@xberg-io/xberg-wasm 1.2.3 → 1.2.6

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@xberg-io/xberg-wasm",
-  "version": "1.2.3",
+  "version": "1.2.6",
   "private": false,
   "description": "High-performance document intelligence library",
   "license": "MIT",

package/pkg/web/package.json

@@ -2,7 +2,7 @@
   "name": "xberg-wasm",
   "type": "module",
   "description": "High-performance document intelligence library",
-  "version": "1.2.3",
+  "version": "1.2.6",
   "license": "MIT",
   "repository": {
     "type": "git",

package/pkg/web/xberg_wasm.d.ts

@@ -10,6 +10,8 @@ export interface WasmPdfMetadataCamelWire {
     pageCount: number | undefined;
     scannedConfidence: number | undefined;
     scannedPages: number[] | undefined;
+    fabricatedTextPages: number[] | undefined;
+    implausibleTextPages: number[] | undefined;
     layoutGatedPages: number[] | undefined;
     layoutGateReasons: string[] | undefined;
 }
@@ -2491,11 +2493,21 @@ export class WasmExtractionConfig {
      * For text-only extractions (no OCR, no image extraction), skipping image
      * decompression can improve CPU utilization by 5-10% by avoiding wasteful
      * image I/O and processing when results won't be used.
-     * Returns `true` when image binary data should be extracted.
+     * Returns `true` when a CONTAINER document (DOCX, PPT, PPTX, HTML, ...) needs
+     * to read an embedded image's binary data out of its archive or source.
      *
-     * True when `config.images.extract_images` is set, captioning is configured, or QR-code
-     * detection is enabled. Captioning and QR-code detection both require image bytes
-     * regardless of whether the caller also requested image extraction.
+     * True when `config.images.extract_images` is set, captioning is configured, QR-code
+     * detection is enabled, or embedded images get OCR'd. All four consume the bytes:
+     * captioning and QR-code detection regardless of whether the caller also requested
+     * image extraction, and embedded-image OCR because the OCR path reads the same bytes
+     * the container would otherwise skip (GH#1662): before the OCR disjunct existed, a
+     * container asked this question, was told no, and attached an image with an empty
+     * buffer, which OCR then ran on and reported `Could not determine image format`.
+     *
+     * This governs whether a CONTAINER reads bytes it does not otherwise have. A
+     * standalone image extractor already holds the whole file's bytes regardless of this
+     * answer, so it must not use this method to decide whether to put its OWN bytes in the
+     * result's `images` array; that is `wants_own_bytes_in_result`.
      */
     needsImageData(): boolean;
     /**
@@ -2509,6 +2521,18 @@ export class WasmExtractionConfig {
      */
     needsImageProcessing(): boolean;
     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, 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.
+     *
+     * This mirrors the condition the pipeline itself uses before it calls
+     * `image_ocr.process_images_with_ocr` (`core/pipeline/mod.rs`), and it is
+     * deliberately the same predicate rather than an equivalent one: when the
+     * two drifted apart, a container extractor asked `needs_image_data` and was
+     * told no, so it attached an image with an empty buffer, and the OCR path
+     * then ran on those zero bytes and reported `Could not determine image
+     * format` (GH#1662).
+     */
+    runsOcrOnEmbeddedImages(): boolean;
     /**
      * Validate the configuration, returning an error if any settings are invalid.
      *
@@ -2538,6 +2562,21 @@ export class WasmExtractionConfig {
      * Returns `XbergError.Validation` describing the first invalid setting found.
      */
     validate(): void;
+    /**
+     * Returns `true` when a standalone image extraction (the whole input document IS
+     * the image, not a container's embedded picture) should attach its own already-read
+     * bytes to the result's `images` array.
+     *
+     * Deliberately narrower than `needs_image_data`: a standalone
+     * image's content is read in full before this question is ever asked, whether or not OCR
+     * runs, so OCR being configured is not by itself a reason to echo those bytes back in the
+     * public output. Using `needs_image_data` here (as `needs_image_data`'s own OCR disjunct
+     * briefly did) would silently start returning every OCR'd standalone image's raw bytes in
+     * `images`, a public output contract change with nothing asking for it. Same three
+     * conditions as before that disjunct existed: explicit image extraction, captioning, or
+     * QR-code detection.
+     */
+    wantsOwnBytesInResult(): boolean;
     get acceleration(): WasmAccelerationConfig | undefined;
     set acceleration(value: WasmAccelerationConfig | null | undefined);
     applyNotebookCellTags: boolean;
@@ -3867,7 +3906,7 @@ export class WasmLlmConfig {
     free(): void;
     [Symbol.dispose](): void;
     static default(): WasmLlmConfig;
-    constructor(model?: string | null, apiKey?: string | null, baseUrl?: string | null, timeoutSecs?: bigint | null, maxRetries?: number | null, temperature?: number | null, maxTokens?: bigint | null, topP?: number | null, stop?: string[] | null, seed?: bigint | null, presencePenalty?: number | null, frequencyPenalty?: number | null, reasoningEffort?: string | null, extraBody?: any | null, loadEnv?: boolean | null, headers?: any | null, providers?: WasmLlmProviderConfig[] | null, cache?: WasmLlmCacheConfig | null, budget?: WasmLlmBudgetConfig | null, rateLimit?: WasmLlmRateLimitConfig | null, costTracking?: boolean | null, tracing?: boolean | null, cooldownSecs?: bigint | null, healthCheckSecs?: bigint | null, bedrock?: WasmBedrockConfig | null, credentialProvider?: any | null, maxConcurrency?: number | null);
+    constructor(model?: string | null, apiKey?: string | null, baseUrl?: string | null, timeoutSecs?: bigint | null, maxRetries?: number | null, temperature?: number | null, maxTokens?: bigint | null, topP?: number | null, stop?: string[] | null, seed?: bigint | null, presencePenalty?: number | null, frequencyPenalty?: number | null, reasoningEffort?: string | null, extraBody?: any | null, loadEnv?: boolean | null, headers?: any | null, providers?: WasmLlmProviderConfig[] | null, cache?: WasmLlmCacheConfig | null, budget?: WasmLlmBudgetConfig | null, rateLimit?: WasmLlmRateLimitConfig | null, costTracking?: boolean | null, tracing?: boolean | null, cooldownSecs?: bigint | null, healthCheckSecs?: bigint | null, bedrock?: WasmBedrockConfig | null, credentialProvider?: any | null, maxConcurrency?: number | null, maxResponseBytes?: number | null);
     /**
      * Validate the request-time sampling parameters that have a documented range:
      * `top_p` (`[0.0, 1.0]`), `presence_penalty`, and `frequency_penalty` (both
@@ -3907,6 +3946,8 @@ export class WasmLlmConfig {
     set loadEnv(value: boolean | null | undefined);
     get maxConcurrency(): number | undefined;
     set maxConcurrency(value: number | null | undefined);
+    get maxResponseBytes(): number | undefined;
+    set maxResponseBytes(value: number | null | undefined);
     get maxRetries(): number | undefined;
     set maxRetries(value: number | null | undefined);
     get maxTokens(): bigint | undefined;
@@ -4593,10 +4634,11 @@ export class WasmOcrQualityThresholds {
     free(): void;
     [Symbol.dispose](): void;
     static default(): WasmOcrQualityThresholds;
-    constructor(minTotalNonWhitespace?: number | null, minNonWhitespacePerPage?: number | null, minMeaningfulWordLen?: number | null, minMeaningfulWords?: number | null, minAlnumRatio?: number | null, minGarbageChars?: number | null, maxFragmentedWordRatio?: number | null, criticalFragmentedWordRatio?: number | null, maxOcrOutputFragmentedWordRatio?: number | null, minOcrMeanConfidence?: number | null, minWordsForOcrOutputCheck?: number | null, maxOcrOutputDictInvalidWordRatio?: number | null, discardSuspectedOcrNoise?: boolean | null, minAvgWordLength?: number | null, minWordsForAvgLengthCheck?: number | null, minConsecutiveRepeatRatio?: number | null, minWordsForRepeatCheck?: number | null, substantiveMinChars?: number | null, nonTextMinChars?: number | null, alnumWsRatioThreshold?: number | null, pipelineMinQuality?: number | null, minUndecodableRatio?: number | null, enableProvenanceOcrRouting?: boolean | null, minProvenanceFallbackRatio?: number | null);
