resultFormat, outputFormat, escapeMarkdown, tableAnchors, jupyterCellRendering, applyNotebookCellTags, useLayoutForMarkdown, includeDocumentStructure, strictParse, url, maxArchiveDepth, ocr, forceOcrPages, ocrNearEmptyFallback, ocrScannedPageQualityGate, ocrEmbeddedImages, chunking, contentFilter, images, pdfOptions, tokenReduction, languageDetection, pages, keywords, postprocessor, htmlOptions, htmlOutput, extractionTimeoutSecs, maxConcurrentExtractions, securityLimits, maxEmbeddedFileBytes, layout, transcription, acceleration, cacheNamespace, cacheTtlSecs, email, csv, geojson, concurrency, structuredExtraction, ner, redaction, summarization, translation, pageClassification, chunkClassification, captioning, qrCodes) {
         let ptr0 = 0;
         if (!isLikeNone(url)) {
             _assertClass(url, WasmUrlExtractionConfig);
@@ -16044,7 +16045,7 @@ export class WasmExtractionConfig {
             _assertClass(captioning, WasmCaptioningConfig);
             ptr30 = captioning.__destroy_into_raw();
         }
-        const ret = wasm.wasmextractionconfig_new(isLikeNone(mimeDetectionPolicy) ? 3 : mimeDetectionPolicy, isLikeNone(useCache) ? 0xFFFFFF : useCache ? 1 : 0, isLikeNone(enableQualityProcessing) ? 0xFFFFFF : enableQualityProcessing ? 1 : 0, isLikeNone(forceOcr) ? 0xFFFFFF : forceOcr ? 1 : 0, isLikeNone(ocrStrategy) ? 0 : addToExternrefTable0(ocrStrategy), isLikeNone(disableOcr) ? 0xFFFFFF : disableOcr ? 1 : 0, isLikeNone(resultFormat) ? 2 : resultFormat, isLikeNone(outputFormat) ? 0 : addToExternrefTable0(outputFormat), isLikeNone(escapeMarkdown) ? 0xFFFFFF : escapeMarkdown ? 1 : 0, isLikeNone(tableAnchors) ? 0xFFFFFF : tableAnchors ? 1 : 0, isLikeNone(jupyterCellRendering) ? 3 : jupyterCellRendering, isLikeNone(applyNotebookCellTags) ? 0xFFFFFF : applyNotebookCellTags ? 1 : 0, isLikeNone(useLayoutForMarkdown) ? 0xFFFFFF : useLayoutForMarkdown ? 1 : 0, isLikeNone(includeDocumentStructure) ? 0xFFFFFF : includeDocumentStructure ? 1 : 0, ptr0, isLikeNone(maxArchiveDepth) ? Number.MAX_SAFE_INTEGER : (maxArchiveDepth) >>> 0, ptr1, ptr2, len2, isLikeNone(ocrNearEmptyFallback) ? 0xFFFFFF : ocrNearEmptyFallback ? 1 : 0, isLikeNone(ocrScannedPageQualityGate) ? 0xFFFFFF : ocrScannedPageQualityGate ? 1 : 0, isLikeNone(ocrEmbeddedImages) ? 0xFFFFFF : ocrEmbeddedImages ? 1 : 0, ptr3, ptr4, ptr5, ptr6, ptr7, ptr8, ptr9, ptr10, ptr11, ptr12, ptr13, !isLikeNone(extractionTimeoutSecs), isLikeNone(extractionTimeoutSecs) ? BigInt(0) : extractionTimeoutSecs, isLikeNone(maxConcurrentExtractions) ? Number.MAX_SAFE_INTEGER : (maxConcurrentExtractions) >>> 0, ptr14, !isLikeNone(maxEmbeddedFileBytes), isLikeNone(maxEmbeddedFileBytes) ? BigInt(0) : maxEmbeddedFileBytes, ptr15, ptr16, ptr17, ptr18, len18, !isLikeNone(cacheTtlSecs), isLikeNone(cacheTtlSecs) ? BigInt(0) : cacheTtlSecs, ptr19, ptr20, ptr21, ptr22, ptr23, ptr24, ptr25, ptr26, ptr27, ptr28, ptr29, ptr30, isLikeNone(qrCodes) ? 0xFFFFFF : qrCodes ? 1 : 0);
