@kreuzberg/wasm 4.0.0-rc.6 → 4.0.1

package/dist/kreuzberg_wasm.mjs

@@ -1,48 +0,0 @@
-import { createRequire } from 'module';
-import { fileURLToPath } from 'url';
-import { dirname, join } from 'path';
-
-const __filename = fileURLToPath(import.meta.url);
-const __dirname = dirname(__filename);
-const require = createRequire(import.meta.url);
-
-// Import the CommonJS module
-const wasmModule = require('./kreuzberg_wasm.js');
-
-// Re-export everything from the CommonJS module as ESM
-export const {
-  memory,
-  extractBytes,
-  extractBytesSync,
-  batchExtractBytes,
-  batchExtractBytesSync,
-  extractFile,
-  batchExtractFiles,
-  detectMimeFromBytes,
-  normalizeMimeType,
-  getMimeFromExtension,
-  getExtensionsForMime,
-  loadConfigFromString,
-  discoverConfig,
-  version,
-  get_module_info,
-  register_ocr_backend,
-  unregister_ocr_backend,
-  list_ocr_backends,
-  clear_ocr_backends,
-  register_post_processor,
-  unregister_post_processor,
-  list_post_processors,
-  clear_post_processors,
-  register_validator,
-  unregister_validator,
-  list_validators,
-  clear_validators,
-  initialize_pdfium_render,
-  read_block_from_callback_wasm,
-  write_block_from_callback_wasm,
-  default: wasmDefault
-} = wasmModule;
-
-// Support default export pattern
-export default wasmModule;

package/dist/ocr/registry.d.mts

@@ -1,102 +0,0 @@
-import { f as OcrBackendProtocol } from '../types-GJVIvbPy.mjs';
-
-/**
- * OCR Backend Registry
- *
- * Provides a registry for OCR backends in the WASM environment.
- * This enables auto-registration and management of OCR backends.
- *
- * Note: The WASM package provides a lightweight registry in the browser.
- * For more advanced features like Rust integration, use @kreuzberg/node or @kreuzberg/deno.
- *
- * @example
- * ```typescript
- * import { TesseractWasmBackend } from '@kreuzberg/wasm/ocr/tesseract-wasm-backend';
- * import { enableOcr } from '@kreuzberg/wasm';
- *
- * // Simple auto-registration
- * await enableOcr();
- * ```
- */
-
-/**
- * Register an OCR backend
- *
- * Registers an OCR backend with the WASM extraction pipeline.
- * If a backend with the same name is already registered, it will be replaced.
- *
- * @param backend - OCR backend implementing OcrBackendProtocol
- * @throws {Error} If backend validation fails
- *
- * @example
- * ```typescript
- * import { TesseractWasmBackend } from '@kreuzberg/wasm/ocr/tesseract-wasm-backend';
- * import { registerOcrBackend } from '@kreuzberg/wasm/ocr/registry';
- *
- * const backend = new TesseractWasmBackend();
- * await backend.initialize();
- * registerOcrBackend(backend);
- * ```
- */
-declare function registerOcrBackend(backend: OcrBackendProtocol): void;
-/**
- * Get a registered OCR backend by name
- *
- * @param name - Backend name
- * @returns The OCR backend or undefined if not found
- *
- * @example
- * ```typescript
- * import { getOcrBackend } from '@kreuzberg/wasm/ocr/registry';
- *
- * const backend = getOcrBackend('tesseract-wasm');
- * if (backend) {
- *   console.log('Available languages:', backend.supportedLanguages());
- * }
- * ```
- */
-declare function getOcrBackend(name: string): OcrBackendProtocol | undefined;
-/**
- * List all registered OCR backends
- *
- * @returns Array of registered backend names
- *
- * @example
- * ```typescript
- * import { listOcrBackends } from '@kreuzberg/wasm/ocr/registry';
- *
- * const backends = listOcrBackends();
- * console.log('Available OCR backends:', backends);
- * ```
- */
-declare function listOcrBackends(): string[];
-/**
- * Unregister an OCR backend
- *
- * @param name - Backend name to unregister
- * @throws {Error} If backend is not found
- *
- * @example
- * ```typescript
- * import { unregisterOcrBackend } from '@kreuzberg/wasm/ocr/registry';
- *
- * unregisterOcrBackend('tesseract-wasm');
- * ```
- */
-declare function unregisterOcrBackend(name: string): Promise<void>;
-/**
- * Clear all registered OCR backends
- *
- * Unregisters all OCR backends and calls their shutdown methods.
