@kreuzberg/wasm 4.0.0-rc.6

package/dist/index.js

@@ -0,0 +1,383 @@
+"use strict";
+var __create = Object.create;
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __getProtoOf = Object.getPrototypeOf;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toESM = (mod, isNodeMode, target) => (target = mod != null ? __create(__getProtoOf(mod)) : {}, __copyProps(
+  // If the importer is in node compatibility mode or this is not an ESM
+  // file that has been converted to a CommonJS file using a Babel-
+  // compatible transform (i.e. "__esModule" has not been set), then set
+  // "default" to the CommonJS "module.exports" for node compatibility.
+  isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target,
+  mod
+));
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+var index_exports = {};
+__export(index_exports, {
+  TesseractWasmBackend: () => import_tesseract_wasm_backend2.TesseractWasmBackend,
+  batchExtractBytes: () => batchExtractBytes,
+  batchExtractBytesSync: () => batchExtractBytesSync,
+  batchExtractFiles: () => batchExtractFiles,
+  clearOcrBackends: () => import_registry2.clearOcrBackends,
+  configToJS: () => import_wasm_adapter2.configToJS,
+  detectRuntime: () => import_runtime2.detectRuntime,
+  enableOcr: () => enableOcr,
+  extractBytes: () => extractBytes,
+  extractBytesSync: () => extractBytesSync,
+  extractFile: () => extractFile,
+  extractFromFile: () => extractFromFile,
+  fileToUint8Array: () => import_wasm_adapter2.fileToUint8Array,
+  getInitializationError: () => getInitializationError,
+  getOcrBackend: () => import_registry2.getOcrBackend,
+  getRuntimeInfo: () => import_runtime2.getRuntimeInfo,
+  getRuntimeVersion: () => import_runtime2.getRuntimeVersion,
+  getVersion: () => getVersion,
+  getWasmCapabilities: () => import_runtime2.getWasmCapabilities,
+  hasBigInt: () => import_runtime2.hasBigInt,
+  hasBlob: () => import_runtime2.hasBlob,
+  hasFileApi: () => import_runtime2.hasFileApi,
+  hasModuleWorkers: () => import_runtime2.hasModuleWorkers,
+  hasSharedArrayBuffer: () => import_runtime2.hasSharedArrayBuffer,
+  hasWasm: () => import_runtime2.hasWasm,
+  hasWasmStreaming: () => import_runtime2.hasWasmStreaming,
+  hasWorkers: () => import_runtime2.hasWorkers,
+  initWasm: () => initWasm,
+  isBrowser: () => import_runtime2.isBrowser,
+  isBun: () => import_runtime2.isBun,
+  isDeno: () => import_runtime2.isDeno,
+  isInitialized: () => isInitialized,
+  isNode: () => import_runtime2.isNode,
+  isServerEnvironment: () => import_runtime2.isServerEnvironment,
+  isValidExtractionResult: () => import_wasm_adapter2.isValidExtractionResult,
+  isWebEnvironment: () => import_runtime2.isWebEnvironment,
+  jsToExtractionResult: () => import_wasm_adapter2.jsToExtractionResult,
+  listOcrBackends: () => import_registry2.listOcrBackends,
+  registerOcrBackend: () => import_registry2.registerOcrBackend,
+  unregisterOcrBackend: () => import_registry2.unregisterOcrBackend,
+  wrapWasmError: () => import_wasm_adapter2.wrapWasmError
+});
+module.exports = __toCommonJS(index_exports);
+var import_wasm_adapter = require("./adapters/wasm-adapter.js");
+var import_registry = require("./ocr/registry.js");
+var import_tesseract_wasm_backend = require("./ocr/tesseract-wasm-backend.js");
+var import_runtime = require("./runtime.js");
+var import_wasm_adapter2 = require("./adapters/wasm-adapter.js");
+var import_registry2 = require("./ocr/registry.js");
+var import_tesseract_wasm_backend2 = require("./ocr/tesseract-wasm-backend.js");
+var import_runtime2 = require("./runtime.js");
+let wasm = null;
+let initialized = false;
+let initializationError = null;
+let initializationPromise = null;
+async function initWasm() {
+  if (initialized) {
+    return;
+  }
+  if (initializationPromise) {
+    return initializationPromise;
+  }
+  initializationPromise = (async () => {
+    try {
+      if (!(0, import_runtime.hasWasm)()) {
+        throw new Error("WebAssembly is not supported in this environment");
+      }
+      let wasmModule;
+      try {
+        wasmModule = await import("../../pkg/kreuzberg_wasm");
+      } catch {
+        wasmModule = await import("./kreuzberg_wasm");
+      }
+      wasm = wasmModule;
+      if (wasm && typeof wasm.default === "function") {
+        await wasm.default();
+      }
+      initialized = true;
+      initializationError = null;
+    } catch (error) {
+      initializationError = error instanceof Error ? error : new Error(String(error));
+      throw (0, import_wasm_adapter.wrapWasmError)(error, "initializing Kreuzberg WASM module");
+    }
+  })();
+  return initializationPromise;
+}
+function isInitialized() {
+  return initialized;
+}
+function getVersion() {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  return wasm.version();
+}
+function getInitializationError() {
+  return initializationError;
+}
+async function extractBytes(data, mimeType, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    if (!data || data.length === 0) {
+      throw new Error("Document data cannot be empty");
+    }
+    if (!mimeType) {
+      throw new Error("MIME type is required");
+    }
+    const normalizedConfig = (0, import_wasm_adapter.configToJS)(config ?? null);
+    const result = await wasm.extractBytes(data, mimeType, normalizedConfig);
+    if (!result) {
+      throw new Error("Invalid extraction result: no result from WASM module");
+    }
+    return (0, import_wasm_adapter.jsToExtractionResult)(result);
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, "extracting from bytes");
+  }
+}
+async function extractFile(path, mimeType, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    if (!path) {
+      throw new Error("File path is required");
+    }
+    const runtime = (0, import_runtime.detectRuntime)();
+    if (runtime === "browser") {
+      throw new Error("Use extractBytes with fileToUint8Array for browser environments");
+    }
+    let fileData;
+    if (runtime === "node") {
+      const { readFile } = await import("node:fs/promises");
+      const buffer = await readFile(path);
+      fileData = new Uint8Array(buffer);
+    } else if (runtime === "deno") {
+      const deno = globalThis.Deno;
+      fileData = await deno.readFile(path);
+    } else if (runtime === "bun") {
+      const { readFile } = await import("node:fs/promises");
+      const buffer = await readFile(path);
+      fileData = new Uint8Array(buffer);
+    } else {