+    constructor(minTotalNonWhitespace?: number | null, minNonWhitespacePerPage?: number | null, minMeaningfulWordLen?: number | null, minMeaningfulWords?: number | null, minAlnumRatio?: number | null, minGarbageChars?: number | null, maxFragmentedWordRatio?: number | null, criticalFragmentedWordRatio?: number | null, maxOcrOutputFragmentedWordRatio?: number | null, minOcrMeanConfidence?: number | null, minWordsForOcrOutputCheck?: number | null, maxOcrOutputDictInvalidWordRatio?: number | null, discardSuspectedOcrNoise?: boolean | null, minAvgWordLength?: number | null, minWordsForAvgLengthCheck?: number | null, minConsecutiveRepeatRatio?: number | null, minWordsForRepeatCheck?: number | null, substantiveMinChars?: number | null, nonTextMinChars?: number | null, alnumWsRatioThreshold?: number | null, pipelineMinQuality?: number | null, minUndecodableRatio?: number | null, enableProvenanceOcrRouting?: boolean | null, minProvenanceFallbackRatio?: number | null, enablePlausibilityOcrRouting?: boolean | null, minReliableLanguageChunkRatio?: number | null);
     alnumWsRatioThreshold: number;
     criticalFragmentedWordRatio: number;
     discardSuspectedOcrNoise: boolean;
+    enablePlausibilityOcrRouting: boolean;
     enableProvenanceOcrRouting: boolean;
     maxFragmentedWordRatio: number;
     maxOcrOutputDictInvalidWordRatio: number;
@@ -4610,6 +4652,7 @@ export class WasmOcrQualityThresholds {
     minNonWhitespacePerPage: number;
     minOcrMeanConfidence: number;
     minProvenanceFallbackRatio: number;
+    minReliableLanguageChunkRatio: number;
     minTotalNonWhitespace: number;
     minUndecodableRatio: number;
     minWordsForAvgLengthCheck: number;
@@ -5375,9 +5418,13 @@ export class WasmPdfMetadata {
     free(): void;
     [Symbol.dispose](): void;
     static default(): WasmPdfMetadata;
-    constructor(pdfVersion?: string | null, producer?: string | null, isEncrypted?: boolean | null, width?: bigint | null, height?: bigint | null, pageCount?: number | null, scannedConfidence?: number | null, scannedPages?: Uint32Array | null, layoutGatedPages?: Uint32Array | null, layoutGateReasons?: string[] | null);
+    constructor(pdfVersion?: string | null, producer?: string | null, isEncrypted?: boolean | null, width?: bigint | null, height?: bigint | null, pageCount?: number | null, scannedConfidence?: number | null, scannedPages?: Uint32Array | null, fabricatedTextPages?: Uint32Array | null, implausibleTextPages?: Uint32Array | null, layoutGatedPages?: Uint32Array | null, layoutGateReasons?: string[] | null);
+    get fabricatedTextPages(): Uint32Array | undefined;
+    set fabricatedTextPages(value: Uint32Array | null | undefined);
     get height(): bigint | undefined;
     set height(value: bigint | null | undefined);
+    get implausibleTextPages(): Uint32Array | undefined;
+    set implausibleTextPages(value: Uint32Array | null | undefined);
     get isEncrypted(): boolean | undefined;
     set isEncrypted(value: boolean | null | undefined);
     get layoutGateReasons(): string[] | undefined;
@@ -7246,9 +7293,10 @@ export function listEmbeddingBackends(): string[];
  * # Cost
  *
  * Calling this is not free for every backend. In particular, `TesseractBackend`'s
- * `supported_languages()` allocates a Tesseract API and initializes it (`init("", "eng")`) the
- * first time it is called, to enumerate installed tessdata languages; subsequent calls are
- * served from a cache.
+ * `supported_languages()` allocates a Tesseract API and initializes it against the same
+ * tessdata directory a real OCR job resolves (`resolve_tessdata_path`), the first time it
+ * is called, to enumerate installed tessdata languages; subsequent calls are served from a
+ * cache.
  *
  * # Errors
  *
@@ -8972,6 +9020,7 @@ export interface InitOutput {
     readonly wasmextractionconfig_qrCodes: (a: number) => number;
     readonly wasmextractionconfig_redaction: (a: number) => number;
     readonly wasmextractionconfig_resultFormat: (a: number) => [number, number];
+    readonly wasmextractionconfig_runsOcrOnEmbeddedImages: (a: number) => number;
     readonly wasmextractionconfig_securityLimits: (a: number) => number;
     readonly wasmextractionconfig_set_acceleration: (a: number, b: number) => void;
     readonly wasmextractionconfig_set_applyNotebookCellTags: (a: number, b: number) => void;
@@ -9034,6 +9083,7 @@ export interface InitOutput {
     readonly wasmextractionconfig_useCache: (a: number) => number;
     readonly wasmextractionconfig_useLayoutForMarkdown: (a: number) => number;
     readonly wasmextractionconfig_validate: (a: number) => [number, number];
+    readonly wasmextractionconfig_wantsOwnBytesInResult: (a: number) => number;
     readonly wasmextractionerroritem_code: (a: number) => number;
     readonly wasmextractionerroritem_default: () => number;
     readonly wasmextractionerroritem_errorType: (a: number) => [number, number];
@@ -9680,10 +9730,11 @@ export interface InitOutput {
     readonly wasmllmconfig_healthCheckSecs: (a: number) => [number, bigint];
     readonly wasmllmconfig_loadEnv: (a: number) => number;
     readonly wasmllmconfig_maxConcurrency: (a: number) => number;
+    readonly wasmllmconfig_maxResponseBytes: (a: number) => number;
     readonly wasmllmconfig_maxRetries: (a: number) => number;
     readonly wasmllmconfig_maxTokens: (a: number) => [number, bigint];
     readonly wasmllmconfig_model: (a: number) => [number, number];
-    readonly wasmllmconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint, i: number, j: number, k: number, l: number, m: bigint, n: number, o: number, p: number, q: number, r: number, s: bigint, 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: number, k1: bigint, l1: number, m1: bigint, n1: number, o1: number, p1: number) => number;
+    readonly wasmllmconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint, i: number, j: number, k: number, l: number, m: bigint, n: number, o: number, p: number, q: number, r: number, s: bigint, 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: number, k1: bigint, l1: number, m1: bigint, n1: number, o1: number, p1: number, q1: number) => number;
     readonly wasmllmconfig_presencePenalty: (a: number) => [number, number];
     readonly wasmllmconfig_providers: (a: number) => any;
     readonly wasmllmconfig_rateLimit: (a: number) => number;
@@ -9703,6 +9754,7 @@ export interface InitOutput {
     readonly wasmllmconfig_set_healthCheckSecs: (a: number, b: number, c: bigint) => void;
     readonly wasmllmconfig_set_loadEnv: (a: number, b: number) => void;
     readonly wasmllmconfig_set_maxConcurrency: (a: number, b: number) => void;
+    readonly wasmllmconfig_set_maxResponseBytes: (a: number, b: number) => void;
     readonly wasmllmconfig_set_maxRetries: (a: number, b: number) => void;
     readonly wasmllmconfig_set_maxTokens: (a: number, b: number, c: bigint) => void;
     readonly wasmllmconfig_set_model: (a: number, b: number, c: number) => void;
@@ -10082,6 +10134,7 @@ export interface InitOutput {
     readonly wasmocrqualitythresholds_criticalFragmentedWordRatio: (a: number) => number;
     readonly wasmocrqualitythresholds_default: () => number;
     readonly wasmocrqualitythresholds_discardSuspectedOcrNoise: (a: number) => number;
+    readonly wasmocrqualitythresholds_enablePlausibilityOcrRouting: (a: number) => number;
     readonly wasmocrqualitythresholds_enableProvenanceOcrRouting: (a: number) => number;
     readonly wasmocrqualitythresholds_maxFragmentedWordRatio: (a: number) => number;
     readonly wasmocrqualitythresholds_maxOcrOutputDictInvalidWordRatio: (a: number) => number;
@@ -10095,17 +10148,19 @@ export interface InitOutput {
     readonly wasmocrqualitythresholds_minNonWhitespacePerPage: (a: number) => number;
     readonly wasmocrqualitythresholds_minOcrMeanConfidence: (a: number) => number;
