@kreuzberg/wasm 4.0.0-rc.6 → 4.0.0

package/dist/types.d.ts

@@ -0,0 +1,363 @@
+/**
+ * Type definitions for Kreuzberg WASM bindings
+ *
+ * These types are generated from the Rust core library and define
+ * the interface for extraction, configuration, and results.
+ */
+/**
+ * Token reduction configuration
+ */
+export interface TokenReductionConfig {
+    /** Token reduction mode */
+    mode?: string;
+    /** Preserve important words during reduction */
+    preserveImportantWords?: boolean;
+}
+/**
+ * Post-processor configuration
+ */
+export interface PostProcessorConfig {
+    /** Whether post-processing is enabled */
+    enabled?: boolean;
+    /** List of enabled processors */
+    enabledProcessors?: string[];
+    /** List of disabled processors */
+    disabledProcessors?: string[];
+}
+/**
+ * Keyword extraction algorithm type
+ *
+ * Supported algorithms:
+ * - "yake": YAKE (Yet Another Keyword Extractor) - statistical approach
+ * - "rake": RAKE (Rapid Automatic Keyword Extraction) - co-occurrence based
+ */
+export type KeywordAlgorithm = "yake" | "rake";
+/**
+ * YAKE algorithm-specific parameters
+ */
+export interface YakeParams {
+    /** Window size for co-occurrence analysis (default: 2) */
+    windowSize?: number;
+}
+/**
+ * RAKE algorithm-specific parameters
+ */
+export interface RakeParams {
+    /** Minimum word length to consider (default: 1) */
+    minWordLength?: number;
+    /** Maximum words in a keyword phrase (default: 3) */
+    maxWordsPerPhrase?: number;
+}
+/**
+ * Keyword extraction configuration
+ *
+ * Controls how keywords are extracted from text, including algorithm selection,
+ * scoring thresholds, n-gram ranges, and language-specific settings.
+ */
+export interface KeywordConfig {
+    /** Algorithm to use for extraction (default: "yake") */
+    algorithm?: KeywordAlgorithm;
+    /** Maximum number of keywords to extract (default: 10) */
+    maxKeywords?: number;
+    /** Minimum score threshold 0.0-1.0 (default: 0.0) */
+    minScore?: number;
+    /** N-gram range [min, max] for keyword extraction (default: [1, 3]) */
+    ngramRange?: [number, number];
+    /** Language code for stopword filtering (e.g., "en", "de", "fr") */
+    language?: string;
+    /** YAKE-specific tuning parameters */
+    yakeParams?: YakeParams;
+    /** RAKE-specific tuning parameters */
+    rakeParams?: RakeParams;
+}
+/**
+ * Extracted keyword with relevance metadata
+ *
+ * Represents a single keyword extracted from text along with its relevance score,
+ * the algorithm that extracted it, and optional position information.
+ */
+export interface ExtractedKeyword {
+    /** The keyword text */
+    text: string;
+    /** Relevance score (higher is better, algorithm-specific range) */
+    score: number;
+    /** Algorithm that extracted this keyword */
+    algorithm: KeywordAlgorithm;
+    /** Optional positions where keyword appears in text (character offsets) */
+    positions?: number[];
+}
+/**
+ * Configuration for document extraction
+ */
+export interface ExtractionConfig {
+    /** OCR configuration */
+    ocr?: OcrConfig;
+    /** Chunking configuration */
+    chunking?: ChunkingConfig;
+    /** Image extraction configuration */
+    images?: ImageExtractionConfig;
+    /** Page extraction configuration */
+    pages?: PageExtractionConfig;
+    /** Language detection configuration */
+    languageDetection?: LanguageDetectionConfig;
+    /** PDF extraction options */
+    pdfOptions?: PdfConfig;
+    /** Token reduction configuration */
+    tokenReduction?: TokenReductionConfig;
+    /** Post-processor configuration */
+    postprocessor?: PostProcessorConfig;
+    /** Keyword extraction configuration */
+    keywords?: KeywordConfig;
+    /** Whether to use caching */
+    useCache?: boolean;
+    /** Enable quality processing */
+    enableQualityProcessing?: boolean;
+    /** Force OCR even if text is available */