+      throw new Error(`Unsupported runtime for file extraction: ${runtime}`);
+    }
+    let detectedMimeType = mimeType;
+    if (!detectedMimeType) {
+      detectedMimeType = wasm.detectMimeFromBytes(fileData);
+    }
+    if (!detectedMimeType) {
+      throw new Error("Could not detect MIME type for file. Please provide mimeType parameter.");
+    }
+    detectedMimeType = wasm.normalizeMimeType(detectedMimeType);
+    return await extractBytes(fileData, detectedMimeType, config);
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, `extracting from file: ${path}`);
+  }
+}
+async function extractFromFile(file, mimeType, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    const bytes = await (0, import_wasm_adapter.fileToUint8Array)(file);
+    let type = mimeType ?? (file instanceof File ? file.type : "application/octet-stream");
+    type = wasm.normalizeMimeType(type);
+    return await extractBytes(bytes, type, config);
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, `extracting from ${file instanceof File ? "file" : "blob"}`);
+  }
+}
+function extractBytesSync(data, mimeType, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    if (!data || data.length === 0) {
+      throw new Error("Document data cannot be empty");
+    }
+    if (!mimeType) {
+      throw new Error("MIME type is required");
+    }
+    const normalizedConfig = (0, import_wasm_adapter.configToJS)(config ?? null);
+    const result = wasm.extractBytesSync(data, mimeType, normalizedConfig);
+    if (!result) {
+      throw new Error("Invalid extraction result: no result from WASM module");
+    }
+    return (0, import_wasm_adapter.jsToExtractionResult)(result);
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, "extracting from bytes (sync)");
+  }
+}
+async function batchExtractBytes(files, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    if (!Array.isArray(files)) {
+      throw new Error("Files parameter must be an array");
+    }
+    if (files.length === 0) {
+      throw new Error("Files array cannot be empty");
+    }
+    const dataList = [];
+    const mimeTypes = [];
+    for (let i = 0; i < files.length; i += 1) {
+      const file = files[i];
+      if (!file || typeof file !== "object") {
+        throw new Error(`Invalid file at index ${i}: must be an object with data and mimeType`);
+      }
+      const f = file;
+      if (!(f.data instanceof Uint8Array)) {
+        throw new Error(`Invalid file at index ${i}: data must be Uint8Array`);
+      }
+      if (typeof f.mimeType !== "string") {
+        throw new Error(`Invalid file at index ${i}: mimeType must be a string`);
+      }
+      if (f.data.length === 0) {
+        throw new Error(`Invalid file at index ${i}: data cannot be empty`);
+      }
+      dataList.push(f.data);
+      mimeTypes.push(f.mimeType);
+    }
+    const normalizedConfig = (0, import_wasm_adapter.configToJS)(config ?? null);
+    const results = await wasm.batchExtractBytes(dataList, mimeTypes, normalizedConfig);
+    if (!Array.isArray(results)) {
+      throw new Error("Invalid batch extraction result: expected array");
+    }
+    return results.map((result, index) => {
+      if (!result) {
+        throw new Error(`Invalid extraction result at index ${index}: no result from WASM module`);
+      }
+      return (0, import_wasm_adapter.jsToExtractionResult)(result);
+    });
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, "batch extracting from bytes");
+  }
+}
+function batchExtractBytesSync(files, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!wasm) {
+    throw new Error("WASM module not loaded. Call initWasm() first.");
+  }
+  try {
+    if (!Array.isArray(files)) {
+      throw new Error("Files parameter must be an array");
+    }
+    if (files.length === 0) {
+      throw new Error("Files array cannot be empty");
+    }
+    const dataList = [];
+    const mimeTypes = [];
+    for (let i = 0; i < files.length; i += 1) {
+      const file = files[i];
+      if (!file || typeof file !== "object") {
+        throw new Error(`Invalid file at index ${i}: must be an object with data and mimeType`);
+      }
+      const f = file;
+      if (!(f.data instanceof Uint8Array)) {
+        throw new Error(`Invalid file at index ${i}: data must be Uint8Array`);
+      }
+      if (typeof f.mimeType !== "string") {
+        throw new Error(`Invalid file at index ${i}: mimeType must be a string`);
+      }
+      if (f.data.length === 0) {
+        throw new Error(`Invalid file at index ${i}: data cannot be empty`);
+      }
+      dataList.push(f.data);
+      mimeTypes.push(f.mimeType);
+    }
+    const normalizedConfig = (0, import_wasm_adapter.configToJS)(config ?? null);
+    const results = wasm.batchExtractBytesSync(dataList, mimeTypes, normalizedConfig);
+    if (!Array.isArray(results)) {
+      throw new Error("Invalid batch extraction result: expected array");
+    }
+    return results.map((result, index) => {
+      if (!result) {
+        throw new Error(`Invalid extraction result at index ${index}: no result from WASM module`);
+      }
+      return (0, import_wasm_adapter.jsToExtractionResult)(result);
+    });
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, "batch extracting from bytes (sync)");
+  }
+}
+async function batchExtractFiles(files, config) {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  try {
+    if (!Array.isArray(files)) {
+      throw new Error("Files parameter must be an array");
+    }
+    if (files.length === 0) {
+      throw new Error("Files array cannot be empty");
+    }
+    const byteFiles = [];
+    for (let i = 0; i < files.length; i += 1) {
+      const file = files[i];
+      if (!(file instanceof File)) {
+        throw new Error(`Invalid file at index ${i}: must be a File object`);
+      }
+      const bytes = await (0, import_wasm_adapter.fileToUint8Array)(file);
+      byteFiles.push({
+        data: bytes,
+        mimeType: file.type || "application/octet-stream"
+      });
+    }
+    return await batchExtractBytes(byteFiles, config);
+  } catch (error) {
+    throw (0, import_wasm_adapter.wrapWasmError)(error, "batch extracting from files");
+  }
+}
+async function enableOcr() {
+  if (!initialized) {
+    throw new Error("WASM module not initialized. Call initWasm() first.");
+  }
+  if (!(0, import_runtime.isBrowser)()) {
+    throw new Error(
+      "OCR is only available in browser environments. TesseractWasmBackend requires Web Workers and createImageBitmap."