+        const ret = wasm.wasmextractionconfig_new(isLikeNone(mimeDetectionPolicy) ? 3 : mimeDetectionPolicy, isLikeNone(useCache) ? 0xFFFFFF : useCache ? 1 : 0, isLikeNone(enableQualityProcessing) ? 0xFFFFFF : enableQualityProcessing ? 1 : 0, isLikeNone(forceOcr) ? 0xFFFFFF : forceOcr ? 1 : 0, isLikeNone(ocrStrategy) ? 0 : addToExternrefTable0(ocrStrategy), isLikeNone(disableOcr) ? 0xFFFFFF : disableOcr ? 1 : 0, isLikeNone(resultFormat) ? 2 : resultFormat, isLikeNone(outputFormat) ? 0 : addToExternrefTable0(outputFormat), isLikeNone(escapeMarkdown) ? 0xFFFFFF : escapeMarkdown ? 1 : 0, isLikeNone(tableAnchors) ? 0xFFFFFF : tableAnchors ? 1 : 0, isLikeNone(jupyterCellRendering) ? 3 : jupyterCellRendering, isLikeNone(applyNotebookCellTags) ? 0xFFFFFF : applyNotebookCellTags ? 1 : 0, isLikeNone(useLayoutForMarkdown) ? 0xFFFFFF : useLayoutForMarkdown ? 1 : 0, isLikeNone(includeDocumentStructure) ? 0xFFFFFF : includeDocumentStructure ? 1 : 0, isLikeNone(strictParse) ? 0xFFFFFF : strictParse ? 1 : 0, ptr0, isLikeNone(maxArchiveDepth) ? Number.MAX_SAFE_INTEGER : (maxArchiveDepth) >>> 0, ptr1, ptr2, len2, isLikeNone(ocrNearEmptyFallback) ? 0xFFFFFF : ocrNearEmptyFallback ? 1 : 0, isLikeNone(ocrScannedPageQualityGate) ? 0xFFFFFF : ocrScannedPageQualityGate ? 1 : 0, isLikeNone(ocrEmbeddedImages) ? 0xFFFFFF : ocrEmbeddedImages ? 1 : 0, ptr3, ptr4, ptr5, ptr6, ptr7, ptr8, ptr9, ptr10, ptr11, ptr12, ptr13, !isLikeNone(extractionTimeoutSecs), isLikeNone(extractionTimeoutSecs) ? BigInt(0) : extractionTimeoutSecs, isLikeNone(maxConcurrentExtractions) ? Number.MAX_SAFE_INTEGER : (maxConcurrentExtractions) >>> 0, ptr14, !isLikeNone(maxEmbeddedFileBytes), isLikeNone(maxEmbeddedFileBytes) ? BigInt(0) : maxEmbeddedFileBytes, ptr15, ptr16, ptr17, ptr18, len18, !isLikeNone(cacheTtlSecs), isLikeNone(cacheTtlSecs) ? BigInt(0) : cacheTtlSecs, ptr19, ptr20, ptr21, ptr22, ptr23, ptr24, ptr25, ptr26, ptr27, ptr28, ptr29, ptr30, isLikeNone(qrCodes) ? 0xFFFFFF : qrCodes ? 1 : 0);
         this.__wbg_ptr = ret;
         WasmExtractionConfigFinalization.register(this, this.__wbg_ptr, this);
         return this;
@@ -16507,6 +16508,12 @@ export class WasmExtractionConfig {
         _assertClass(value, WasmSecurityLimits);
         wasm.wasmextractionconfig_set_securityLimits(this.__wbg_ptr, value.__wbg_ptr);
     }
+    /**
+     * @param {boolean} value
+     */
+    set strictParse(value) {
+        wasm.wasmextractionconfig_set_strictParse(this.__wbg_ptr, value);
+    }
     /**
      * @param {WasmStructuredExtractionConfig} value
      */
@@ -16567,6 +16574,13 @@ export class WasmExtractionConfig {
     set useLayoutForMarkdown(value) {
         wasm.wasmextractionconfig_set_useLayoutForMarkdown(this.__wbg_ptr, value);
     }
+    /**
+     * @returns {boolean}
+     */
+    get strictParse() {
+        const ret = wasm.wasmextractionconfig_strictParse(this.__wbg_ptr);
+        return ret !== 0;
+    }
     /**
      * Returns a detached copy of `structuredExtraction`.