+    forceOcr?: boolean;
+    /** Maximum concurrent extractions */
+    maxConcurrentExtractions?: number;
+}
+/**
+ * Tesseract OCR configuration
+ */
+export interface TesseractConfig {
+    /** Tesseract page segmentation mode */
+    psm?: number;
+    /** Enable table detection */
+    enableTableDetection?: boolean;
+    /** Character whitelist for recognition */
+    tesseditCharWhitelist?: string;
+}
+/**
+ * OCR configuration
+ */
+export interface OcrConfig {
+    /** OCR backend to use */
+    backend?: string;
+    /** Language codes (ISO 639) */
+    languages?: string[];
+    /** Whether to perform OCR */
+    enabled?: boolean;
+    /** Tesseract-specific configuration */
+    tesseractConfig?: TesseractConfig;
+    /** Language code for OCR */
+    language?: string;
+}
+/**
+ * Chunking configuration
+ */
+export interface ChunkingConfig {
+    /** Maximum characters per chunk */
+    maxChars?: number;
+    /** Overlap between chunks */
+    maxOverlap?: number;
+}
+/**
+ * Image extraction configuration
+ */
+export interface ImageExtractionConfig {
+    /** Whether to extract images */
+    enabled?: boolean;
+    /** Target DPI for image extraction */
+    targetDpi?: number;
+    /** Maximum image dimension in pixels */
+    maxImageDimension?: number;
+    /** Automatically adjust DPI */
+    autoAdjustDpi?: boolean;
+    /** Minimum DPI threshold */
+    minDpi?: number;
+    /** Maximum DPI threshold */
+    maxDpi?: number;
+}
+/**
+ * PDF extraction configuration
+ */
+export interface PdfConfig {
+    /** Whether to extract images from PDF */
+    extractImages?: boolean;
+    /** Passwords for encrypted PDFs */
+    passwords?: string[];
+    /** Whether to extract metadata */
+    extractMetadata?: boolean;
+}
+/**
+ * Page extraction configuration
+ */
+export interface PageExtractionConfig {
+    /** Extract pages as separate array (ExtractionResult.pages) */
+    extractPages?: boolean;
+    /** Insert page markers in main content string */
+    insertPageMarkers?: boolean;
+    /** Page marker format (use {page_num} placeholder) */
+    markerFormat?: string;
+}
+/**
+ * Language detection configuration
+ */
+export interface LanguageDetectionConfig {
+    /** Whether to detect languages */
+    enabled?: boolean;
+}
+/**
+ * Result of document extraction
+ */
+export interface ExtractionResult {
+    /** Extracted text content */
+    content: string;
+    /** MIME type of the document */
+    mimeType: string;
+    /** Document metadata */
+    metadata: Metadata;
+    /** Extracted tables */
+    tables: Table[];
+    /** Detected languages (ISO 639 codes) */
+    detectedLanguages?: string[] | null;
+    /** Text chunks when chunking is enabled */
+    chunks?: Chunk[] | null;
+    /** Extracted images */
+    images?: ExtractedImage[] | null;
+    /** Per-page content */
+    pages?: PageContent[] | null;
+    /** Extracted keywords when keyword extraction is enabled */
+    keywords?: ExtractedKeyword[] | null;
+}
+/**
+ * Document metadata
+ */
+export interface Metadata {
+    /** Document title */
+    title?: string;
+    /** Document subject or description */
+    subject?: string;
+    /** Document author(s) */
+    authors?: string[];
+    /** Keywords/tags */
+    keywords?: string[];
+    /** Primary language (ISO 639 code) */
+    language?: string;
+    /** Creation timestamp (ISO 8601 format) */
+    createdAt?: string;
+    /** Last modification timestamp (ISO 8601 format) */
+    modifiedAt?: string;
+    /** User who created the document */
+    creator?: string;
+    /** User who last modified the document */
+    lastModifiedBy?: string;
+    /** Number of pages/slides */
+    pageCount?: number;
+    /** Format-specific metadata */
+    formatMetadata?: unknown;
+    /**
+     * Additional fields may be added at runtime by postprocessors.
+     * Use bracket notation to safely access unexpected properties.