+    );
+  }
+  try {
+    const backend = new import_tesseract_wasm_backend.TesseractWasmBackend();
+    await backend.initialize();
+    (0, import_registry.registerOcrBackend)(backend);
+  } catch (error) {
+    const message = error instanceof Error ? error.message : String(error);
+    throw new Error(`Failed to enable OCR: ${message}`);
+  }
+}
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../typescript/index.ts"],"sourcesContent":["/**\n * Kreuzberg - WebAssembly Bindings for Browser and Runtime Environments\n *\n * This module provides WebAssembly bindings for Kreuzberg document intelligence,\n * enabling high-performance document extraction in browser and JavaScript runtime environments.\n *\n * ## Features\n *\n * - Extract text, metadata, and tables from documents\n * - Support for multiple document formats (PDF, Office, images, etc.)\n * - Browser and runtime-compatible WASM bindings\n * - Type-safe TypeScript interfaces\n * - Runtime detection and feature capability checking\n * - Automatic type conversion and error handling\n *\n * ## Installation\n *\n * ```bash\n * npm install @kreuzberg/wasm\n * ```\n *\n * ## Basic Usage\n *\n * ```typescript\n * import { extractBytes, initWasm } from '@kreuzberg/wasm';\n *\n * // Initialize WASM module once at app startup\n * await initWasm();\n *\n * // Extract from bytes\n * const bytes = new Uint8Array(buffer);\n * const result = await extractBytes(bytes, 'application/pdf');\n * console.log(result.content);\n * ```\n *\n * ## Browser Usage with File Input\n *\n * ```typescript\n * import { extractBytes, initWasm } from '@kreuzberg/wasm';\n * import { fileToUint8Array } from '@kreuzberg/wasm/adapters/wasm-adapter';\n *\n * // Initialize once at app startup\n * await initWasm();\n *\n * // Handle file input\n * const fileInput = document.getElementById('file');\n * fileInput.addEventListener('change', async (e) => {\n *   const file = e.target.files?.[0];\n *   if (file) {\n *     const bytes = await fileToUint8Array(file);\n *     const result = await extractBytes(bytes, file.type);\n *     console.log(result.content);\n *   }\n * });\n * ```\n *\n * ## Runtime Detection\n *\n * ```typescript\n * import { detectRuntime, getWasmCapabilities } from '@kreuzberg/wasm/runtime';\n *\n * const runtime = detectRuntime();\n * const caps = getWasmCapabilities();\n *\n * if (caps.hasWorkers) {\n *   // Can use Web Workers for parallel processing\n * }\n * ```\n *\n * ## Configuration\n *\n * ```typescript\n * import { extractBytes, initWasm } from '@kreuzberg/wasm';\n * import type { ExtractionConfig } from '@kreuzberg/wasm';\n *\n * await initWasm();\n *\n * const config: ExtractionConfig = {\n *   ocr: {\n *     backend: 'tesseract',\n *     language: 'eng'\n *   },\n *   chunking: {\n *     maxChars: 1000,\n *     chunkOverlap: 100\n *   },\n *   images: {\n *     extractImages: true,\n *     targetDpi: 150\n *   }\n * };\n *\n * const result = await extractBytes(bytes, 'application/pdf', config);\n * ```\n */\n\nimport type { ExtractionConfig as ExtractionConfigType, ExtractionResult } from \"./types.js\";\n\nimport { configToJS, fileToUint8Array, jsToExtractionResult, wrapWasmError } from \"./adapters/wasm-adapter.js\";\nimport { registerOcrBackend } from \"./ocr/registry.js\";\nimport { TesseractWasmBackend } from \"./ocr/tesseract-wasm-backend.js\";\nimport { detectRuntime, hasWasm, isBrowser } from \"./runtime.js\";\n\nexport type {\n\tChunk,\n\tChunkingConfig,\n\tExtractedImage,\n\tExtractionConfig,\n\tExtractionResult,\n\tTable,\n} from \"./types.js\";\n\nexport type * from \"./types.js\";\n\n// Re-export adapter utilities for convenient access\nexport {\n\tconfigToJS,\n\tfileToUint8Array,\n\tisValidExtractionResult,\n\tjsToExtractionResult,\n\twrapWasmError,\n} from \"./adapters/wasm-adapter.js\";\n\n// Re-export OCR registry\nexport {\n\tclearOcrBackends,\n\tgetOcrBackend,\n\tlistOcrBackends,\n\tregisterOcrBackend,\n\tunregisterOcrBackend,\n} from \"./ocr/registry.js\";\n\n// Re-export OCR backends\nexport { TesseractWasmBackend } from \"./ocr/tesseract-wasm-backend.js\";\n\n// Re-export runtime utilities\nexport {\n\tdetectRuntime,\n\tgetRuntimeInfo,\n\tgetRuntimeVersion,\n\tgetWasmCapabilities,\n\thasBigInt,\n\thasBlob,\n\thasFileApi,\n\thasModuleWorkers,\n\thasSharedArrayBuffer,\n\thasWasm,\n\thasWasmStreaming,\n\thasWorkers,\n\tisBrowser,\n\tisBun,\n\tisDeno,\n\tisNode,\n\tisServerEnvironment,\n\tisWebEnvironment,\n\ttype RuntimeType,\n\ttype WasmCapabilities,\n} from \"./runtime.js\";\n\n// WASM binding imports (to be populated during build)\n// Type definition for the WASM module interface\ntype WasmModule = {\n\t// Extraction functions\n\textractBytes: (data: Uint8Array, mimeType: string, config: Record<string, unknown> | null) => Promise<unknown>;\n\textractBytesSync: (data: Uint8Array, mimeType: string, config: Record<string, unknown> | null) => unknown;\n\tbatchExtractBytes: (\n\t\tdataList: Uint8Array[],\n\t\tmimeTypes: string[],\n\t\tconfig: Record<string, unknown> | null,\n\t) => Promise<unknown>;\n\tbatchExtractBytesSync: (\n\t\tdataList: Uint8Array[],\n\t\tmimeTypes: string[],\n\t\tconfig: Record<string, unknown> | null,\n\t) => unknown;\n\textractFile: (file: File, mimeType: string | null, config: Record<string, unknown> | null) => Promise<unknown>;\n\tbatchExtractFiles: (files: File[], config: Record<string, unknown> | null) => Promise<unknown>;\n\n\t// MIME and file type utilities\n\tdetectMimeFromBytes: (data: Uint8Array) => string;\n\tnormalizeMimeType: (mimeType: string) => string;\n\tgetMimeFromExtension: (extension: string) => string | null;\n\tgetExtensionsForMime: (mimeType: string) => string[];\n\n\t// Configuration\n\tloadConfigFromString: (content: string, format: string) => Record<string, unknown>;\n\tdiscoverConfig: () => Record<string, unknown>;\n\n\t// Module information\n\tversion: () => string;\n\tget_module_info: () => ModuleInfo;\n\n\t// OCR backend management\n\tregister_ocr_backend: (backend: unknown) => void;\n\tunregister_ocr_backend: (name: string) => void;\n\tlist_ocr_backends: () => string[];\n\tclear_ocr_backends: () => void;\n\n\t// Post-processor management\n\tregister_post_processor: (processor: unknown) => void;\n\tunregister_post_processor: (name: string) => void;\n\tlist_post_processors: () => string[];\n\tclear_post_processors: () => void;\n\n\t// Validator management\n\tregister_validator: (validator: unknown) => void;\n\tunregister_validator: (name: string) => void;\n\tlist_validators: () => string[];\n\tclear_validators: () => void;\n\n\t// PDF/WASM utilities\n\tinitialize_pdfium_render: (pdfiumWasmModule: unknown, localWasmModule: unknown, debug: boolean) => boolean;\n\tread_block_from_callback_wasm: (param: number, position: number, pBuf: number, size: number) => number;\n\twrite_block_from_callback_wasm: (param: number, buf: number, size: number) => number;\n\n\t// Initialization\n\tdefault?: () => Promise<void>;\n};\n\n// ModuleInfo type definition\ntype ModuleInfo = {\n\tname: () => string;\n\tversion: () => string;\n\tfree: () => void;\n};\n\nlet wasm: WasmModule | null = null;\n\n/** Initialize flag */\nlet initialized = false;\n\n/** Initialization error (if any) */\nlet initializationError: Error | null = null;\n\n/** Initialization promise for handling concurrent init calls */\nlet initializationPromise: Promise<void> | null = null;\n\n/**\n * Initialize the WASM module\n *\n * This function must be called once before using any extraction functions.\n * It loads and initializes the WASM module in the current runtime environment,\n * automatically selecting the appropriate WASM variant for the detected runtime.\n *\n * Multiple calls to initWasm() are safe and will return immediately if already initialized.\n *\n * @throws {Error} If WASM module fails to load or is not supported in the current environment\n *\n * @example Basic Usage\n * ```typescript\n * import { initWasm } from '@kreuzberg/wasm';\n *\n * async function main() {\n *   await initWasm();\n *   // Now you can use extraction functions\n * }\n *\n * main().catch(console.error);\n * ```\n *\n * @example With Error Handling\n * ```typescript\n * import { initWasm, getWasmCapabilities } from '@kreuzberg/wasm';\n *\n * async function initializeKreuzberg() {\n *   const caps = getWasmCapabilities();\n *   if (!caps.hasWasm) {\n *     throw new Error('WebAssembly is not supported in this environment');\n *   }\n *\n *   try {\n *     await initWasm();\n *     console.log('Kreuzberg initialized successfully');\n *   } catch (error) {\n *     console.error('Failed to initialize Kreuzberg:', error);\n *     throw error;\n *   }\n * }\n * ```\n */\nexport async function initWasm(): Promise<void> {\n\tif (initialized) {\n\t\treturn;\n\t}\n\n\t// Handle concurrent init calls\n\tif (initializationPromise) {\n\t\treturn initializationPromise;\n\t}\n\n\tinitializationPromise = (async () => {\n\t\ttry {\n\t\t\t// Check WASM support\n\t\t\tif (!hasWasm()) {\n\t\t\t\tthrow new Error(\"WebAssembly is not supported in this environment\");\n\t\t\t}\n\n\t\t\t// Dynamic WASM import and initialization\n\t\t\tlet wasmModule: unknown;\n\t\t\ttry {\n\t\t\t\t// Try importing from pkg directory (development/monorepo layout)\n\t\t\t\t// @ts-expect-error - Path may not exist at type-check time\n\t\t\t\twasmModule = await import(\"../../pkg/kreuzberg_wasm\");\n\t\t\t} catch {\n\t\t\t\t// Fallback to dist-relative path (published package layout)\n\t\t\t\t// @ts-expect-error - Dynamic import path\n\t\t\t\twasmModule = await import(\"./kreuzberg_wasm\");\n\t\t\t}\n\t\t\twasm = wasmModule as unknown as WasmModule;\n\n\t\t\t// Call default initialization if available (for some wasm-pack targets)\n\t\t\tif (wasm && typeof wasm.default === \"function\") {\n\t\t\t\tawait wasm.default();\n\t\t\t}\n\n\t\t\tinitialized = true;\n\t\t\tinitializationError = null;\n\t\t} catch (error) {\n\t\t\tinitializationError = error instanceof Error ? error : new Error(String(error));\n\t\t\tthrow wrapWasmError(error, \"initializing