      *
@@ -17648,6 +17662,7 @@ export class WasmFileExtractionConfig {
      * @param {WasmResultFormat | null} [resultFormat]
      * @param {any | null} [outputFormat]
      * @param {boolean | null} [includeDocumentStructure]
+     * @param {boolean | null} [strictParse]
      * @param {WasmLayoutDetectionConfig | null} [layout]
      * @param {WasmTranscriptionConfig | null} [transcription]
      * @param {bigint | null} [timeoutSecs]
@@ -17662,7 +17677,7 @@ export class WasmFileExtractionConfig {
      * @param {WasmCaptioningConfig | null} [captioning]
      * @param {boolean | null} [qrCodes]
      */
-    constructor(mimeDetectionPolicy, enableQualityProcessing, ocr, forceOcr, ocrStrategy, forceOcrPages, disableOcr, chunking, contentFilter, images, pdfOptions, tokenReduction, languageDetection, pages, keywords, postprocessor, htmlOptions, htmlOutput, resultFormat, outputFormat, includeDocumentStructure, layout, transcription, timeoutSecs, structuredExtraction, url, ner, redaction, summarization, translation, pageClassification, chunkClassification, captioning, qrCodes) {
+    constructor(mimeDetectionPolicy, enableQualityProcessing, ocr, forceOcr, ocrStrategy, forceOcrPages, disableOcr, chunking, contentFilter, images, pdfOptions, tokenReduction, languageDetection, pages, keywords, postprocessor, htmlOptions, htmlOutput, resultFormat, outputFormat, includeDocumentStructure, strictParse, layout, transcription, timeoutSecs, structuredExtraction, url, ner, redaction, summarization, translation, pageClassification, chunkClassification, captioning, qrCodes) {
         let ptr0 = 0;
         if (!isLikeNone(ocr)) {
             _assertClass(ocr, WasmOcrConfig);
@@ -17780,7 +17795,7 @@ export class WasmFileExtractionConfig {
             _assertClass(captioning, WasmCaptioningConfig);
             ptr23 = captioning.__destroy_into_raw();
         }
-        const ret = wasm.wasmfileextractionconfig_new(isLikeNone(mimeDetectionPolicy) ? 3 : mimeDetectionPolicy, isLikeNone(enableQualityProcessing) ? 0xFFFFFF : enableQualityProcessing ? 1 : 0, ptr0, isLikeNone(forceOcr) ? 0xFFFFFF : forceOcr ? 1 : 0, isLikeNone(ocrStrategy) ? 0 : addToExternrefTable0(ocrStrategy), ptr1, len1, isLikeNone(disableOcr) ? 0xFFFFFF : disableOcr ? 1 : 0, ptr2, ptr3, ptr4, ptr5, ptr6, ptr7, ptr8, ptr9, ptr10, ptr11, ptr12, isLikeNone(resultFormat) ? 2 : resultFormat, isLikeNone(outputFormat) ? 0 : addToExternrefTable0(outputFormat), isLikeNone(includeDocumentStructure) ? 0xFFFFFF : includeDocumentStructure ? 1 : 0, ptr13, ptr14, !isLikeNone(timeoutSecs), isLikeNone(timeoutSecs) ? BigInt(0) : timeoutSecs, ptr15, ptr16, ptr17, ptr18, ptr19, ptr20, ptr21, ptr22, ptr23, isLikeNone(qrCodes) ? 0xFFFFFF : qrCodes ? 1 : 0);
+        const ret = wasm.wasmfileextractionconfig_new(isLikeNone(mimeDetectionPolicy) ? 3 : mimeDetectionPolicy, isLikeNone(enableQualityProcessing) ? 0xFFFFFF : enableQualityProcessing ? 1 : 0, ptr0, isLikeNone(forceOcr) ? 0xFFFFFF : forceOcr ? 1 : 0, isLikeNone(ocrStrategy) ? 0 : addToExternrefTable0(ocrStrategy), ptr1, len1, isLikeNone(disableOcr) ? 0xFFFFFF : disableOcr ? 1 : 0, ptr2, ptr3, ptr4, ptr5, ptr6, ptr7, ptr8, ptr9, ptr10, ptr11, ptr12, isLikeNone(resultFormat) ? 2 : resultFormat, isLikeNone(outputFormat) ? 0 : addToExternrefTable0(outputFormat), isLikeNone(includeDocumentStructure) ? 0xFFFFFF : includeDocumentStructure ? 1 : 0, isLikeNone(strictParse) ? 0xFFFFFF : strictParse ? 1 : 0, ptr13, ptr14, !isLikeNone(timeoutSecs), isLikeNone(timeoutSecs) ? BigInt(0) : timeoutSecs, ptr15, ptr16, ptr17, ptr18, ptr19, ptr20, ptr21, ptr22, ptr23, isLikeNone(qrCodes) ? 0xFFFFFF : qrCodes ? 1 : 0);
         this.__wbg_ptr = ret;
         WasmFileExtractionConfigFinalization.register(this, this.__wbg_ptr, this);
         return this;
@@ -18065,6 +18080,12 @@ export class WasmFileExtractionConfig {
     set resultFormat(value) {