+     */
+    [key: string]: unknown;
+}
+/**
+ * Extracted table
+ */
+export interface Table {
+    /** Table cells/rows */
+    cells?: string[][];
+    /** Table markdown representation */
+    markdown?: string;
+    /** Page number if available */
+    pageNumber?: number;
+    /** Table headers */
+    headers?: string[];
+    /** Table rows */
+    rows?: string[][];
+}
+/**
+ * Chunk metadata
+ */
+export interface ChunkMetadata {
+    /** Character start position in original content */
+    charStart: number;
+    /** Character end position in original content */
+    charEnd: number;
+    /** Token count if available */
+    tokenCount: number | null;
+    /** Index of this chunk */
+    chunkIndex: number;
+    /** Total number of chunks */
+    totalChunks: number;
+}
+/**
+ * Text chunk from chunked content
+ */
+export interface Chunk {
+    /** Chunk text content */
+    content: string;
+    /** Chunk metadata */
+    metadata?: ChunkMetadata;
+    /** Character position in original content (legacy) */
+    charIndex?: number;
+    /** Token count if available (legacy) */
+    tokenCount?: number;
+    /** Embedding vector if computed */
+    embedding?: number[] | null;
+}
+/**
+ * Extracted image from document
+ */
+export interface ExtractedImage {
+    /** Image data as Uint8Array or base64 string */
+    data: Uint8Array | string;
+    /** Image format/MIME type */
+    format?: string;
+    /** MIME type of the image */
+    mimeType?: string;
+    /** Image index in document */
+    imageIndex?: number;
+    /** Page number if available */
+    pageNumber?: number | null;
+    /** Image width in pixels */
+    width?: number | null;
+    /** Image height in pixels */
+    height?: number | null;
+    /** Color space of the image */
+    colorspace?: string | null;
+    /** Bits per color component */
+    bitsPerComponent?: number | null;
+    /** Whether this is a mask image */
+    isMask?: boolean;
+    /** Image description */
+    description?: string | null;
+    /** Optional OCR result from the image */
+    ocrResult?: ExtractionResult | string | null;
+}
+/**
+ * Per-page content
+ */
+export interface PageContent {
+    /** Page number (1-indexed) */
+    pageNumber: number;
+    /** Text content of the page */
+    content: string;
+    /** Tables on this page */
+    tables?: Table[];
+    /** Images on this page */
+    images?: ExtractedImage[];
+}
+/**
+ * OCR backend protocol/interface
+ */
+export interface OcrBackendProtocol {
+    /** Get the backend name */
+    name(): string;
+    /** Get supported language codes */
+    supportedLanguages?(): string[];
+    /** Initialize the backend */
+    initialize(options?: Record<string, unknown>): void | Promise<void>;
+    /** Shutdown the backend */
+    shutdown?(): void | Promise<void>;
+    /** Process an image with OCR */
+    processImage(imageData: Uint8Array | string, language?: string): Promise<{
+        content: string;
+        mime_type: string;
+        metadata?: Record<string, unknown>;
+        tables?: unknown[];