Kreuzberg WASM module\");\n\t\t}\n\t})();\n\n\treturn initializationPromise;\n}\n\n/**\n * Check if WASM module is initialized\n *\n * @returns True if WASM module is initialized, false otherwise\n *\n * @example\n * ```typescript\n * if (!isInitialized()) {\n *   await initWasm();\n * }\n * ```\n */\nexport function isInitialized(): boolean {\n\treturn initialized;\n}\n\n/**\n * Get WASM module version\n *\n * @throws {Error} If WASM module is not initialized\n * @returns The version string of the WASM module\n *\n * @example\n * ```typescript\n * const version = getVersion();\n * console.log(`Using Kreuzberg ${version}`);\n * ```\n */\nexport function getVersion(): string {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\treturn wasm.version();\n}\n\n/**\n * Get initialization error if module failed to load\n *\n * @returns The error that occurred during initialization, or null if no error\n *\n * @internal\n */\nexport function getInitializationError(): Error | null {\n\treturn initializationError;\n}\n\n/**\n * Extract content from bytes (document data)\n *\n * Extracts text, metadata, tables, images, and other content from document bytes.\n * Automatically detects document type from MIME type and applies appropriate extraction logic.\n *\n * @param data - The document bytes to extract from\n * @param mimeType - MIME type of the document (e.g., 'application/pdf', 'image/jpeg')\n * @param config - Optional extraction configuration\n * @returns Promise resolving to the extraction result\n * @throws {Error} If WASM module is not initialized or extraction fails\n *\n * @example Extract PDF\n * ```typescript\n * const bytes = new Uint8Array(buffer);\n * const result = await extractBytes(bytes, 'application/pdf');\n * console.log(result.content);\n * console.log(result.tables);\n * ```\n *\n * @example Extract with Configuration\n * ```typescript\n * const result = await extractBytes(bytes, 'application/pdf', {\n *   ocr: {\n *     backend: 'tesseract',\n *     language: 'deu' // German\n *   },\n *   images: {\n *     extractImages: true,\n *     targetDpi: 200\n *   }\n * });\n * ```\n *\n * @example Extract from File\n * ```typescript\n * const file = inputEvent.target.files[0];\n * const bytes = await fileToUint8Array(file);\n * const result = await extractBytes(bytes, file.type);\n * ```\n */\nexport async function extractBytes(\n\tdata: Uint8Array,\n\tmimeType: string,\n\tconfig?: ExtractionConfigType | null,\n): Promise<ExtractionResult> {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\t// Validate input\n\t\tif (!data || data.length === 0) {\n\t\t\tthrow new Error(\"Document data cannot be empty\");\n\t\t}\n\n\t\tif (!mimeType) {\n\t\t\tthrow new Error(\"MIME type is required\");\n\t\t}\n\n\t\t// Normalize config for WASM\n\t\tconst normalizedConfig = configToJS(config ?? null);\n\n\t\t// Call WASM function\n\t\tconst result = await wasm.extractBytes(data, mimeType, normalizedConfig);\n\n\t\t// Validate result\n\t\tif (!result) {\n\t\t\tthrow new Error(\"Invalid extraction result: no result from WASM module\");\n\t\t}\n\n\t\t// Convert and return result\n\t\treturn jsToExtractionResult(result);\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, \"extracting from bytes\");\n\t}\n}\n\n/**\n * Extract content from a file on the file system\n *\n * Node.js and Deno specific function that reads a file from the file system\n * and extracts content from it. Automatically detects MIME type if not provided.\n *\n * @param path - Path to the file to extract from\n * @param mimeType - Optional MIME type of the file. If not provided, will attempt to detect\n * @param config - Optional extraction configuration\n * @returns Promise resolving to the extraction result\n * @throws {Error} If WASM module is not initialized, file doesn't exist, or extraction fails\n *\n * @example Extract with auto-detection\n * ```typescript\n * const result = await extractFile('./document.pdf');\n * console.log(result.content);\n * ```\n *\n * @example Extract with explicit MIME type\n * ```typescript\n * const result = await extractFile('./document.docx', 'application/vnd.openxmlformats-officedocument.wordprocessingml.document');\n * ```\n *\n * @example Extract from Node.js with config\n * ```typescript\n * import { extractFile } from '@kreuzberg/wasm';\n * import { readFile } from 'fs/promises';\n *\n * const result = await extractFile('./report.xlsx', null, {\n *   chunking: {\n *     maxChars: 1000\n *   }\n * });\n * ```\n */\nexport async function extractFile(\n\tpath: string,\n\tmimeType?: string | null,\n\tconfig?: ExtractionConfigType | null,\n): Promise<ExtractionResult> {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\tif (!path) {\n\t\t\tthrow new Error(\"File path is required\");\n\t\t}\n\n\t\t// This function is only suitable for Node.js/Deno/Bun environments\n\t\t// Browser environments should use extractBytes with fileToUint8Array\n\t\tconst runtime = detectRuntime();\n\t\tif (runtime === \"browser\") {\n\t\t\tthrow