- *
- * @example
- * ```typescript
- * import { clearOcrBackends } from '@kreuzberg/wasm/ocr/registry';
- *
- * // Clean up all backends when shutting down
- * await clearOcrBackends();
- * ```
- */
-declare function clearOcrBackends(): Promise<void>;
-
-export { clearOcrBackends, getOcrBackend, listOcrBackends, registerOcrBackend, unregisterOcrBackend };

package/dist/ocr/registry.mjs

@@ -1,70 +0,0 @@
-const ocrBackendRegistry = /* @__PURE__ */ new Map();
-function registerOcrBackend(backend) {
-  if (!backend) {
-    throw new Error("Backend cannot be null or undefined");
-  }
-  if (typeof backend.name !== "function") {
-    throw new Error("Backend must implement name() method");
-  }
-  if (typeof backend.supportedLanguages !== "function") {
-    throw new Error("Backend must implement supportedLanguages() method");
-  }
-  if (typeof backend.processImage !== "function") {
-    throw new Error("Backend must implement processImage() method");
-  }
-  const backendName = backend.name();
-  if (!backendName || typeof backendName !== "string") {
-    throw new Error("Backend name must be a non-empty string");
-  }
-  if (ocrBackendRegistry.has(backendName)) {
-    console.warn(`OCR backend "${backendName}" is already registered and will be replaced`);
-  }
-  ocrBackendRegistry.set(backendName, backend);
-}
-function getOcrBackend(name) {
-  return ocrBackendRegistry.get(name);
-}
-function listOcrBackends() {
-  return Array.from(ocrBackendRegistry.keys());
-}
-async function unregisterOcrBackend(name) {
-  const backend = ocrBackendRegistry.get(name);
-  if (!backend) {
-    throw new Error(
-      `OCR backend "${name}" is not registered. Available backends: ${Array.from(ocrBackendRegistry.keys()).join(", ")}`
-    );
-  }
-  if (typeof backend.shutdown === "function") {
-    try {
-      await backend.shutdown();
-    } catch (error) {
-      console.warn(
-        `Error shutting down OCR backend "${name}": ${error instanceof Error ? error.message : String(error)}`
-      );
-    }
-  }
-  ocrBackendRegistry.delete(name);
-}
-async function clearOcrBackends() {
-  const backends = Array.from(ocrBackendRegistry.entries());
-  for (const [name, backend] of backends) {
-    if (typeof backend.shutdown === "function") {
-      try {
-        await backend.shutdown();
-      } catch (error) {
-        console.warn(
-          `Error shutting down OCR backend "${name}": ${error instanceof Error ? error.message : String(error)}`
-        );
-      }
-    }
-  }
-  ocrBackendRegistry.clear();
-}
-export {
-  clearOcrBackends,
-  getOcrBackend,
-  listOcrBackends,
-  registerOcrBackend,
-  unregisterOcrBackend
-};
-//# sourceMappingURL=registry.mjs.map

package/dist/ocr/registry.mjs.map

@@ -1 +0,0 @@