     readonly wasmocrqualitythresholds_minProvenanceFallbackRatio: (a: number) => number;
+    readonly wasmocrqualitythresholds_minReliableLanguageChunkRatio: (a: number) => number;
     readonly wasmocrqualitythresholds_minTotalNonWhitespace: (a: number) => number;
     readonly wasmocrqualitythresholds_minUndecodableRatio: (a: number) => number;
     readonly wasmocrqualitythresholds_minWordsForAvgLengthCheck: (a: number) => number;
     readonly wasmocrqualitythresholds_minWordsForOcrOutputCheck: (a: number) => number;
     readonly wasmocrqualitythresholds_minWordsForRepeatCheck: (a: number) => number;
-    readonly wasmocrqualitythresholds_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: number, k1: number) => number;
+    readonly wasmocrqualitythresholds_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: number, k1: number, l1: number, m1: number, n1: number) => number;
     readonly wasmocrqualitythresholds_nonTextMinChars: (a: number) => number;
     readonly wasmocrqualitythresholds_pipelineMinQuality: (a: number) => number;
     readonly wasmocrqualitythresholds_set_alnumWsRatioThreshold: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_criticalFragmentedWordRatio: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_discardSuspectedOcrNoise: (a: number, b: number) => void;
+    readonly wasmocrqualitythresholds_set_enablePlausibilityOcrRouting: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_enableProvenanceOcrRouting: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_maxFragmentedWordRatio: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_maxOcrOutputDictInvalidWordRatio: (a: number, b: number) => void;
@@ -10119,6 +10174,7 @@ export interface InitOutput {
     readonly wasmocrqualitythresholds_set_minNonWhitespacePerPage: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_minOcrMeanConfidence: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_minProvenanceFallbackRatio: (a: number, b: number) => void;
+    readonly wasmocrqualitythresholds_set_minReliableLanguageChunkRatio: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_minTotalNonWhitespace: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_minUndecodableRatio: (a: number, b: number) => void;
     readonly wasmocrqualitythresholds_set_minWordsForAvgLengthCheck: (a: number, b: number) => void;
@@ -10426,17 +10482,21 @@ export interface InitOutput {
     readonly wasmpdfformfield_tooltip: (a: number) => [number, number];
     readonly wasmpdfformfield_value: (a: number) => [number, number];
     readonly wasmpdfmetadata_default: () => number;
+    readonly wasmpdfmetadata_fabricatedTextPages: (a: number) => [number, number];
     readonly wasmpdfmetadata_height: (a: number) => [number, bigint];
+    readonly wasmpdfmetadata_implausibleTextPages: (a: number) => [number, number];
     readonly wasmpdfmetadata_isEncrypted: (a: number) => number;
     readonly wasmpdfmetadata_layoutGateReasons: (a: number) => [number, number];
     readonly wasmpdfmetadata_layoutGatedPages: (a: number) => [number, number];
-    readonly wasmpdfmetadata_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint, h: number, i: bigint, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number) => number;
+    readonly wasmpdfmetadata_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint, h: number, i: bigint, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number) => number;
     readonly wasmpdfmetadata_pageCount: (a: number) => number;
     readonly wasmpdfmetadata_pdfVersion: (a: number) => [number, number];
     readonly wasmpdfmetadata_producer: (a: number) => [number, number];
     readonly wasmpdfmetadata_scannedConfidence: (a: number) => number;
     readonly wasmpdfmetadata_scannedPages: (a: number) => [number, number];
+    readonly wasmpdfmetadata_set_fabricatedTextPages: (a: number, b: number, c: number) => void;
     readonly wasmpdfmetadata_set_height: (a: number, b: number, c: bigint) => void;
+    readonly wasmpdfmetadata_set_implausibleTextPages: (a: number, b: number, c: number) => void;
     readonly wasmpdfmetadata_set_isEncrypted: (a: number, b: number) => void;
     readonly wasmpdfmetadata_set_layoutGateReasons: (a: number, b: number, c: number) => void;
     readonly wasmpdfmetadata_set_layoutGatedPages: (a: number, b: number, c: number) => void;
@@ -11260,9 +11320,9 @@ 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__h26802dab031c3123: (a: number, b: number, c: any) => [number, number];
-    readonly wasm_bindgen__convert__closures_____invoke__h44bdeda823314cfc: (a: number, b: number, c: any, d: any) => void;
-    readonly wasm_bindgen__convert__closures_____invoke__ha4d0d7342deba426: (a: number, b: number) => void;
+    readonly wasm_bindgen__convert__closures_____invoke__h1752b33118c61bf6: (a: number, b: number, c: any, d: any) => void;
+    readonly wasm_bindgen__convert__closures_____invoke__h58dcb4b03347de73: (a: number, b: number, c: any) => [number, number];
+    readonly wasm_bindgen__convert__closures_____invoke__hc0b53db039d89c01: (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

@@ -13866,11 +13866,21 @@ export class WasmExtractionConfig {
      * For text-only extractions (no OCR, no image extraction), skipping image
      * decompression can improve CPU utilization by 5-10% by avoiding wasteful
      * image I/O and processing when results won't be used.
-     * Returns `true` when image binary data should be extracted.
+     * Returns `true` when a CONTAINER document (DOCX, PPT, PPTX, HTML, ...) needs
+     * to read an embedded image's binary data out of its archive or source.
      *
-     * True when `config.images.extract_images` is set, captioning is configured, or QR-code
-     * detection is enabled. Captioning and QR-code detection both require image bytes
-     * regardless of whether the caller also requested image extraction.
+     * True when `config.images.extract_images` is set, captioning is configured, QR-code
+     * detection is enabled, or embedded images get OCR'd. All four consume the bytes:
+     * captioning and QR-code detection regardless of whether the caller also requested
+     * image extraction, and embedded-image OCR because the OCR path reads the same bytes
+     * the container would otherwise skip (GH#1662): before the OCR disjunct existed, a
+     * container asked this question, was told no, and attached an image with an empty
+     * buffer, which OCR then ran on and reported `Could not determine image format`.
+     *
+     * This governs whether a CONTAINER reads bytes it does not otherwise have. A
+     * standalone image extractor already holds the whole file's bytes regardless of this
+     * answer, so it must not use this method to decide whether to put its OWN bytes in the
+     * result's `images` array; that is `wants_own_bytes_in_result`.
      * @returns {boolean}
      */
     needsImageData() {
@@ -14184,6 +14194,22 @@ export class WasmExtractionConfig {
             wasm.__wbindgen_free_command_export(deferred1_0, deferred1_1, 1);
         }
     }
+    /**
+     * Whether embedded images get OCR'd.
+     *
+     * This mirrors the condition the pipeline itself uses before it calls
+     * `image_ocr.process_images_with_ocr` (`core/pipeline/mod.rs`), and it is
+     * deliberately the same predicate rather than an equivalent one: when the
+     * two drifted apart, a container extractor asked `needs_image_data` and was
+     * told no, so it attached an image with an empty buffer, and the OCR path
+     * then ran on those zero bytes and reported `Could not determine image
+     * format` (GH#1662).
+     * @returns {boolean}
+     */
+    runsOcrOnEmbeddedImages() {
+        const ret = wasm.wasmextractionconfig_runsOcrOnEmbeddedImages(this.__wbg_ptr);
+        return ret !== 0;
+    }
     /**
      * @returns {WasmSecurityLimits | undefined}
      */
@@ -14740,6 +14766,25 @@ export class WasmExtractionConfig {
             throw takeFromExternrefTable0(ret[0]);
         }
     }
+    /**
+     * Returns `true` when a standalone image extraction (the whole input document IS
+     * the image, not a container's embedded picture) should attach its own already-read
+     * bytes to the result's `images` array.