new Error(\"Use extractBytes with fileToUint8Array for browser environments\");\n\t\t}\n\n\t\t// Read file based on runtime\n\t\tlet fileData: Uint8Array;\n\n\t\tif (runtime === \"node\") {\n\t\t\t// Node.js: use dynamic import to avoid issues in non-Node.js environments\n\t\t\tconst { readFile } = await import(\"node:fs/promises\");\n\t\t\tconst buffer = await readFile(path);\n\t\t\tfileData = new Uint8Array(buffer);\n\t\t} else if (runtime === \"deno\") {\n\t\t\t// Deno: use Deno.readFile\n\t\t\tconst deno = (globalThis as Record<string, unknown>).Deno as {\n\t\t\t\treadFile: (path: string) => Promise<Uint8Array>;\n\t\t\t};\n\t\t\tfileData = await deno.readFile(path);\n\t\t} else if (runtime === \"bun\") {\n\t\t\t// Bun: use dynamic import for fs/promises (compatible with Node.js API)\n\t\t\tconst { readFile } = await import(\"node:fs/promises\");\n\t\t\tconst buffer = await readFile(path);\n\t\t\tfileData = new Uint8Array(buffer);\n\t\t} else {\n\t\t\tthrow new Error(`Unsupported runtime for file extraction: ${runtime}`);\n\t\t}\n\n\t\t// Detect MIME type if not provided\n\t\tlet detectedMimeType = mimeType;\n\t\tif (!detectedMimeType) {\n\t\t\tdetectedMimeType = wasm.detectMimeFromBytes(fileData);\n\t\t}\n\n\t\tif (!detectedMimeType) {\n\t\t\tthrow new Error(\"Could not detect MIME type for file. Please provide mimeType parameter.\");\n\t\t}\n\n\t\t// Normalize MIME type\n\t\tdetectedMimeType = wasm.normalizeMimeType(detectedMimeType);\n\n\t\t// Call extractBytes with the file contents\n\t\treturn await extractBytes(fileData, detectedMimeType, config);\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, `extracting from file: ${path}`);\n\t}\n}\n\n/**\n * Extract content from a File or Blob (browser-friendly wrapper)\n *\n * Convenience function that wraps fileToUint8Array and extractBytes,\n * providing a streamlined API for browser applications handling file inputs.\n *\n * @param file - The File or Blob to extract from\n * @param mimeType - Optional MIME type. If not provided, uses file.type if available\n * @param config - Optional extraction configuration\n * @returns Promise resolving to the extraction result\n * @throws {Error} If WASM module is not initialized or extraction fails\n *\n * @example Simple file extraction\n * ```typescript\n * const fileInput = document.getElementById('file');\n * fileInput.addEventListener('change', async (e) => {\n *   const file = e.target.files?.[0];\n *   if (file) {\n *     const result = await extractFromFile(file);\n *     console.log(result.content);\n *   }\n * });\n * ```\n *\n * @example With configuration\n * ```typescript\n * const result = await extractFromFile(file, file.type, {\n *   chunking: { maxChars: 1000 },\n *   images: { extractImages: true }\n * });\n * ```\n */\nexport async function extractFromFile(\n\tfile: File | Blob,\n\tmimeType?: string | null,\n\tconfig?: ExtractionConfigType | null,\n): Promise<ExtractionResult> {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\tconst bytes = await fileToUint8Array(file);\n\t\tlet type = mimeType ?? (file instanceof File ? file.type : \"application/octet-stream\");\n\n\t\t// Normalize MIME type\n\t\ttype = wasm.normalizeMimeType(type);\n\n\t\treturn await extractBytes(bytes, type, config);\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, `extracting from ${file instanceof File ? \"file\" : \"blob\"}`);\n\t}\n}\n\n/**\n * Extract content from bytes synchronously\n *\n * Synchronous version of extractBytes. Performs extraction without async operations.\n * Note: Some extraction features may still be async internally, but the wrapper is synchronous.\n *\n * @param data - The document bytes to extract from\n * @param mimeType - MIME type of the document\n * @param config - Optional extraction configuration\n * @returns The extraction result\n * @throws {Error} If WASM module is not initialized or extraction fails\n *\n * @example\n * ```typescript\n * const bytes = new Uint8Array(buffer);\n * const result = extractBytesSync(bytes, 'application/pdf');\n * console.log(result.content);\n * ```\n */\nexport function extractBytesSync(\n\tdata: Uint8Array,\n\tmimeType: string,\n\tconfig?: ExtractionConfigType | null,\n): ExtractionResult {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\t// Validate input\n\t\tif (!data || data.length === 0) {\n\t\t\tthrow new Error(\"Document data cannot be empty\");\n\t\t}\n\n\t\tif (!mimeType) {\n\t\t\tthrow new Error(\"MIME type is required\");\n\t\t}\n\n\t\t// Normalize config for WASM\n\t\tconst normalizedConfig = configToJS(config ?? null);\n\n\t\t// Call WASM function\n\t\tconst result = wasm.extractBytesSync(data, mimeType, normalizedConfig);\n\n\t\t// Validate result\n\t\tif (!result) {\n\t\t\tthrow new Error(\"Invalid extraction result: no result from WASM module\");\n\t\t}\n\n\t\t// Convert and return result\n\t\treturn jsToExtractionResult(result);\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, \"extracting from bytes (sync)\");\n\t}\n}\n\n/**\n * Batch extract content from multiple byte arrays asynchronously\n *\n * Extracts content from multiple documents in a single batch operation,\n * allowing for more efficient processing of multiple files.