         wasm.wasmfileextractionconfig_set_resultFormat(this.__wbg_ptr, isLikeNone(value) ? 2 : value);
     }
+    /**
+     * @param {boolean | null} [value]
+     */
+    set strictParse(value) {
+        wasm.wasmfileextractionconfig_set_strictParse(this.__wbg_ptr, isLikeNone(value) ? 0xFFFFFF : value ? 1 : 0);
+    }
     /**
      * @param {WasmStructuredExtractionConfig} value
      */
@@ -18113,6 +18134,13 @@ export class WasmFileExtractionConfig {
         _assertClass(value, WasmUrlExtractionConfig);
         wasm.wasmfileextractionconfig_set_url(this.__wbg_ptr, value.__wbg_ptr);
     }
+    /**
+     * @returns {boolean | undefined}
+     */
+    get strictParse() {
+        const ret = wasm.wasmfileextractionconfig_strictParse(this.__wbg_ptr);
+        return ret === 0xFFFFFF ? undefined : ret !== 0;
+    }
     /**
      * Returns a detached copy of `structuredExtraction`.
      *
@@ -29573,9 +29601,9 @@ if (Symbol.dispose) WasmOcrMetadata.prototype[Symbol.dispose] = WasmOcrMetadata.
  *
  * The warning in `ExtractedDocument.processing_warnings` stays available. One warning
  * can cover several pages: when an automatic OCR run of selected pages fails as a whole,
- * it adds one warning and one record for each selected page. A failure of the OCR fallback
- * for the whole document adds a warning and no record. Repeated failures on one page make
- * one record.
+ * it adds one warning and one record for each selected page. When the OCR fallback for the
+ * whole document fails, it adds one warning and one record for each page that has no native
+ * text. Repeated failures on one page make one record.
  */
 export class WasmOcrPageFailure {
     static __wrap(ptr) {
@@ -46328,7 +46356,7 @@ function __wbg_get_imports() {
                     const a = state0.a;
                     state0.a = 0;
                     try {
-                        return wasm_bindgen__convert__closures_____invoke__h0577a323e04f4984(a, state0.b, arg0, arg1);
+                        return wasm_bindgen__convert__closures_____invoke__h1697dacc3af5e769(a, state0.b, arg0, arg1);
                     } finally {
                         state0.a = a;
                     }
@@ -46488,7 +46516,7 @@ function __wbg_get_imports() {
             const ret = arg0.value;
             return ret;
         },
-        __wbg_warn_b16dffa679542e9c: function(arg0, arg1) {
+        __wbg_warn_b222f740e9ddf70d: function(arg0, arg1) {
             console.warn(getStringFromWasm0(arg0, arg1));
         },
         __wbg_wasmarchiveentry_new: function(arg0) {
@@ -47068,13 +47096,13 @@ function __wbg_get_imports() {
             return ret;
         },
         __wbindgen_generic_0000000000000001: function(arg0, arg1) {
-            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [Externref], shim_idx: 692, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
-            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__h4bf1310a73163971);
+            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [Externref], shim_idx: 684, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
+            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__hcbfea0afcfcd2f5c);
             return ret;
         },
         __wbindgen_generic_0000000000000002: function(arg0, arg1) {
-            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 1136, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
-            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__hc34ab747265d79ec);
+            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 1131, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
+            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__h697ad9e8f3481d94);
             return ret;
         },
         __wbindgen_generic_0000000000000003: function(arg0) {
@@ -47142,19 +47170,19 @@ function __wbg_get_imports() {
     };
 }
 