-{"version":3,"sources":["../../typescript/ocr/registry.ts"],"sourcesContent":["/**\n * OCR Backend Registry\n *\n * Provides a registry for OCR backends in the WASM environment.\n * This enables auto-registration and management of OCR backends.\n *\n * Note: The WASM package provides a lightweight registry in the browser.\n * For more advanced features like Rust integration, use @kreuzberg/node or @kreuzberg/deno.\n *\n * @example\n * ```typescript\n * import { TesseractWasmBackend } from '@kreuzberg/wasm/ocr/tesseract-wasm-backend';\n * import { enableOcr } from '@kreuzberg/wasm';\n *\n * // Simple auto-registration\n * await enableOcr();\n * ```\n */\n\nimport type { OcrBackendProtocol } from \"../types.js\";\n\n/** Global registry of OCR backends */\nconst ocrBackendRegistry = new Map<string, OcrBackendProtocol>();\n\n/**\n * Register an OCR backend\n *\n * Registers an OCR backend with the WASM extraction pipeline.\n * If a backend with the same name is already registered, it will be replaced.\n *\n * @param backend - OCR backend implementing OcrBackendProtocol\n * @throws {Error} If backend validation fails\n *\n * @example\n * ```typescript\n * import { TesseractWasmBackend } from '@kreuzberg/wasm/ocr/tesseract-wasm-backend';\n * import { registerOcrBackend } from '@kreuzberg/wasm/ocr/registry';\n *\n * const backend = new TesseractWasmBackend();\n * await backend.initialize();\n * registerOcrBackend(backend);\n * ```\n */\nexport function registerOcrBackend(backend: OcrBackendProtocol): void {\n\t// Validate backend\n\tif (!backend) {\n\t\tthrow new Error(\"Backend cannot be null or undefined\");\n\t}\n\n\tif (typeof backend.name !== \"function\") {\n\t\tthrow new Error(\"Backend must implement name() method\");\n\t}\n\n\tif (typeof backend.supportedLanguages !== \"function\") {\n\t\tthrow new Error(\"Backend must implement supportedLanguages() method\");\n\t}\n\n\tif (typeof backend.processImage !== \"function\") {\n\t\tthrow new Error(\"Backend must implement processImage() method\");\n\t}\n\n\tconst backendName = backend.name();\n\n\tif (!backendName || typeof backendName !== \"string\") {\n\t\tthrow new Error(\"Backend name must be a non-empty string\");\n\t}\n\n\t// Check for duplicate registration (allow overwriting with warning)\n\tif (ocrBackendRegistry.has(backendName)) {\n\t\tconsole.warn(`OCR backend \"${backendName}\" is already registered and will be replaced`);\n\t}\n\n\t// Register the backend\n\tocrBackendRegistry.set(backendName, backend);\n}\n\n/**\n * Get a registered OCR backend by name\n *\n * @param name - Backend name\n * @returns The OCR backend or undefined if not found\n *\n * @example\n * ```typescript\n * import { getOcrBackend } from '@kreuzberg/wasm/ocr/registry';\n *\n * const backend = getOcrBackend('tesseract-wasm');\n * if (backend) {\n *   console.log('Available languages:', backend.supportedLanguages());\n * }\n * ```\n */\nexport function getOcrBackend(name: string): OcrBackendProtocol | undefined {\n\treturn ocrBackendRegistry.get(name);\n}\n\n/**\n * List all registered OCR backends\n *\n * @returns Array of registered backend names\n *\n * @example\n * ```typescript\n * import { listOcrBackends } from '@kreuzberg/wasm/ocr/registry';\n *\n * const backends = listOcrBackends();\n * console.log('Available OCR backends:', backends);\n * ```\n */\nexport function listOcrBackends(): string[] {\n\treturn Array.from(ocrBackendRegistry.keys());\n}\n\n/**\n * Unregister an OCR backend\n *\n * @param name - Backend name to unregister\n * @throws {Error} If backend is not found\n *\n * @example\n * ```typescript\n * import { unregisterOcrBackend } from '@kreuzberg/wasm/ocr/registry';\n *\n * unregisterOcrBackend('tesseract-wasm');\n * ```\n */\nexport async function unregisterOcrBackend(name: string): Promise<void> {\n\tconst backend = ocrBackendRegistry.get(name);\n\n\tif (!backend) {\n\t\tthrow new Error(\n\t\t\t`OCR backend \"${name}\" is not registered. Available backends: ${Array.from(ocrBackendRegistry.keys()).join(\", \")}`,\n\t\t);\n\t}\n\n\t// Call shutdown if available\n\tif (typeof backend.shutdown === \"function\") {\n\t\ttry {\n\t\t\tawait backend.shutdown();\n\t\t} catch (error) {\n\t\t\tconsole.warn(\n\t\t\t\t`Error shutting down OCR backend \"${name}\": ${error instanceof Error ? error.message : String(error)}`,\n\t\t\t);\n\t\t}\n\t}\n\n\tocrBackendRegistry.delete(name);\n}\n\n/**\n * Clear all registered OCR backends\n *\n * Unregisters all OCR backends and calls their shutdown methods.\n *\n * @example\n * ```typescript\n * import { clearOcrBackends } from '@kreuzberg/wasm/ocr/registry';\n *\n * // Clean up all backends when shutting down\n * await clearOcrBackends();\n * ```\n */\nexport async function clearOcrBackends(): Promise<void> {\n\tconst backends = Array.from(ocrBackendRegistry.entries());\n\n\tfor (const [name, backend] of backends) {\n\t\tif (typeof backend.shutdown === \"function\") {\n\t\t\ttry {\n\t\t\t\tawait backend.shutdown();\n\t\t\t} catch (error) {\n\t\t\t\tconsole.warn(\n\t\t\t\t\t`Error shutting down OCR backend \"${name}\": ${error instanceof Error ? error.message : String(error)}`,\n\t\t\t\t);\n\t\t\t}\n\t\t}\n\t}\n\n\tocrBackendRegistry.clear();\n}\n"],"mappings":"AAsBA,MAAM,qBAAqB,oBAAI,IAAgC;AAqBxD,SAAS,mBAAmB,SAAmC;AAErE,MAAI,CAAC,SAAS;AACb,UAAM,IAAI,MAAM,qCAAqC;AAAA,EACtD;AAEA,MAAI,OAAO,QAAQ,SAAS,YAAY;AACvC,UAAM,IAAI,MAAM,sCAAsC;AAAA,EACvD;AAEA,MAAI,OAAO,QAAQ,uBAAuB,YAAY;AACrD,UAAM,IAAI,MAAM,oDAAoD;AAAA,EACrE;AAEA,MAAI,OAAO,QAAQ,iBAAiB,YAAY;AAC/C,UAAM,IAAI,MAAM,8CAA8C;AAAA,EAC/D;AAEA,QAAM,cAAc,QAAQ,KAAK;AAEjC,MAAI,CAAC,eAAe,OAAO,gBAAgB,UAAU;AACpD,UAAM,IAAI,MAAM,yCAAyC;AAAA,EAC1D;AAGA,MAAI,mBAAmB,IAAI,WAAW,GAAG;AACxC,YAAQ,KAAK,gBAAgB,WAAW,8CAA8C;AAAA,EACvF;AAGA,qBAAmB,IAAI,aAAa,OAAO;AAC5C;AAkBO,SAAS,cAAc,MAA8C;AAC3E,SAAO,mBAAmB,IAAI,IAAI;AACnC;AAeO,SAAS,kBAA4B;AAC3C,SAAO,MAAM,KAAK,mBAAmB,KAAK,CAAC;AAC5C;AAeA,eAAsB,qBAAqB,MAA6B;AACvE,QAAM,UAAU,mBAAmB,IAAI,IAAI;AAE3C,MAAI,CAAC,SAAS;AACb,UAAM,IAAI;AAAA,MACT,gBAAgB,IAAI,4CAA4C,MAAM,KAAK,mBAAmB,KAAK,CAAC,EAAE,KAAK,IAAI,CAAC;AAAA,IACjH;AAAA,EACD;AAGA,MAAI,OAAO,QAAQ,aAAa,YAAY;AAC3C,QAAI;AACH,YAAM,QAAQ,SAAS;AAAA,IACxB,SAAS,OAAO;AACf,cAAQ;AAAA,QACP,oCAAoC,IAAI,MAAM,iBAAiB,QAAQ,MAAM,UAAU,OAAO,KAAK,CAAC;AAAA,MACrG;AAAA,IACD;AAAA,EACD;AAEA,qBAAmB,OAAO,IAAI;AAC/B;AAeA,eAAsB,mBAAkC;AACvD,QAAM,WAAW,MAAM,KAAK,mBAAmB,QAAQ,CAAC;AAExD,aAAW,CAAC,MAAM,OAAO,KAAK,UAAU;AACvC,QAAI,OAAO,QAAQ,aAAa,YAAY;AAC3C,UAAI;AACH,cAAM,QAAQ,SAAS;AAAA,MACxB,SAAS,OAAO;AACf,gBAAQ;AAAA,UACP,oCAAoC,IAAI,MAAM,iBAAiB,QAAQ,MAAM,UAAU,OAAO,KAAK,CAAC;AAAA,QACrG;AAAA,MACD;AAAA,IACD;AAAA,EACD;AAEA,qBAAmB,MAAM;AAC1B;","names":[]}

package/dist/ocr/tesseract-wasm-backend.d.mts

@@ -1,257 +0,0 @@
-import { f as OcrBackendProtocol } from '../types-GJVIvbPy.mjs';
-
-/**
- * Tesseract WASM OCR Backend
- *
- * Provides OCR capabilities using tesseract-wasm library for browser environments.