\n *\n * @param files - Array of objects containing data (Uint8Array) and mimeType (string)\n * @param config - Optional extraction configuration applied to all files\n * @returns Promise resolving to array of extraction results\n * @throws {Error} If WASM module is not initialized or extraction fails\n *\n * @example\n * ```typescript\n * const files = [\n *   { data: pdfBytes, mimeType: 'application/pdf' },\n *   { data: docxBytes, mimeType: 'application/vnd.openxmlformats-officedocument.wordprocessingml.document' }\n * ];\n * const results = await batchExtractBytes(files);\n * results.forEach((result) => console.log(result.content));\n * ```\n */\nexport async function batchExtractBytes(\n\tfiles: Array<{ data: Uint8Array; mimeType: string }>,\n\tconfig?: ExtractionConfigType | null,\n): Promise<ExtractionResult[]> {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\t// Validate input\n\t\tif (!Array.isArray(files)) {\n\t\t\tthrow new Error(\"Files parameter must be an array\");\n\t\t}\n\n\t\tif (files.length === 0) {\n\t\t\tthrow new Error(\"Files array cannot be empty\");\n\t\t}\n\n\t\t// Extract data arrays and MIME types\n\t\tconst dataList: Uint8Array[] = [];\n\t\tconst mimeTypes: string[] = [];\n\n\t\tfor (let i = 0; i < files.length; i += 1) {\n\t\t\tconst file = files[i];\n\t\t\tif (!file || typeof file !== \"object\") {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: must be an object with data and mimeType`);\n\t\t\t}\n\n\t\t\tconst f = file as Record<string, unknown>;\n\n\t\t\tif (!(f.data instanceof Uint8Array)) {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: data must be Uint8Array`);\n\t\t\t}\n\n\t\t\tif (typeof f.mimeType !== \"string\") {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: mimeType must be a string`);\n\t\t\t}\n\n\t\t\tif (f.data.length === 0) {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: data cannot be empty`);\n\t\t\t}\n\n\t\t\tdataList.push(f.data);\n\t\t\tmimeTypes.push(f.mimeType);\n\t\t}\n\n\t\t// Normalize config for WASM\n\t\tconst normalizedConfig = configToJS(config ?? null);\n\n\t\t// Call WASM function\n\t\tconst results = await wasm.batchExtractBytes(dataList, mimeTypes, normalizedConfig);\n\n\t\t// Validate results\n\t\tif (!Array.isArray(results)) {\n\t\t\tthrow new Error(\"Invalid batch extraction result: expected array\");\n\t\t}\n\n\t\t// Convert each result\n\t\treturn results.map((result, index) => {\n\t\t\tif (!result) {\n\t\t\t\tthrow new Error(`Invalid extraction result at index ${index}: no result from WASM module`);\n\t\t\t}\n\n\t\t\treturn jsToExtractionResult(result);\n\t\t});\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, \"batch extracting from bytes\");\n\t}\n}\n\n/**\n * Batch extract content from multiple byte arrays synchronously\n *\n * Synchronous version of batchExtractBytes. Extracts content from multiple documents\n * in a single batch operation without async operations.\n *\n * @param files - Array of objects containing data (Uint8Array) and mimeType (string)\n * @param config - Optional extraction configuration applied to all files\n * @returns Array of extraction results\n * @throws {Error} If WASM module is not initialized or extraction fails\n *\n * @example\n * ```typescript\n * const files = [\n *   { data: pdfBytes, mimeType: 'application/pdf' },\n *   { data: docxBytes, mimeType: 'application/vnd.openxmlformats-officedocument.wordprocessingml.document' }\n * ];\n * const results = batchExtractBytesSync(files);\n * results.forEach((result) => console.log(result.content));\n * ```\n */\nexport function batchExtractBytesSync(\n\tfiles: Array<{ data: Uint8Array; mimeType: string }>,\n\tconfig?: ExtractionConfigType | null,\n): ExtractionResult[] {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\tif (!wasm) {\n\t\tthrow new Error(\"WASM module not loaded. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\t// Validate input\n\t\tif (!Array.isArray(files)) {\n\t\t\tthrow new Error(\"Files parameter must be an array\");\n\t\t}\n\n\t\tif (files.length === 0) {\n\t\t\tthrow new Error(\"Files array cannot be empty\");\n\t\t}\n\n\t\t// Extract data arrays and MIME types\n\t\tconst dataList: Uint8Array[] = [];\n\t\tconst mimeTypes: string[] = [];\n\n\t\tfor (let i = 0; i < files.length; i += 1) {\n\t\t\tconst file = files[i];\n\t\t\tif (!file || typeof file !== \"object\") {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: must be an object with data and mimeType`);\n\t\t\t}\n\n\t\t\tconst f = file as Record<string, unknown>;\n\n\t\t\tif (!(f.data instanceof Uint8Array)) {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: data must be Uint8Array`);\n\t\t\t}\n\n\t\t\tif (typeof f.mimeType !