-function wasm_bindgen__convert__closures_____invoke__hc34ab747265d79ec(arg0, arg1) {
-    wasm.wasm_bindgen__convert__closures_____invoke__hc34ab747265d79ec(arg0, arg1);
+function wasm_bindgen__convert__closures_____invoke__h697ad9e8f3481d94(arg0, arg1) {
+    wasm.wasm_bindgen__convert__closures_____invoke__h697ad9e8f3481d94(arg0, arg1);
 }
 
-function wasm_bindgen__convert__closures_____invoke__h4bf1310a73163971(arg0, arg1, arg2) {
-    const ret = wasm.wasm_bindgen__convert__closures_____invoke__h4bf1310a73163971(arg0, arg1, arg2);
+function wasm_bindgen__convert__closures_____invoke__hcbfea0afcfcd2f5c(arg0, arg1, arg2) {
+    const ret = wasm.wasm_bindgen__convert__closures_____invoke__hcbfea0afcfcd2f5c(arg0, arg1, arg2);
     if (ret[1]) {
         throw takeFromExternrefTable0(ret[0]);
     }
 }
 
-function wasm_bindgen__convert__closures_____invoke__h0577a323e04f4984(arg0, arg1, arg2, arg3) {
-    wasm.wasm_bindgen__convert__closures_____invoke__h0577a323e04f4984(arg0, arg1, arg2, arg3);
+function wasm_bindgen__convert__closures_____invoke__h1697dacc3af5e769(arg0, arg1, arg2, arg3) {
+    wasm.wasm_bindgen__convert__closures_____invoke__h1697dacc3af5e769(arg0, arg1, arg2, arg3);
 }
 
 

package/pkg/web/xberg_wasm_bg.wasm

@@ -1,4 +1,4 @@
 [diffend] Oversized file quarantined before diffing.
 name: package/pkg/web/xberg_wasm_bg.wasm
-size: 38141194 bytes
-sha256: 438419e1c0fd85793460c13c31a1069d3ad76a496a123dc54a81c5b65eee32f6
+size: 38155686 bytes
+sha256: 36d9eac1ba5c078a3fe66b091f9385701ed3b0608f1d2db549c990a3d9886001

package/pkg/web/xberg_wasm_bg.wasm.d.ts

@@ -1705,7 +1705,7 @@ export const wasmextractionconfig_mimeDetectionPolicy: (a: number) => [number, n
 export const wasmextractionconfig_needsImageData: (a: number) => number;
 export const wasmextractionconfig_needsImageProcessing: (a: number) => number;
 export const wasmextractionconfig_ner: (a: number) => number;
-export const wasmextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: bigint, j1: number, k1: number, l1: number, m1: bigint, n1: number, o1: number, p1: number, q1: number, r1: number, s1: number, t1: bigint, u1: number, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: number, g2: number) => number;
+export const wasmextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: bigint, k1: number, l1: number, m1: number, n1: bigint, o1: number, p1: number, q1: number, r1: number, s1: number, t1: number, u1: bigint, v1: number, w1: number, x1: number, y1: number, z1: number, a2: number, b2: number, c2: number, d2: number, e2: number, f2: number, g2: number, h2: number) => number;
 export const wasmextractionconfig_ocr: (a: number) => number;
 export const wasmextractionconfig_ocrEmbeddedImages: (a: number) => number;
 export const wasmextractionconfig_ocrNearEmptyFallback: (a: number) => number;
@@ -1766,6 +1766,7 @@ export const wasmextractionconfig_set_qrCodes: (a: number, b: number) => void;
 export const wasmextractionconfig_set_redaction: (a: number, b: number) => void;
 export const wasmextractionconfig_set_resultFormat: (a: number, b: number) => void;
 export const wasmextractionconfig_set_securityLimits: (a: number, b: number) => void;
+export const wasmextractionconfig_set_strictParse: (a: number, b: number) => void;
 export const wasmextractionconfig_set_structuredExtraction: (a: number, b: number) => void;
 export const wasmextractionconfig_set_summarization: (a: number, b: number) => void;
 export const wasmextractionconfig_set_tableAnchors: (a: number, b: number) => void;
@@ -1775,6 +1776,7 @@ export const wasmextractionconfig_set_translation: (a: number, b: number) => voi
 export const wasmextractionconfig_set_url: (a: number, b: number) => void;
 export const wasmextractionconfig_set_useCache: (a: number, b: number) => void;
 export const wasmextractionconfig_set_useLayoutForMarkdown: (a: number, b: number) => void;
+export const wasmextractionconfig_strictParse: (a: number) => number;