+    } | string>;
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../typescript/types.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACpC,2BAA2B;IAC3B,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,gDAAgD;IAChD,sBAAsB,CAAC,EAAE,OAAO,CAAC;CACjC;AAED;;GAEG;AACH,MAAM,WAAW,mBAAmB;IACnC,yCAAyC;IACzC,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,iCAAiC;IACjC,iBAAiB,CAAC,EAAE,MAAM,EAAE,CAAC;IAC7B,kCAAkC;IAClC,kBAAkB,CAAC,EAAE,MAAM,EAAE,CAAC;CAC9B;AAED;;;;;;GAMG;AACH,MAAM,MAAM,gBAAgB,GAAG,MAAM,GAAG,MAAM,CAAC;AAE/C;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,0DAA0D;IAC1D,UAAU,CAAC,EAAE,MAAM,CAAC;CACpB;AAED;;GAEG;AACH,MAAM,WAAW,UAAU;IAC1B,mDAAmD;IACnD,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,qDAAqD;IACrD,iBAAiB,CAAC,EAAE,MAAM,CAAC;CAC3B;AAED;;;;;GAKG;AACH,MAAM,WAAW,aAAa;IAC7B,wDAAwD;IACxD,SAAS,CAAC,EAAE,gBAAgB,CAAC;IAC7B,0DAA0D;IAC1D,WAAW,CAAC,EAAE,MAAM,CAAC;IACrB,qDAAqD;IACrD,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,uEAAuE;IACvE,UAAU,CAAC,EAAE,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAC9B,oEAAoE;IACpE,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,sCAAsC;IACtC,UAAU,CAAC,EAAE,UAAU,CAAC;IACxB,sCAAsC;IACtC,UAAU,CAAC,EAAE,UAAU,CAAC;CACxB;AAED;;;;;GAKG;AACH,MAAM,WAAW,gBAAgB;IAChC,uBAAuB;IACvB,IAAI,EAAE,MAAM,CAAC;IACb,mEAAmE;IACnE,KAAK,EAAE,MAAM,CAAC;IACd,4CAA4C;IAC5C,SAAS,EAAE,gBAAgB,CAAC;IAC5B,2EAA2E;IAC3E,SAAS,CAAC,EAAE,MAAM,EAAE,CAAC;CACrB;AAED;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC,wBAAwB;IACxB,GAAG,CAAC,EAAE,SAAS,CAAC;IAChB,6BAA6B;IAC7B,QAAQ,CAAC,EAAE,cAAc,CAAC;IAC1B,qCAAqC;IACrC,MAAM,CAAC,EAAE,qBAAqB,CAAC;IAC/B,oCAAoC;IACpC,KAAK,CAAC,EAAE,oBAAoB,CAAC;IAC7B,uCAAuC;IACvC,iBAAiB,CAAC,EAAE,uBAAuB,CAAC;IAC5C,6BAA6B;IAC7B,UAAU,CAAC,EAAE,SAAS,CAAC;IACvB,oCAAoC;IACpC,cAAc,CAAC,EAAE,oBAAoB,CAAC;IACtC,mCAAmC;IACnC,aAAa,CAAC,EAAE,mBAAmB,CAAC;IACpC,uCAAuC;IACvC,QAAQ,CAAC,EAAE,aAAa,CAAC;IACzB,6BAA6B;IAC7B,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,gCAAgC;IAChC,uBAAuB,CAAC,EAAE,OAAO,CAAC;IAClC,0CAA0C;IAC1C,QAAQ,CAAC,EAAE,OAAO,CAAC;IACnB,qCAAqC;IACrC,wBAAwB,CAAC,EAAE,MAAM,CAAC;CAClC;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC/B,uCAAuC;IACvC,GAAG,CAAC,EAAE,MAAM,CAAC;IACb,6BAA6B;IAC7B,oBAAoB,CAAC,EAAE,OAAO,CAAC;IAC/B,0CAA0C;IAC1C,qBAAqB,CAAC,EAAE,MAAM,CAAC;CAC/B;AAED;;GAEG;AACH,MAAM,WAAW,SAAS;IACzB,yBAAyB;IACzB,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,+BAA+B;IAC/B,SAAS,CAAC,EAAE,MAAM,EAAE,CAAC;IACrB,6BAA6B;IAC7B,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,uCAAuC;IACvC,eAAe,CAAC,EAAE,eAAe,CAAC;IAClC,4BAA4B;IAC5B,QAAQ,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;GAEG;AACH,MAAM,WAAW,cAAc;IAC9B,mCAAmC;IACnC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,6BAA6B;IAC7B,UAAU,CAAC,EAAE,MAAM,CAAC;CACpB;AAED;;GAEG;AACH,MAAM,WAAW,qBAAqB;IACrC,gCAAgC;IAChC,OAAO,CAAC,EAAE,OAAO,CAAC;IAClB,sCAAsC;IACtC,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,wCAAwC;IACxC,iBAAiB,CAAC,EAAE,MAAM,CAAC;IAC3B,+BAA+B;IAC/B,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,4BAA4B;IAC5B,