+     *
+     * Deliberately narrower than `needs_image_data`: a standalone
+     * image's content is read in full before this question is ever asked, whether or not OCR
+     * runs, so OCR being configured is not by itself a reason to echo those bytes back in the
+     * public output. Using `needs_image_data` here (as `needs_image_data`'s own OCR disjunct
+     * briefly did) would silently start returning every OCR'd standalone image's raw bytes in
+     * `images`, a public output contract change with nothing asking for it. Same three
+     * conditions as before that disjunct existed: explicit image extraction, captioning, or
+     * QR-code detection.
+     * @returns {boolean}
+     */
+    wantsOwnBytesInResult() {
+        const ret = wasm.wasmextractionconfig_wantsOwnBytesInResult(this.__wbg_ptr);
+        return ret !== 0;
+    }
 }
 if (Symbol.dispose) WasmExtractionConfig.prototype[Symbol.dispose] = WasmExtractionConfig.prototype.free;
 
@@ -22194,6 +22239,13 @@ export class WasmLlmConfig {
         const ret = wasm.wasmllmconfig_maxConcurrency(this.__wbg_ptr);
         return ret === Number.MAX_SAFE_INTEGER ? undefined : ret;
     }
+    /**
+     * @returns {number | undefined}
+     */
+    get maxResponseBytes() {
+        const ret = wasm.wasmllmconfig_maxResponseBytes(this.__wbg_ptr);
+        return ret === Number.MAX_SAFE_INTEGER ? undefined : ret;
+    }
     /**
      * @returns {number | undefined}
      */
@@ -22251,8 +22303,9 @@ export class WasmLlmConfig {
      * @param {WasmBedrockConfig | null} [bedrock]
      * @param {any | null} [credentialProvider]
      * @param {number | null} [maxConcurrency]
+     * @param {number | null} [maxResponseBytes]
      */
-    constructor(model, apiKey, baseUrl, timeoutSecs, maxRetries, temperature, maxTokens, topP, stop, seed, presencePenalty, frequencyPenalty, reasoningEffort, extraBody, loadEnv, headers, providers, cache, budget, rateLimit, costTracking, tracing, cooldownSecs, healthCheckSecs, bedrock, credentialProvider, maxConcurrency) {
+    constructor(model, apiKey, baseUrl, timeoutSecs, maxRetries, temperature, maxTokens, topP, stop, seed, presencePenalty, frequencyPenalty, reasoningEffort, extraBody, loadEnv, headers, providers, cache, budget, rateLimit, costTracking, tracing, cooldownSecs, healthCheckSecs, bedrock, credentialProvider, maxConcurrency, maxResponseBytes) {
         var ptr0 = isLikeNone(model) ? 0 : passStringToWasm0(model, wasm.__wbindgen_malloc_command_export, wasm.__wbindgen_realloc_command_export);
         var len0 = WASM_VECTOR_LEN;
         var ptr1 = isLikeNone(apiKey) ? 0 : passStringToWasm0(apiKey, wasm.__wbindgen_malloc_command_export, wasm.__wbindgen_realloc_command_export);
@@ -22285,7 +22338,7 @@ export class WasmLlmConfig {
             _assertClass(bedrock, WasmBedrockConfig);
             ptr9 = bedrock.__destroy_into_raw();
         }
-        const ret = wasm.wasmllmconfig_new(ptr0, len0, ptr1, len1, ptr2, len2, !isLikeNone(timeoutSecs), isLikeNone(timeoutSecs) ? BigInt(0) : timeoutSecs, isLikeNone(maxRetries) ? Number.MAX_SAFE_INTEGER : (maxRetries) >>> 0, !isLikeNone(temperature), isLikeNone(temperature) ? 0 : temperature, !isLikeNone(maxTokens), isLikeNone(maxTokens) ? BigInt(0) : maxTokens, !isLikeNone(topP), isLikeNone(topP) ? 0 : topP, ptr3, len3, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed, !isLikeNone(presencePenalty), isLikeNone(presencePenalty) ? 0 : presencePenalty, !isLikeNone(frequencyPenalty), isLikeNone(frequencyPenalty) ? 0 : frequencyPenalty, ptr4, len4, isLikeNone(extraBody) ? 0 : addToExternrefTable0(extraBody), isLikeNone(loadEnv) ? 0xFFFFFF : loadEnv ? 1 : 0, isLikeNone(headers) ? 0 : addToExternrefTable0(headers), ptr5, len5, ptr6, ptr7, ptr8, isLikeNone(costTracking) ? 0xFFFFFF : costTracking ? 1 : 0, isLikeNone(tracing) ? 0xFFFFFF : tracing ? 1 : 0, !isLikeNone(cooldownSecs), isLikeNone(cooldownSecs) ? BigInt(0) : cooldownSecs, !isLikeNone(healthCheckSecs), isLikeNone(healthCheckSecs) ? BigInt(0) : healthCheckSecs, ptr9, isLikeNone(credentialProvider) ? 0 : addToExternrefTable0(credentialProvider), isLikeNone(maxConcurrency) ? Number.MAX_SAFE_INTEGER : (maxConcurrency) >>> 0);
+        const ret = wasm.wasmllmconfig_new(ptr0, len0, ptr1, len1, ptr2, len2, !isLikeNone(timeoutSecs), isLikeNone(timeoutSecs) ? BigInt(0) : timeoutSecs, isLikeNone(maxRetries) ? Number.MAX_SAFE_INTEGER : (maxRetries) >>> 0, !isLikeNone(temperature), isLikeNone(temperature) ? 0 : temperature, !isLikeNone(maxTokens), isLikeNone(maxTokens) ? BigInt(0) : maxTokens, !isLikeNone(topP), isLikeNone(topP) ? 0 : topP, ptr3, len3, !isLikeNone(seed), isLikeNone(seed) ? BigInt(0) : seed, !isLikeNone(presencePenalty), isLikeNone(presencePenalty) ? 0 : presencePenalty, !isLikeNone(frequencyPenalty), isLikeNone(frequencyPenalty) ? 0 : frequencyPenalty, ptr4, len4, isLikeNone(extraBody) ? 0 : addToExternrefTable0(extraBody), isLikeNone(loadEnv) ? 0xFFFFFF : loadEnv ? 1 : 0, isLikeNone(headers) ? 0 : addToExternrefTable0(headers), ptr5, len5, ptr6, ptr7, ptr8, isLikeNone(costTracking) ? 0xFFFFFF : costTracking ? 1 : 0, isLikeNone(tracing) ? 0xFFFFFF : tracing ? 1 : 0, !isLikeNone(cooldownSecs), isLikeNone(cooldownSecs) ? BigInt(0) : cooldownSecs, !isLikeNone(healthCheckSecs), isLikeNone(healthCheckSecs) ? BigInt(0) : healthCheckSecs, ptr9, isLikeNone(credentialProvider) ? 0 : addToExternrefTable0(credentialProvider), isLikeNone(maxConcurrency) ? Number.MAX_SAFE_INTEGER : (maxConcurrency) >>> 0, isLikeNone(maxResponseBytes) ? Number.MAX_SAFE_INTEGER : (maxResponseBytes) >>> 0);
         this.__wbg_ptr = ret;
         WasmLlmConfigFinalization.register(this, this.__wbg_ptr, this);
         return this;
@@ -22433,6 +22486,12 @@ export class WasmLlmConfig {
     set maxConcurrency(value) {
         wasm.wasmllmconfig_set_maxConcurrency(this.__wbg_ptr, isLikeNone(value) ? Number.MAX_SAFE_INTEGER : (value) >>> 0);
     }
+    /**
+     * @param {number | null} [value]
+     */
+    set maxResponseBytes(value) {
+        wasm.wasmllmconfig_set_maxResponseBytes(this.__wbg_ptr, isLikeNone(value) ? Number.MAX_SAFE_INTEGER : (value) >>> 0);
+    }
     /**
      * @param {number | null} [value]
      */
@@ -26732,6 +26791,13 @@ export class WasmOcrQualityThresholds {
         const ret = wasm.wasmocrqualitythresholds_discardSuspectedOcrNoise(this.__wbg_ptr);
         return ret !== 0;
     }
+    /**
+     * @returns {boolean}
+     */
+    get enablePlausibilityOcrRouting() {
+        const ret = wasm.wasmocrqualitythresholds_enablePlausibilityOcrRouting(this.__wbg_ptr);
+        return ret !== 0;
+    }
     /**
      * @returns {boolean}
      */
@@ -26823,6 +26889,13 @@ export class WasmOcrQualityThresholds {
         const ret = wasm.wasmocrqualitythresholds_minProvenanceFallbackRatio(this.__wbg_ptr);
         return ret;
     }
+    /**
+     * @returns {number}