 export const wasmextractionconfig_structuredExtraction: (a: number) => number;
 export const wasmextractionconfig_summarization: (a: number) => number;
 export const wasmextractionconfig_tableAnchors: (a: number) => number;
@@ -1881,7 +1883,7 @@ export const wasmfileextractionconfig_languageDetection: (a: number) => number;
 export const wasmfileextractionconfig_layout: (a: number) => number;
 export const wasmfileextractionconfig_mimeDetectionPolicy: (a: number) => [number, number];
 export const wasmfileextractionconfig_ner: (a: number) => number;
-export const wasmfileextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: bigint, a1: number, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number) => number;
+export const wasmfileextractionconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: number, i: number, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number, v: number, w: number, x: number, y: number, z: number, a1: bigint, b1: number, c1: number, d1: number, e1: number, f1: number, g1: number, h1: number, i1: number, j1: number, k1: number) => number;
 export const wasmfileextractionconfig_ocr: (a: number) => number;
 export const wasmfileextractionconfig_ocrStrategy: (a: number) => any;
 export const wasmfileextractionconfig_outputFormat: (a: number) => any;
@@ -1919,6 +1921,7 @@ export const wasmfileextractionconfig_set_postprocessor: (a: number, b: number)
 export const wasmfileextractionconfig_set_qrCodes: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_redaction: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_resultFormat: (a: number, b: number) => void;
+export const wasmfileextractionconfig_set_strictParse: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_structuredExtraction: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_summarization: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_timeoutSecs: (a: number, b: number, c: bigint) => void;
@@ -1926,6 +1929,7 @@ export const wasmfileextractionconfig_set_tokenReduction: (a: number, b: number)
 export const wasmfileextractionconfig_set_transcription: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_translation: (a: number, b: number) => void;
 export const wasmfileextractionconfig_set_url: (a: number, b: number) => void;
+export const wasmfileextractionconfig_strictParse: (a: number) => number;
 export const wasmfileextractionconfig_structuredExtraction: (a: number) => number;
 export const wasmfileextractionconfig_summarization: (a: number) => number;
 export const wasmfileextractionconfig_timeoutSecs: (a: number) => [number, bigint];
@@ -4283,9 +4287,9 @@ export const xbergengine_ner: (a: number, b: number, c: number, d: any) => any;
 export const xbergengine_new: (a: any, b: any) => [number, number, number];
 export const qcms_enable_iccv4: () => void;
 export const xbergengine_ocr: (a: number, b: number, c: number, d: any) => any;
-export const wasm_bindgen__convert__closures_____invoke__h0577a323e04f4984: (a: number, b: number, c: any, d: any) => void;
-export const wasm_bindgen__convert__closures_____invoke__h4bf1310a73163971: (a: number, b: number, c: any) => [number, number];
-export const wasm_bindgen__convert__closures_____invoke__hc34ab747265d79ec: (a: number, b: number) => void;
+export const wasm_bindgen__convert__closures_____invoke__h1697dacc3af5e769: (a: number, b: number, c: any, d: any) => void;
+export const wasm_bindgen__convert__closures_____invoke__hcbfea0afcfcd2f5c: (a: number, b: number, c: any) => [number, number];
+export const wasm_bindgen__convert__closures_____invoke__h697ad9e8f3481d94: (a: number, b: number) => void;
 export const __wbindgen_malloc_command_export: (a: number, b: number) => number;
 export const __wbindgen_realloc_command_export: (a: number, b: number, c: number, d: number) => number;
 export const __wbindgen_exn_store_command_export: (a: number) => void;