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,4BAA4B;IAC5B,MAAM,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;GAEG;AACH,MAAM,WAAW,SAAS;IACzB,yCAAyC;IACzC,aAAa,CAAC,EAAE,OAAO,CAAC;IACxB,mCAAmC;IACnC,SAAS,CAAC,EAAE,MAAM,EAAE,CAAC;IACrB,kCAAkC;IAClC,eAAe,CAAC,EAAE,OAAO,CAAC;CAC1B;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACpC,+DAA+D;IAC/D,YAAY,CAAC,EAAE,OAAO,CAAC;IACvB,iDAAiD;IACjD,iBAAiB,CAAC,EAAE,OAAO,CAAC;IAC5B,sDAAsD;IACtD,YAAY,CAAC,EAAE,MAAM,CAAC;CACtB;AAED;;GAEG;AACH,MAAM,WAAW,uBAAuB;IACvC,kCAAkC;IAClC,OAAO,CAAC,EAAE,OAAO,CAAC;CAClB;AAED;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAChC,6BAA6B;IAC7B,OAAO,EAAE,MAAM,CAAC;IAChB,gCAAgC;IAChC,QAAQ,EAAE,MAAM,CAAC;IACjB,wBAAwB;IACxB,QAAQ,EAAE,QAAQ,CAAC;IACnB,uBAAuB;IACvB,MAAM,EAAE,KAAK,EAAE,CAAC;IAChB,yCAAyC;IACzC,iBAAiB,CAAC,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC;IACpC,2CAA2C;IAC3C,MAAM,CAAC,EAAE,KAAK,EAAE,GAAG,IAAI,CAAC;IACxB,uBAAuB;IACvB,MAAM,CAAC,EAAE,cAAc,EAAE,GAAG,IAAI,CAAC;IACjC,uBAAuB;IACvB,KAAK,CAAC,EAAE,WAAW,EAAE,GAAG,IAAI,CAAC;IAC7B,4DAA4D;IAC5D,QAAQ,CAAC,EAAE,gBAAgB,EAAE,GAAG,IAAI,CAAC;CACrC;AAED;;GAEG;AACH,MAAM,WAAW,QAAQ;IACxB,qBAAqB;IACrB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,sCAAsC;IACtC,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,yBAAyB;IACzB,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,oBAAoB;IACpB,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,sCAAsC;IACtC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,2CAA2C;IAC3C,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,oDAAoD;IACpD,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,oCAAoC;IACpC,OAAO,CAAC,EAAE,MAAM,CAAC;IACjB,0CAA0C;IAC1C,cAAc,CAAC,EAAE,MAAM,CAAC;IACxB,6BAA6B;IAC7B,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,+BAA+B;IAC/B,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;;OAGG;IACH,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC;CACvB;AAED;;GAEG;AACH,MAAM,WAAW,KAAK;IACrB,uBAAuB;IACvB,KAAK,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;IACnB,oCAAoC;IACpC,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,+BAA+B;IAC/B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,oBAAoB;IACpB,OAAO,CAAC,EAAE,MAAM,EAAE,CAAC;IACnB,iBAAiB;IACjB,IAAI,CAAC,EAAE,MAAM,EAAE,EAAE,CAAC;CAClB;AAED;;GAEG;AACH,MAAM,WAAW,aAAa;IAC7B,mDAAmD;IACnD,SAAS,EAAE,MAAM,CAAC;IAClB,iDAAiD;IACjD,OAAO,EAAE,MAAM,CAAC;IAChB,+BAA+B;IAC/B,UAAU,EAAE,MAAM,GAAG,IAAI,CAAC;IAC1B,0BAA0B;IAC1B,UAAU,EAAE,MAAM,CAAC;IACnB,6BAA6B;IAC7B,WAAW,EAAE,MAAM,CAAC;CACpB;AAED;;GAEG;AACH,MAAM,WAAW,KAAK;IACrB,yBAAyB;IACzB,OAAO,EAAE,MAAM,CAAC;IAChB,qBAAqB;IACrB,QAAQ,CAAC,EAAE,aAAa,CAAC;IACzB,sDAAsD;IACtD,SAAS,CAAC,EAAE,MAAM,CAAC;IACnB,wCAAwC;IACxC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,mCAAmC;IACnC,SAAS,CAAC,EAAE,MAAM,EAAE,GAAG,IAAI,CAAC;CAC5B;AAED;;GAEG;AACH,MAAM,WAAW,cAAc;IAC9B,gDAAgD;IAChD,IAAI,EAAE,UAAU,GAAG,MAAM,CAAC;IAC1B,6BAA6B;IAC7B,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,6BAA6B;IAC7B,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,8BAA8B;IAC9B,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