+     */
+    get minReliableLanguageChunkRatio() {
+        const ret = wasm.wasmocrqualitythresholds_minReliableLanguageChunkRatio(this.__wbg_ptr);
+        return ret;
+    }
     /**
      * @returns {number}
      */
@@ -26883,9 +26956,11 @@ export class WasmOcrQualityThresholds {
      * @param {number | null} [minUndecodableRatio]
      * @param {boolean | null} [enableProvenanceOcrRouting]
      * @param {number | null} [minProvenanceFallbackRatio]
+     * @param {boolean | null} [enablePlausibilityOcrRouting]
+     * @param {number | null} [minReliableLanguageChunkRatio]
      */
-    constructor(minTotalNonWhitespace, minNonWhitespacePerPage, minMeaningfulWordLen, minMeaningfulWords, minAlnumRatio, minGarbageChars, maxFragmentedWordRatio, criticalFragmentedWordRatio, maxOcrOutputFragmentedWordRatio, minOcrMeanConfidence, minWordsForOcrOutputCheck, maxOcrOutputDictInvalidWordRatio, discardSuspectedOcrNoise, minAvgWordLength, minWordsForAvgLengthCheck, minConsecutiveRepeatRatio, minWordsForRepeatCheck, substantiveMinChars, nonTextMinChars, alnumWsRatioThreshold, pipelineMinQuality, minUndecodableRatio, enableProvenanceOcrRouting, minProvenanceFallbackRatio) {
-        const ret = wasm.wasmocrqualitythresholds_new(isLikeNone(minTotalNonWhitespace) ? Number.MAX_SAFE_INTEGER : (minTotalNonWhitespace) >>> 0, !isLikeNone(minNonWhitespacePerPage), isLikeNone(minNonWhitespacePerPage) ? 0 : minNonWhitespacePerPage, isLikeNone(minMeaningfulWordLen) ? Number.MAX_SAFE_INTEGER : (minMeaningfulWordLen) >>> 0, isLikeNone(minMeaningfulWords) ? Number.MAX_SAFE_INTEGER : (minMeaningfulWords) >>> 0, !isLikeNone(minAlnumRatio), isLikeNone(minAlnumRatio) ? 0 : minAlnumRatio, isLikeNone(minGarbageChars) ? Number.MAX_SAFE_INTEGER : (minGarbageChars) >>> 0, !isLikeNone(maxFragmentedWordRatio), isLikeNone(maxFragmentedWordRatio) ? 0 : maxFragmentedWordRatio, !isLikeNone(criticalFragmentedWordRatio), isLikeNone(criticalFragmentedWordRatio) ? 0 : criticalFragmentedWordRatio, !isLikeNone(maxOcrOutputFragmentedWordRatio), isLikeNone(maxOcrOutputFragmentedWordRatio) ? 0 : maxOcrOutputFragmentedWordRatio, !isLikeNone(minOcrMeanConfidence), isLikeNone(minOcrMeanConfidence) ? 0 : minOcrMeanConfidence, isLikeNone(minWordsForOcrOutputCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForOcrOutputCheck) >>> 0, !isLikeNone(maxOcrOutputDictInvalidWordRatio), isLikeNone(maxOcrOutputDictInvalidWordRatio) ? 0 : maxOcrOutputDictInvalidWordRatio, isLikeNone(discardSuspectedOcrNoise) ? 0xFFFFFF : discardSuspectedOcrNoise ? 1 : 0, !isLikeNone(minAvgWordLength), isLikeNone(minAvgWordLength) ? 0 : minAvgWordLength, isLikeNone(minWordsForAvgLengthCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForAvgLengthCheck) >>> 0, !isLikeNone(minConsecutiveRepeatRatio), isLikeNone(minConsecutiveRepeatRatio) ? 0 : minConsecutiveRepeatRatio, isLikeNone(minWordsForRepeatCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForRepeatCheck) >>> 0, isLikeNone(substantiveMinChars) ? Number.MAX_SAFE_INTEGER : (substantiveMinChars) >>> 0, isLikeNone(nonTextMinChars) ? Number.MAX_SAFE_INTEGER : (nonTextMinChars) >>> 0, !isLikeNone(alnumWsRatioThreshold), isLikeNone(alnumWsRatioThreshold) ? 0 : alnumWsRatioThreshold, !isLikeNone(pipelineMinQuality), isLikeNone(pipelineMinQuality) ? 0 : pipelineMinQuality, !isLikeNone(minUndecodableRatio), isLikeNone(minUndecodableRatio) ? 0 : minUndecodableRatio, isLikeNone(enableProvenanceOcrRouting) ? 0xFFFFFF : enableProvenanceOcrRouting ? 1 : 0, !isLikeNone(minProvenanceFallbackRatio), isLikeNone(minProvenanceFallbackRatio) ? 0 : minProvenanceFallbackRatio);
+    constructor(minTotalNonWhitespace, minNonWhitespacePerPage, minMeaningfulWordLen, minMeaningfulWords, minAlnumRatio, minGarbageChars, maxFragmentedWordRatio, criticalFragmentedWordRatio, maxOcrOutputFragmentedWordRatio, minOcrMeanConfidence, minWordsForOcrOutputCheck, maxOcrOutputDictInvalidWordRatio, discardSuspectedOcrNoise, minAvgWordLength, minWordsForAvgLengthCheck, minConsecutiveRepeatRatio, minWordsForRepeatCheck, substantiveMinChars, nonTextMinChars, alnumWsRatioThreshold, pipelineMinQuality, minUndecodableRatio, enableProvenanceOcrRouting, minProvenanceFallbackRatio, enablePlausibilityOcrRouting, minReliableLanguageChunkRatio) {
+        const ret = wasm.wasmocrqualitythresholds_new(isLikeNone(minTotalNonWhitespace) ? Number.MAX_SAFE_INTEGER : (minTotalNonWhitespace) >>> 0, !isLikeNone(minNonWhitespacePerPage), isLikeNone(minNonWhitespacePerPage) ? 0 : minNonWhitespacePerPage, isLikeNone(minMeaningfulWordLen) ? Number.MAX_SAFE_INTEGER : (minMeaningfulWordLen) >>> 0, isLikeNone(minMeaningfulWords) ? Number.MAX_SAFE_INTEGER : (minMeaningfulWords) >>> 0, !isLikeNone(minAlnumRatio), isLikeNone(minAlnumRatio) ? 0 : minAlnumRatio, isLikeNone(minGarbageChars) ? Number.MAX_SAFE_INTEGER : (minGarbageChars) >>> 0, !isLikeNone(maxFragmentedWordRatio), isLikeNone(maxFragmentedWordRatio) ? 0 : maxFragmentedWordRatio, !isLikeNone(criticalFragmentedWordRatio), isLikeNone(criticalFragmentedWordRatio) ? 0 : criticalFragmentedWordRatio, !isLikeNone(maxOcrOutputFragmentedWordRatio), isLikeNone(maxOcrOutputFragmentedWordRatio) ? 0 : maxOcrOutputFragmentedWordRatio, !isLikeNone(minOcrMeanConfidence), isLikeNone(minOcrMeanConfidence) ? 0 : minOcrMeanConfidence, isLikeNone(minWordsForOcrOutputCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForOcrOutputCheck) >>> 0, !isLikeNone(maxOcrOutputDictInvalidWordRatio), isLikeNone(maxOcrOutputDictInvalidWordRatio) ? 0 : maxOcrOutputDictInvalidWordRatio, isLikeNone(discardSuspectedOcrNoise) ? 0xFFFFFF : discardSuspectedOcrNoise ? 1 : 0, !isLikeNone(minAvgWordLength), isLikeNone(minAvgWordLength) ? 0 : minAvgWordLength, isLikeNone(minWordsForAvgLengthCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForAvgLengthCheck) >>> 0, !isLikeNone(minConsecutiveRepeatRatio), isLikeNone(minConsecutiveRepeatRatio) ? 0 : minConsecutiveRepeatRatio, isLikeNone(minWordsForRepeatCheck) ? Number.MAX_SAFE_INTEGER : (minWordsForRepeatCheck) >>> 0, isLikeNone(substantiveMinChars) ? Number.MAX_SAFE_INTEGER : (substantiveMinChars) >>> 0, isLikeNone(nonTextMinChars) ? Number.MAX_SAFE_INTEGER : (nonTextMinChars) >>> 0, !isLikeNone(alnumWsRatioThreshold), isLikeNone(alnumWsRatioThreshold) ? 0 : alnumWsRatioThreshold, !isLikeNone(pipelineMinQuality), isLikeNone(pipelineMinQuality) ? 0 : pipelineMinQuality, !isLikeNone(minUndecodableRatio), isLikeNone(minUndecodableRatio) ? 0 : minUndecodableRatio, isLikeNone(enableProvenanceOcrRouting) ? 0xFFFFFF : enableProvenanceOcrRouting ? 1 : 0, !isLikeNone(minProvenanceFallbackRatio), isLikeNone(minProvenanceFallbackRatio) ? 0 : minProvenanceFallbackRatio, isLikeNone(enablePlausibilityOcrRouting) ? 0xFFFFFF : enablePlausibilityOcrRouting ? 1 : 0, !isLikeNone(minReliableLanguageChunkRatio), isLikeNone(minReliableLanguageChunkRatio) ? 0 : minReliableLanguageChunkRatio);