- * Loads training data on-demand from jsDelivr CDN and implements the OcrBackendProtocol.
- *
- * ## Browser-Only Requirement
- *
- * This backend requires browser APIs like createImageBitmap and Web Workers.
- * It will NOT work in Node.js environments without additional canvas polyfills.
- *
- * ## Supported Languages
- *
- * Common ISO 639-1 and ISO 639-2 codes:
- * - English: "eng"
- * - German: "deu"
- * - French: "fra"
- * - Spanish: "spa"
- * - Italian: "ita"
- * - Portuguese: "por"
- * - Dutch: "nld"
- * - Russian: "rus"
- * - Chinese (Simplified): "chi_sim"
- * - Chinese (Traditional): "chi_tra"
- * - Japanese: "jpn"
- * - Korean: "kor"
- * - Arabic: "ara"
- * - Hindi: "hin"
- *
- * For complete language list, see: https://github.com/naptha/tesseract.js
- *
- * @example Basic Usage
- * ```typescript
- * import { TesseractWasmBackend } from '@kreuzberg/wasm/ocr/tesseract-wasm-backend';
- * import { registerOcrBackend, extractBytes, initWasm } from '@kreuzberg/wasm';
- *
- * // Initialize
- * await initWasm();
- * const backend = new TesseractWasmBackend();
- * await backend.initialize();
- * registerOcrBackend(backend);
- *
- * // Use in extraction
- * const imageBytes = new Uint8Array(buffer);
- * const result = await extractBytes(imageBytes, 'image/png', {
- *   ocr: { backend: 'tesseract-wasm', language: 'eng' }
- * });
- * console.log(result.content); // Extracted text
- * ```
- *
- * @example With Language Auto-Detection
- * ```typescript
- * const backend = new TesseractWasmBackend();
- * await backend.initialize();
- * registerOcrBackend(backend);
- *
- * // Extract without specifying language - backend will auto-detect
- * const result = await extractBytes(imageBytes, 'image/png', {
- *   ocr: { backend: 'tesseract-wasm' } // language will auto-detect
- * });
- * ```
- */
-
-/**
- * TesseractWasmBackend - OCR backend using tesseract-wasm library
- *
- * Implements the OcrBackendProtocol for Kreuzberg document extraction pipeline.
- * Provides comprehensive OCR support with model caching, error handling, and progress reporting.
- */
-declare class TesseractWasmBackend implements OcrBackendProtocol {
-    /** Tesseract WASM client instance */
-    private client;
-    /** Track which models are currently loaded to avoid redundant loads */
-    private loadedLanguages;
-    /** Cache for language availability validation */
-    private supportedLangsCache;
-    /** Progress callback for UI updates */
-    private progressCallback;
-    /** Base URL for training data CDN */
-    private readonly CDN_BASE_URL;
-    /**
-     * Return the unique name of this OCR backend
-     *
-     * @returns Backend identifier "tesseract-wasm"
-     */
-    name(): string;
-    /**
-     * Return list of supported language codes
-     *
-     * Returns a curated list of commonly available Tesseract language models.
-     * Tesseract supports many more languages through custom models.
-     *
-     * @returns Array of ISO 639-1/2/3 language codes
-     */
-    supportedLanguages(): string[];
-    /**
-     * Initialize the OCR backend
-     *
-     * Creates the Tesseract WASM client instance. This is called once when
-     * the backend is registered with the extraction pipeline.
-     *
-     * The actual model loading happens in processImage() on-demand to avoid
-     * loading all models upfront.
-     *
-     * @throws {Error} If tesseract-wasm is not available or initialization fails
-     *
-     * @example
-     * ```typescript
-     * const backend = new TesseractWasmBackend();
-     * try {
-     *   await backend.initialize();
-     * } catch (error) {
-     *   console.error('Failed to initialize OCR:', error);
-     * }
-     * ```
-     */
-    initialize(): Promise<void>;
-    /**
-     * Process image bytes and extract text via OCR
-     *
-     * Handles image loading, model loading, OCR processing, and result formatting.
-     * Automatically loads the language model on first use and caches it for subsequent calls.