,+BAA+B;IAC/B,UAAU,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,4BAA4B;IAC5B,KAAK,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACtB,6BAA6B;IAC7B,MAAM,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACvB,+BAA+B;IAC/B,UAAU,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC3B,+BAA+B;IAC/B,gBAAgB,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IACjC,mCAAmC;IACnC,MAAM,CAAC,EAAE,OAAO,CAAC;IACjB,wBAAwB;IACxB,WAAW,CAAC,EAAE,MAAM,GAAG,IAAI,CAAC;IAC5B,yCAAyC;IACzC,SAAS,CAAC,EAAE,gBAAgB,GAAG,MAAM,GAAG,IAAI,CAAC;CAC7C;AAED;;GAEG;AACH,MAAM,WAAW,WAAW;IAC3B,8BAA8B;IAC9B,UAAU,EAAE,MAAM,CAAC;IACnB,+BAA+B;IAC/B,OAAO,EAAE,MAAM,CAAC;IAChB,0BAA0B;IAC1B,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC;IACjB,0BAA0B;IAC1B,MAAM,CAAC,EAAE,cAAc,EAAE,CAAC;CAC1B;AAED;;GAEG;AACH,MAAM,WAAW,kBAAkB;IAClC,2BAA2B;IAC3B,IAAI,IAAI,MAAM,CAAC;IACf,mCAAmC;IACnC,kBAAkB,CAAC,IAAI,MAAM,EAAE,CAAC;IAChC,6BAA6B;IAC7B,UAAU,CAAC,OAAO,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IACpE,2BAA2B;IAC3B,QAAQ,CAAC,IAAI,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAClC,gCAAgC;IAChC,YAAY,CACX,SAAS,EAAE,UAAU,GAAG,MAAM,EAC9B,QAAQ,CAAC,EAAE,MAAM,GACf,OAAO,CACP;QACA,OAAO,EAAE,MAAM,CAAC;QAChB,SAAS,EAAE,MAAM,CAAC;QAClB,QAAQ,CAAC,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;QACnC,MAAM,CAAC,EAAE,OAAO,EAAE,CAAC;KAClB,GACD,MAAM,CACR,CAAC;CACF"}

package/package.json

@@ -1,6 +1,8 @@
 {
 	"name": "@kreuzberg/wasm",
-	"version": "4.0.0-rc.6",
+	"version": "4.0.0",
+	"type": "module",
+	"packageManager": "pnpm@10.17.0",
 	"description": "Kreuzberg document intelligence - WebAssembly bindings",
 	"author": {
 		"name": "Na'aman Hirschfeld",
@@ -12,58 +14,35 @@
 		"url": "https://github.com/kreuzberg-dev/kreuzberg/issues"
 	},
 	"main": "dist/index.js",
-	"module": "dist/index.mjs",
+	"module": "dist/index.js",
 	"types": "dist/index.d.ts",
 	"exports": {
 		".": {
-			"import": {
-				"types": "./dist/index.d.mts",
-				"default": "./dist/index.mjs"
-			},
-			"require": {
-				"types": "./dist/index.d.ts",
-				"default": "./dist/index.js"
-			}
+			"types": "./dist/index.d.ts",
+			"import": "./dist/index.js",
+			"default": "./dist/index.js"
 		},
+		"./dist/pkg/*": "./dist/pkg/*",
+		"./dist/kreuzberg_wasm.js": "./dist/kreuzberg_wasm.js",
 		"./runtime": {
-			"import": {
-				"types": "./dist/runtime.d.mts",
-				"default": "./dist/runtime.mjs"
-			},
-			"require": {
-				"types": "./dist/runtime.d.ts",
-				"default": "./dist/runtime.js"
-			}
+			"types": "./dist/runtime.d.ts",
+			"import": "./dist/runtime.js",
+			"default": "./dist/runtime.js"
 		},
 		"./adapters/wasm-adapter": {
-			"import": {
-				"types": "./dist/adapters/wasm-adapter.d.mts",
-				"default": "./dist/adapters/wasm-adapter.mjs"
-			},
-			"require": {
-				"types": "./dist/adapters/wasm-adapter.d.ts",
-				"default": "./dist/adapters/wasm-adapter.js"
-			}
+			"types": "./dist/adapters/wasm-adapter.d.ts",
+			"import": "./dist/adapters/wasm-adapter.js",
+			"default": "./dist/adapters/wasm-adapter.js"
 		},
 		"./ocr/registry": {