         this.__wbg_ptr = ret;
         WasmOcrQualityThresholdsFinalization.register(this, this.__wbg_ptr, this);
         return this;
@@ -26922,6 +26997,12 @@ export class WasmOcrQualityThresholds {
     set discardSuspectedOcrNoise(value) {
         wasm.wasmocrqualitythresholds_set_discardSuspectedOcrNoise(this.__wbg_ptr, value);
     }
+    /**
+     * @param {boolean} value
+     */
+    set enablePlausibilityOcrRouting(value) {
+        wasm.wasmocrqualitythresholds_set_enablePlausibilityOcrRouting(this.__wbg_ptr, value);
+    }
     /**
      * @param {boolean} value
      */
@@ -27000,6 +27081,12 @@ export class WasmOcrQualityThresholds {
     set minProvenanceFallbackRatio(value) {
         wasm.wasmocrqualitythresholds_set_minProvenanceFallbackRatio(this.__wbg_ptr, value);
     }
+    /**
+     * @param {number} value
+     */
+    set minReliableLanguageChunkRatio(value) {
+        wasm.wasmocrqualitythresholds_set_minReliableLanguageChunkRatio(this.__wbg_ptr, value);
+    }
     /**
      * @param {number} value
      */
@@ -30573,6 +30660,18 @@ export class WasmPdfMetadata {
         const ret = wasm.wasmpdfmetadata_default();
         return WasmPdfMetadata.__wrap(ret);
     }
+    /**
+     * @returns {Uint32Array | undefined}
+     */
+    get fabricatedTextPages() {
+        const ret = wasm.wasmpdfmetadata_fabricatedTextPages(this.__wbg_ptr);
+        let v1;
+        if (ret[0] !== 0) {
+            v1 = getArrayU32FromWasm0(ret[0], ret[1]).slice();
+            wasm.__wbindgen_free_command_export(ret[0], ret[1] * 4, 4);
+        }
+        return v1;
+    }
     /**
      * @returns {bigint | undefined}
      */
@@ -30580,6 +30679,18 @@ export class WasmPdfMetadata {
         const ret = wasm.wasmpdfmetadata_height(this.__wbg_ptr);
         return ret[0] === 0 ? undefined : ret[1];
     }
+    /**
+     * @returns {Uint32Array | undefined}
+     */
+    get implausibleTextPages() {
+        const ret = wasm.wasmpdfmetadata_implausibleTextPages(this.__wbg_ptr);
+        let v1;
+        if (ret[0] !== 0) {
+            v1 = getArrayU32FromWasm0(ret[0], ret[1]).slice();
+            wasm.__wbindgen_free_command_export(ret[0], ret[1] * 4, 4);
+        }
+        return v1;
+    }
     /**
      * @returns {boolean | undefined}
      */
@@ -30620,21 +30731,27 @@ export class WasmPdfMetadata {
      * @param {number | null} [pageCount]
      * @param {number | null} [scannedConfidence]
      * @param {Uint32Array | null} [scannedPages]
+     * @param {Uint32Array | null} [fabricatedTextPages]
+     * @param {Uint32Array | null} [implausibleTextPages]
      * @param {Uint32Array | null} [layoutGatedPages]
      * @param {string[] | null} [layoutGateReasons]
      */
-    constructor(pdfVersion, producer, isEncrypted, width, height, pageCount, scannedConfidence, scannedPages, layoutGatedPages, layoutGateReasons) {
+    constructor(pdfVersion, producer, isEncrypted, width, height, pageCount, scannedConfidence, scannedPages, fabricatedTextPages, implausibleTextPages, layoutGatedPages, layoutGateReasons) {
         var ptr0 = isLikeNone(pdfVersion) ? 0 : passStringToWasm0(pdfVersion, wasm.__wbindgen_malloc_command_export, wasm.__wbindgen_realloc_command_export);
         var len0 = WASM_VECTOR_LEN;
         var ptr1 = isLikeNone(producer) ? 0 : passStringToWasm0(producer, wasm.__wbindgen_malloc_command_export, wasm.__wbindgen_realloc_command_export);
         var len1 = WASM_VECTOR_LEN;
         var ptr2 = isLikeNone(scannedPages) ? 0 : passArray32ToWasm0(scannedPages, wasm.__wbindgen_malloc_command_export);
         var len2 = WASM_VECTOR_LEN;
-        var ptr3 = isLikeNone(layoutGatedPages) ? 0 : passArray32ToWasm0(layoutGatedPages, wasm.__wbindgen_malloc_command_export);
+        var ptr3 = isLikeNone(fabricatedTextPages) ? 0 : passArray32ToWasm0(fabricatedTextPages, wasm.__wbindgen_malloc_command_export);
         var len3 = WASM_VECTOR_LEN;
-        var ptr4 = isLikeNone(layoutGateReasons) ? 0 : passArrayJsValueToWasm0(layoutGateReasons, wasm.__wbindgen_malloc_command_export);
+        var ptr4 = isLikeNone(implausibleTextPages) ? 0 : passArray32ToWasm0(implausibleTextPages, wasm.__wbindgen_malloc_command_export);
         var len4 = WASM_VECTOR_LEN;
-        const ret = wasm.wasmpdfmetadata_new(ptr0, len0, ptr1, len1, isLikeNone(isEncrypted) ? 0xFFFFFF : isEncrypted ? 1 : 0, !isLikeNone(width), isLikeNone(width) ? BigInt(0) : width, !isLikeNone(height), isLikeNone(height) ? BigInt(0) : height, isLikeNone(pageCount) ? Number.MAX_SAFE_INTEGER : (pageCount) >>> 0, isLikeNone(scannedConfidence) ? Number.MAX_SAFE_INTEGER : Math.fround(scannedConfidence), ptr2, len2, ptr3, len3, ptr4, len4);
+        var ptr5 = isLikeNone(layoutGatedPages) ? 0 : passArray32ToWasm0(layoutGatedPages, wasm.__wbindgen_malloc_command_export);
+        var len5 = WASM_VECTOR_LEN;
+        var ptr6 = isLikeNone(layoutGateReasons) ? 0 : passArrayJsValueToWasm0(layoutGateReasons, wasm.__wbindgen_malloc_command_export);
+        var len6 = WASM_VECTOR_LEN;
+        const ret = wasm.wasmpdfmetadata_new(ptr0, len0, ptr1, len1, isLikeNone(isEncrypted) ? 0xFFFFFF : isEncrypted ? 1 : 0, !isLikeNone(width), isLikeNone(width) ? BigInt(0) : width, !isLikeNone(height), isLikeNone(height) ? BigInt(0) : height, isLikeNone(pageCount) ? Number.MAX_SAFE_INTEGER : (pageCount) >>> 0, isLikeNone(scannedConfidence) ? Number.MAX_SAFE_INTEGER : Math.fround(scannedConfidence), ptr2, len2, ptr3, len3, ptr4, len4, ptr5, len5, ptr6, len6);
         this.__wbg_ptr = ret;
         WasmPdfMetadataFinalization.register(this, this.__wbg_ptr, this);
         return this;
@@ -30689,12 +30806,28 @@ export class WasmPdfMetadata {
         }
         return v1;
     }
+    /**
+     * @param {Uint32Array | null} [value]
+     */
+    set fabricatedTextPages(value) {
+        var ptr0 = isLikeNone(value) ? 0 : passArray32ToWasm0(value, wasm.__wbindgen_malloc_command_export);
+        var len0 = WASM_VECTOR_LEN;
+        wasm.wasmpdfmetadata_set_fabricatedTextPages(this.__wbg_ptr, ptr0, len0);
+    }
     /**
      * @param {bigint | null} [value]
      */
     set height(value) {
         wasm.wasmpdfmetadata_set_height(this.__wbg_ptr, !isLikeNone(value), isLikeNone(value) ? BigInt(0) : value);
     }
+    /**
+     * @param {Uint32Array | null} [value]
+     */
+    set implausibleTextPages(value) {
+        var ptr0 = isLikeNone(value) ? 0 : passArray32ToWasm0(value, wasm.__wbindgen_malloc_command_export);
+        var len0 = WASM_VECTOR_LEN;
+        wasm.wasmpdfmetadata_set_implausibleTextPages(this.__wbg_ptr, ptr0, len0);
+    }
     /**
      * @param {boolean | null} [value]
      */
@@ -40435,9 +40568,10 @@ export function listEmbeddingBackends() {
  * # Cost
  *
  * Calling this is not free for every backend. In particular, `TesseractBackend`'s
- * `supported_languages()` allocates a Tesseract API and initializes it (`init("", "eng")`) the