-     *
-     * @param imageBytes - Raw image data (Uint8Array) or Base64-encoded string
-     * @param language - ISO 639-2/3 language code (e.g., "eng", "deu")
-     * @returns Promise resolving to OCR result with content and metadata
-     * @throws {Error} If image processing fails, model loading fails, or language is unsupported
-     *
-     * @example
-     * ```typescript
-     * const backend = new TesseractWasmBackend();
-     * await backend.initialize();
-     *
-     * const imageBuffer = fs.readFileSync('scanned.png');
-     * const result = await backend.processImage(
-     *   new Uint8Array(imageBuffer),
-     *   'eng'
-     * );
-     *
-     * console.log(result.content); // Extracted text
-     * console.log(result.metadata.confidence); // OCR confidence score
-     * ```
-     */
-    processImage(imageBytes: Uint8Array | string, language: string): Promise<{
-        content: string;
-        mime_type: string;
-        metadata: Record<string, unknown>;
-        tables: unknown[];
-    }>;
-    /**
-     * Shutdown the OCR backend and release resources
-     *
-     * Properly cleans up the Tesseract WASM client, freeing memory and Web Workers.
-     * Called when the backend is unregistered or the application shuts down.
-     *
-     * @throws {Error} If cleanup fails (errors are logged but not critical)
-     *
-     * @example
-     * ```typescript
-     * const backend = new TesseractWasmBackend();
-     * await backend.initialize();
-     * // ... use backend ...
-     * await backend.shutdown(); // Clean up resources
-     * ```
-     */
-    shutdown(): Promise<void>;
-    /**
-     * Set a progress callback for UI updates
-     *
-     * Allows the UI to display progress during OCR processing.
-     * The callback will be called with values from 0 to 100.
-     *
-     * @param callback - Function to call with progress percentage
-     *
-     * @example
-     * ```typescript
-     * const backend = new TesseractWasmBackend();
-     * backend.setProgressCallback((progress) => {
-     *   console.log(`OCR Progress: ${progress}%`);
-     *   document.getElementById('progress-bar').style.width = `${progress}%`;
-     * });
-     * ```
-     */
-    setProgressCallback(callback: (progress: number) => void): void;
-    /**
-     * Load language model from CDN
-     *
-     * Fetches the training data for a specific language from jsDelivr CDN.
-     * This is an MVP approach - models are cached by the browser.
-     *
-     * @param language - ISO 639-2/3 language code
-     * @throws {Error} If model download fails or language is not available
-     *
-     * @internal
-     */
-    private loadLanguageModel;
-    /**
-     * Convert image bytes or Base64 string to ImageBitmap
-     *
-     * Handles both Uint8Array and Base64-encoded image data, converting to
-     * ImageBitmap format required by Tesseract WASM.
-     *
-     * @param imageBytes - Image data as Uint8Array or Base64 string
-     * @returns Promise resolving to ImageBitmap
-     * @throws {Error} If conversion fails (browser API not available or invalid image data)
-     *
-     * @internal
-     */
-    private convertToImageBitmap;
-    /**
-     * Get confidence score from OCR result
-     *
-     * Attempts to retrieve confidence score from Tesseract.
-     * Returns a safe default if unavailable.
-     *
-     * @returns Confidence score between 0 and 1
-     *
-     * @internal
-     */
-    private getConfidenceScore;
-    /**
-     * Get page metadata from OCR result
-     *
-     * Retrieves additional metadata like image dimensions and processing info.
-     *
-     * @returns Metadata object (may be empty if unavailable)
-     *
-     * @internal
-     */
-    private getPageMetadata;
-    /**
-     * Dynamically load tesseract-wasm module
-     *
-     * Uses dynamic import to load tesseract-wasm only when needed,
-     * avoiding hard dependency in browser environments where it may not be bundled.
-     *
-     * @returns tesseract-wasm module object
-     * @throws {Error} If module cannot be imported
-     *
-     * @internal
-     */
-    private loadTesseractWasm;
-    /**
-     * Report progress to progress callback
-     *
-     * Internal helper for notifying progress updates during OCR processing.
-     *
-     * @param progress - Progress percentage (0-100)
-     *
-     * @internal
-     */
-    private reportProgress;
-}
-
-export { TesseractWasmBackend };