-			"import": {
-				"types": "./dist/ocr/registry.d.mts",
-				"default": "./dist/ocr/registry.mjs"
-			},
-			"require": {
-				"types": "./dist/ocr/registry.d.ts",
-				"default": "./dist/ocr/registry.js"
-			}
+			"types": "./dist/ocr/registry.d.ts",
+			"import": "./dist/ocr/registry.js",
+			"default": "./dist/ocr/registry.js"
 		},
 		"./ocr/tesseract-wasm-backend": {
-			"import": {
-				"types": "./dist/ocr/tesseract-wasm-backend.d.mts",
-				"default": "./dist/ocr/tesseract-wasm-backend.mjs"
-			},
-			"require": {
-				"types": "./dist/ocr/tesseract-wasm-backend.d.ts",
-				"default": "./dist/ocr/tesseract-wasm-backend.js"
-			}
+			"types": "./dist/ocr/tesseract-wasm-backend.d.ts",
+			"import": "./dist/ocr/tesseract-wasm-backend.js",
+			"default": "./dist/ocr/tesseract-wasm-backend.js"
 		}
 	},
 	"repository": {
@@ -95,6 +74,7 @@
 		"pkg",
 		"*.wasm",
 		"*.d.ts",
+		"pdfium.js",
 		"README.md"
 	],
 	"engines": {
@@ -109,9 +89,12 @@
 		"build:wasm:bundler": "wasm-pack build --target bundler --out-dir pkg --release",
 		"build:wasm:nodejs": "wasm-pack build --target nodejs --out-dir pkg --release",
 		"build:wasm:deno": "wasm-pack build --target deno --out-dir pkg --release",
-		"build:ts": "tsup",
-		"build": "npm run build:wasm:nodejs && npm run build:ts",
-		"build:all": "npm run build:wasm:web && npm run build:wasm:bundler && npm run build:wasm:nodejs && npm run build:wasm:deno && npm run build:ts",
+		"build:ts": "tsup && tsc --emitDeclarationOnly && node scripts/fix-type-exports.js",
+		"build:ts:minify": "tsup --minify && tsc --emitDeclarationOnly && node scripts/fix-type-exports.js",
+		"copy:pkg": "node scripts/copy-pkg.js",
+		"build": "npm run build:wasm:nodejs && npm run build:ts && npm run copy:pkg",
+		"build:ci": "npm run build:wasm:nodejs && npm run build:ts:minify && npm run copy:pkg",
+		"build:all": "npm run build:wasm:web && npm run build:wasm:bundler && npm run build:wasm:nodejs && npm run build:wasm:deno && npm run build:ts && npm run copy:pkg",
 		"typecheck": "tsc --noEmit",
 		"lint": "biome check typescript && oxlint typescript",
 		"lint:fix": "biome check --write typescript",
@@ -123,14 +106,14 @@
 		"prepublishOnly": "npm run build"
 	},
 	"devDependencies": {
-		"@types/node": "^24.10.1",
-		"@vitest/coverage-v8": "^1.4.0",
-		"@vitest/ui": "^1.4.0",
-		"jsdom": "^24.0.0",
-		"oxlint": "^1.31.0",
+		"@types/node": "^25.0.6",
+		"@vitest/coverage-v8": "^4.0.16",
+		"@vitest/ui": "^4.0.16",
+		"jsdom": "^27.4.0",
+		"oxlint": "^1.38.0",
 		"tsup": "^8.5.1",
 		"typescript": "^5.9.3",
-		"vitest": "^1.4.0"
+		"vitest": "^4.0.16"
 	},
 	"optionalDependencies": {
 		"tesseract-wasm": "^0.11.0"

package/dist/adapters/wasm-adapter.d.mts

@@ -1,121 +0,0 @@
-import { E as ExtractionConfig, a as ExtractionResult } from '../types-GJVIvbPy.mjs';
-
-/**
- * WASM Type Adapter
- *
- * This module provides type adapters for converting between JavaScript/TypeScript
- * types and WASM-compatible types, handling File/Blob conversions, config normalization,
- * and result parsing.