- * first time it is called, to enumerate installed tessdata languages; subsequent calls are
- * served from a cache.
+ * `supported_languages()` allocates a Tesseract API and initializes it against the same
+ * tessdata directory a real OCR job resolves (`resolve_tessdata_path`), the first time it
+ * is called, to enumerate installed tessdata languages; subsequent calls are served from a
+ * cache.
  *
  * # Errors
  *
@@ -41277,7 +41411,7 @@ function __wbg_get_imports() {
                     const a = state0.a;
                     state0.a = 0;
                     try {
-                        return wasm_bindgen__convert__closures_____invoke__h44bdeda823314cfc(a, state0.b, arg0, arg1);
+                        return wasm_bindgen__convert__closures_____invoke__h1752b33118c61bf6(a, state0.b, arg0, arg1);
                     } finally {
                         state0.a = a;
                     }
@@ -41437,7 +41571,7 @@ function __wbg_get_imports() {
             const ret = arg0.value;
             return ret;
         },
-        __wbg_warn_3b898cbca4e507de: function(arg0, arg1) {
+        __wbg_warn_2404f6e2ec92fd86: function(arg0, arg1) {
             console.warn(getStringFromWasm0(arg0, arg1));
         },
         __wbg_wasmarchiveentry_new: function(arg0) {
@@ -42001,13 +42135,13 @@ 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: 673, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
-            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__h26802dab031c3123);
+            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [Externref], shim_idx: 677, ret: Result(Unit), inner_ret: Some(Result(Unit)) }, mutable: true }) -> Externref`.
+            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__h58dcb4b03347de73);
             return ret;
         },
         __wbindgen_generic_0000000000000002: function(arg0, arg1) {
-            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 1170, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
-            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__ha4d0d7342deba426);
+            // Cast intrinsic for `Closure(Closure { owned: true, function: Function { arguments: [], shim_idx: 1171, ret: Unit, inner_ret: Some(Unit) }, mutable: true }) -> Externref`.
+            const ret = makeMutClosure(arg0, arg1, wasm_bindgen__convert__closures_____invoke__hc0b53db039d89c01);
             return ret;
         },
         __wbindgen_generic_0000000000000003: function(arg0) {
@@ -42075,19 +42209,19 @@ function __wbg_get_imports() {
     };
 }
 
-function wasm_bindgen__convert__closures_____invoke__ha4d0d7342deba426(arg0, arg1) {
-    wasm.wasm_bindgen__convert__closures_____invoke__ha4d0d7342deba426(arg0, arg1);
+function wasm_bindgen__convert__closures_____invoke__hc0b53db039d89c01(arg0, arg1) {
+    wasm.wasm_bindgen__convert__closures_____invoke__hc0b53db039d89c01(arg0, arg1);
 }
 
-function wasm_bindgen__convert__closures_____invoke__h26802dab031c3123(arg0, arg1, arg2) {
-    const ret = wasm.wasm_bindgen__convert__closures_____invoke__h26802dab031c3123(arg0, arg1, arg2);
+function wasm_bindgen__convert__closures_____invoke__h58dcb4b03347de73(arg0, arg1, arg2) {
+    const ret = wasm.wasm_bindgen__convert__closures_____invoke__h58dcb4b03347de73(arg0, arg1, arg2);
     if (ret[1]) {
         throw takeFromExternrefTable0(ret[0]);
     }
 }
 
-function wasm_bindgen__convert__closures_____invoke__h44bdeda823314cfc(arg0, arg1, arg2, arg3) {
-    wasm.wasm_bindgen__convert__closures_____invoke__h44bdeda823314cfc(arg0, arg1, arg2, arg3);
+function wasm_bindgen__convert__closures_____invoke__h1752b33118c61bf6(arg0, arg1, arg2, arg3) {
+    wasm.wasm_bindgen__convert__closures_____invoke__h1752b33118c61bf6(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: 36443440 bytes
-sha256: 7d6a595c701378198ac0c6d747c094faf0e4bc6825e0041c14eeb3a1906e39f3
+size: 36531009 bytes
+sha256: ba9a867b795a54358bbe0964fa583b54fca80be2f5b7735754892c882989b853

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

@@ -1533,6 +1533,7 @@ export const wasmextractionconfig_postprocessor: (a: number) => number;
 export const wasmextractionconfig_qrCodes: (a: number) => number;
 export const wasmextractionconfig_redaction: (a: number) => number;
 export const wasmextractionconfig_resultFormat: (a: number) => [number, number];
+export const wasmextractionconfig_runsOcrOnEmbeddedImages: (a: number) => number;
 export const wasmextractionconfig_securityLimits: (a: number) => number;
 export const wasmextractionconfig_set_acceleration: (a: number, b: number) => void;
 export const wasmextractionconfig_set_applyNotebookCellTags: (a: number, b: number) => void;
@@ -1595,6 +1596,7 @@ export const wasmextractionconfig_url: (a: number) => number;
 export const wasmextractionconfig_useCache: (a: number) => number;
 export const wasmextractionconfig_useLayoutForMarkdown: (a: number) => number;
 export const wasmextractionconfig_validate: (a: number) => [number, number];
+export const wasmextractionconfig_wantsOwnBytesInResult: (a: number) => number;
 export const wasmextractionerroritem_code: (a: number) => number;
 export const wasmextractionerroritem_default: () => number;
 export const wasmextractionerroritem_errorType: (a: number) => [number, number];
@@ -2241,10 +2243,11 @@ export const wasmllmconfig_headers: (a: number) => any;
 export const wasmllmconfig_healthCheckSecs: (a: number) => [number, bigint];
 export const wasmllmconfig_loadEnv: (a: number) => number;
 export const wasmllmconfig_maxConcurrency: (a: number) => number;
+export const wasmllmconfig_maxResponseBytes: (a: number) => number;
 export const wasmllmconfig_maxRetries: (a: number) => number;
 export const wasmllmconfig_maxTokens: (a: number) => [number, bigint];
 export const wasmllmconfig_model: (a: number) => [number, number];
-export const wasmllmconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint, i: number, j: number, k: number, l: number, m: bigint, n: number, o: number, p: number, q: number, r: number, s: bigint, 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: number, k1: bigint, l1: number, m1: bigint, n1: number, o1: number, p1: number) => number;
+export const wasmllmconfig_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: number, h: bigint, i: number, j: number, k: number, l: number, m: bigint, n: number, o: number, p: number, q: number, r: number, s: bigint, 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: number, k1: bigint, l1: number, m1: bigint, n1: number, o1: number, p1: number, q1: number) => number;
 export const wasmllmconfig_presencePenalty: (a: number) => [number, number];
 export const wasmllmconfig_providers: (a: number) => any;
 export const wasmllmconfig_rateLimit: (a: number) => number;
@@ -2264,6 +2267,7 @@ export const wasmllmconfig_set_headers: (a: number, b: number) => void;