- *
- * @example File Conversion
- * ```typescript
- * import { fileToUint8Array } from '@kreuzberg/wasm/adapters/wasm-adapter';
- *
- * const file = event.target.files[0];
- * const bytes = await fileToUint8Array(file);
- * const result = await extractBytes(bytes, file.type);
- * ```
- *
- * @example Config Normalization
- * ```typescript
- * import { configToJS } from '@kreuzberg/wasm/adapters/wasm-adapter';
- *
- * const config = {
- *   ocr: { backend: 'tesseract', language: 'eng' },
- *   chunking: { maxChars: 1000 }
- * };
- * const normalized = configToJS(config);
- * ```
- */
-
-/**
- * Convert a File or Blob to Uint8Array
- *
- * Handles both browser File API and server-side Blob-like objects,
- * providing a unified interface for reading binary data.
- *
- * @param file - The File or Blob to convert
- * @returns Promise resolving to the byte array
- * @throws {Error} If the file cannot be read or exceeds size limit
- *
- * @example
- * ```typescript
- * const file = document.getElementById('input').files[0];
- * const bytes = await fileToUint8Array(file);
- * const result = await extractBytes(bytes, 'application/pdf');
- * ```
- */
-declare function fileToUint8Array(file: File | Blob): Promise<Uint8Array>;
-/**
- * Normalize ExtractionConfig for WASM processing
- *
- * Converts TypeScript configuration objects to a WASM-compatible format,
- * handling null values, undefined properties, and nested structures.
- *
- * @param config - The extraction configuration or null
- * @returns Normalized configuration object suitable for WASM
- *
- * @example
- * ```typescript
- * const config: ExtractionConfig = {
- *   ocr: { backend: 'tesseract' },
- *   chunking: { maxChars: 1000 }
- * };
- * const wasmConfig = configToJS(config);
- * ```
- */
-declare function configToJS(config: ExtractionConfig | null): Record<string, unknown>;
-/**
- * Parse WASM extraction result and convert to TypeScript type
- *
- * Handles conversion of WASM-returned objects to proper ExtractionResult types,
- * including proper array conversions and type assertions for tables, chunks, and images.
- *
- * @param jsValue - The raw WASM result value
- * @returns Properly typed ExtractionResult
- * @throws {Error} If the result structure is invalid
- *
- * @example
- * ```typescript
- * const wasmResult = await wasmExtract(bytes, mimeType, config);
- * const result = jsToExtractionResult(wasmResult);
- * console.log(result.content);
- * ```
- */
-declare function jsToExtractionResult(jsValue: unknown): ExtractionResult;
-/**
- * Wrap and format WASM errors with context
- *
- * Converts WASM error messages to JavaScript Error objects with proper context
- * and stack trace information when available.
- *
- * @param error - The error from WASM
- * @param context - Additional context about what operation failed
- * @returns A formatted Error object
- *
- * @internal
- *
- * @example
- * ```typescript
- * try {
- *   await wasmExtract(bytes, mimeType);
- * } catch (error) {
- *   throw wrapWasmError(error, 'extracting document');
- * }
- * ```
- */
-declare function wrapWasmError(error: unknown, context: string): Error;
-/**
- * Validate that a WASM-returned value conforms to ExtractionResult structure
- *
- * Performs structural validation without full type checking,
- * useful for runtime validation of WASM output.
- *
- * @param value - The value to validate
- * @returns True if value appears to be a valid ExtractionResult
- *
- * @internal
- */
-declare function isValidExtractionResult(value: unknown): value is ExtractionResult;
-
-export { configToJS, fileToUint8Array, isValidExtractionResult, jsToExtractionResult, wrapWasmError };