 export const wasmllmconfig_set_healthCheckSecs: (a: number, b: number, c: bigint) => void;
 export const wasmllmconfig_set_loadEnv: (a: number, b: number) => void;
 export const wasmllmconfig_set_maxConcurrency: (a: number, b: number) => void;
+export const wasmllmconfig_set_maxResponseBytes: (a: number, b: number) => void;
 export const wasmllmconfig_set_maxRetries: (a: number, b: number) => void;
 export const wasmllmconfig_set_maxTokens: (a: number, b: number, c: bigint) => void;
 export const wasmllmconfig_set_model: (a: number, b: number, c: number) => void;
@@ -2643,6 +2647,7 @@ export const wasmocrqualitythresholds_alnumWsRatioThreshold: (a: number) => numb
 export const wasmocrqualitythresholds_criticalFragmentedWordRatio: (a: number) => number;
 export const wasmocrqualitythresholds_default: () => number;
 export const wasmocrqualitythresholds_discardSuspectedOcrNoise: (a: number) => number;
+export const wasmocrqualitythresholds_enablePlausibilityOcrRouting: (a: number) => number;
 export const wasmocrqualitythresholds_enableProvenanceOcrRouting: (a: number) => number;
 export const wasmocrqualitythresholds_maxFragmentedWordRatio: (a: number) => number;
 export const wasmocrqualitythresholds_maxOcrOutputDictInvalidWordRatio: (a: number) => number;
@@ -2656,17 +2661,19 @@ export const wasmocrqualitythresholds_minMeaningfulWords: (a: number) => number;
 export const wasmocrqualitythresholds_minNonWhitespacePerPage: (a: number) => number;
 export const wasmocrqualitythresholds_minOcrMeanConfidence: (a: number) => number;
 export const wasmocrqualitythresholds_minProvenanceFallbackRatio: (a: number) => number;
+export const wasmocrqualitythresholds_minReliableLanguageChunkRatio: (a: number) => number;
 export const wasmocrqualitythresholds_minTotalNonWhitespace: (a: number) => number;
 export const wasmocrqualitythresholds_minUndecodableRatio: (a: number) => number;
 export const wasmocrqualitythresholds_minWordsForAvgLengthCheck: (a: number) => number;
 export const wasmocrqualitythresholds_minWordsForOcrOutputCheck: (a: number) => number;
 export const wasmocrqualitythresholds_minWordsForRepeatCheck: (a: number) => number;
-export const wasmocrqualitythresholds_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: number, k1: number) => number;
+export const wasmocrqualitythresholds_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: number, k1: number, l1: number, m1: number, n1: number) => number;
 export const wasmocrqualitythresholds_nonTextMinChars: (a: number) => number;
 export const wasmocrqualitythresholds_pipelineMinQuality: (a: number) => number;
 export const wasmocrqualitythresholds_set_alnumWsRatioThreshold: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_criticalFragmentedWordRatio: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_discardSuspectedOcrNoise: (a: number, b: number) => void;
+export const wasmocrqualitythresholds_set_enablePlausibilityOcrRouting: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_enableProvenanceOcrRouting: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_maxFragmentedWordRatio: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_maxOcrOutputDictInvalidWordRatio: (a: number, b: number) => void;
@@ -2680,6 +2687,7 @@ export const wasmocrqualitythresholds_set_minMeaningfulWords: (a: number, b: num
 export const wasmocrqualitythresholds_set_minNonWhitespacePerPage: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_minOcrMeanConfidence: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_minProvenanceFallbackRatio: (a: number, b: number) => void;
+export const wasmocrqualitythresholds_set_minReliableLanguageChunkRatio: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_minTotalNonWhitespace: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_minUndecodableRatio: (a: number, b: number) => void;
 export const wasmocrqualitythresholds_set_minWordsForAvgLengthCheck: (a: number, b: number) => void;
@@ -2987,17 +2995,21 @@ export const wasmpdfformfield_set_value: (a: number, b: number, c: number) => vo
 export const wasmpdfformfield_tooltip: (a: number) => [number, number];
 export const wasmpdfformfield_value: (a: number) => [number, number];
 export const wasmpdfmetadata_default: () => number;
+export const wasmpdfmetadata_fabricatedTextPages: (a: number) => [number, number];
 export const wasmpdfmetadata_height: (a: number) => [number, bigint];
+export const wasmpdfmetadata_implausibleTextPages: (a: number) => [number, number];
 export const wasmpdfmetadata_isEncrypted: (a: number) => number;
 export const wasmpdfmetadata_layoutGateReasons: (a: number) => [number, number];
 export const wasmpdfmetadata_layoutGatedPages: (a: number) => [number, number];
-export const wasmpdfmetadata_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint, h: number, i: bigint, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number) => number;
+export const wasmpdfmetadata_new: (a: number, b: number, c: number, d: number, e: number, f: number, g: bigint, h: number, i: bigint, j: number, k: number, l: number, m: number, n: number, o: number, p: number, q: number, r: number, s: number, t: number, u: number) => number;
 export const wasmpdfmetadata_pageCount: (a: number) => number;
 export const wasmpdfmetadata_pdfVersion: (a: number) => [number, number];
 export const wasmpdfmetadata_producer: (a: number) => [number, number];
 export const wasmpdfmetadata_scannedConfidence: (a: number) => number;
 export const wasmpdfmetadata_scannedPages: (a: number) => [number, number];
+export const wasmpdfmetadata_set_fabricatedTextPages: (a: number, b: number, c: number) => void;
 export const wasmpdfmetadata_set_height: (a: number, b: number, c: bigint) => void;
+export const wasmpdfmetadata_set_implausibleTextPages: (a: number, b: number, c: number) => void;
 export const wasmpdfmetadata_set_isEncrypted: (a: number, b: number) => void;
 export const wasmpdfmetadata_set_layoutGateReasons: (a: number, b: number, c: number) => void;
 export const wasmpdfmetadata_set_layoutGatedPages: (a: number, b: number, c: number) => void;
@@ -3821,9 +3833,9 @@ 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__h26802dab031c3123: (a: number, b: number, c: any) => [number, number];
-export const wasm_bindgen__convert__closures_____invoke__h44bdeda823314cfc: (a: number, b: number, c: any, d: any) => void;
-export const wasm_bindgen__convert__closures_____invoke__ha4d0d7342deba426: (a: number, b: number) => void;
+export const wasm_bindgen__convert__closures_____invoke__h1752b33118c61bf6: (a: number, b: number, c: any, d: any) => void;
+export const wasm_bindgen__convert__closures_____invoke__h58dcb4b03347de73: (a: number, b: number, c: any) => [number, number];
+export const wasm_bindgen__convert__closures_____invoke__hc0b53db039d89c01: (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;