== \"string\") {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: mimeType must be a string`);\n\t\t\t}\n\n\t\t\tif (f.data.length === 0) {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: data cannot be empty`);\n\t\t\t}\n\n\t\t\tdataList.push(f.data);\n\t\t\tmimeTypes.push(f.mimeType);\n\t\t}\n\n\t\t// Normalize config for WASM\n\t\tconst normalizedConfig = configToJS(config ?? null);\n\n\t\t// Call WASM function\n\t\tconst results = wasm.batchExtractBytesSync(dataList, mimeTypes, normalizedConfig);\n\n\t\t// Validate results\n\t\tif (!Array.isArray(results)) {\n\t\t\tthrow new Error(\"Invalid batch extraction result: expected array\");\n\t\t}\n\n\t\t// Convert each result\n\t\treturn results.map((result, index) => {\n\t\t\tif (!result) {\n\t\t\t\tthrow new Error(`Invalid extraction result at index ${index}: no result from WASM module`);\n\t\t\t}\n\n\t\t\treturn jsToExtractionResult(result);\n\t\t});\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, \"batch extracting from bytes (sync)\");\n\t}\n}\n\n/**\n * Batch extract content from multiple File objects asynchronously\n *\n * Convenience function that converts File objects to Uint8Array and calls batchExtractBytes.\n * Automatically uses the file.type as MIME type if available.\n *\n * @param files - Array of File objects to extract from\n * @param config - Optional extraction configuration applied to all files\n * @returns Promise resolving to array of extraction results\n * @throws {Error} If WASM module is not initialized, files cannot be read, or extraction fails\n *\n * @example\n * ```typescript\n * const fileInput = document.getElementById('files');\n * const files = Array.from(fileInput.files ?? []);\n * const results = await batchExtractFiles(files);\n * results.forEach((result, index) => {\n *   console.log(`File ${index}: ${result.content.substring(0, 50)}...`);\n * });\n * ```\n */\nexport async function batchExtractFiles(\n\tfiles: File[],\n\tconfig?: ExtractionConfigType | null,\n): Promise<ExtractionResult[]> {\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\ttry {\n\t\t// Validate input\n\t\tif (!Array.isArray(files)) {\n\t\t\tthrow new Error(\"Files parameter must be an array\");\n\t\t}\n\n\t\tif (files.length === 0) {\n\t\t\tthrow new Error(\"Files array cannot be empty\");\n\t\t}\n\n\t\t// Convert all files to Uint8Array and collect MIME types\n\t\tconst byteFiles: Array<{ data: Uint8Array; mimeType: string }> = [];\n\n\t\tfor (let i = 0; i < files.length; i += 1) {\n\t\t\tconst file = files[i];\n\t\t\tif (!(file instanceof File)) {\n\t\t\t\tthrow new Error(`Invalid file at index ${i}: must be a File object`);\n\t\t\t}\n\n\t\t\tconst bytes = await fileToUint8Array(file);\n\t\t\tbyteFiles.push({\n\t\t\t\tdata: bytes,\n\t\t\t\tmimeType: file.type || \"application/octet-stream\",\n\t\t\t});\n\t\t}\n\n\t\t// Call batchExtractBytes with converted files\n\t\treturn await batchExtractBytes(byteFiles, config);\n\t} catch (error) {\n\t\tthrow wrapWasmError(error, \"batch extracting from files\");\n\t}\n}\n\n/**\n * Enable OCR functionality with tesseract-wasm backend\n *\n * Convenience function that automatically initializes and registers the Tesseract WASM backend.\n * This is the recommended approach for enabling OCR in WASM-based applications.\n *\n * ## Browser Requirement\n *\n * This function requires a browser environment with support for:\n * - WebWorkers (for Tesseract processing)\n * - createImageBitmap (for image conversion)\n * - Blob API\n *\n * ## Network Requirement\n *\n * Training data will be loaded from jsDelivr CDN on first use of each language.\n * Ensure network access to cdn.jsdelivr.net is available.\n *\n * @throws {Error} If not in browser environment or tesseract-wasm is not available\n *\n * @example Basic Usage\n * ```typescript\n * import { enableOcr, extractBytes, initWasm } from '@kreuzberg/wasm';\n *\n * async function main() {\n *   // Initialize WASM module\n *   await initWasm();\n *\n *   // Enable OCR with tesseract-wasm\n *   await enableOcr();\n *\n *   // Now you can use OCR in extraction\n *   const imageBytes = new Uint8Array(buffer);\n *   const result = await extractBytes(imageBytes, 'image/png', {\n *     ocr: { backend: 'tesseract-wasm', language: 'eng' }\n *   });\n *\n *   console.log(result.content); // Extracted text\n * }\n *\n * main().catch(console.error);\n * ```\n *\n * @example With Progress Tracking\n * ```typescript\n * import { enableOcr, TesseractWasmBackend } from '@kreuzberg/wasm';\n *\n * async function setupOcrWithProgress() {\n *   const backend = new TesseractWasmBackend();\n *   backend.setProgressCallback((progress) => {\n *     console.log(`OCR Progress: ${progress}%`);\n *     updateProgressBar(progress);\n *   });\n *\n *   await backend.initialize();\n *   registerOcrBackend(backend);\n * }\n *\n * setupOcrWithProgress().catch(console.error);\n * ```\n *\n * @example Multiple Languages\n * ```typescript\n * import { enableOcr, extractBytes, initWasm } from '@kreuzberg/wasm';\n *\n * await initWasm();\n * await enableOcr();\n *\n * // Extract English text\n * const englishResult = await extractBytes(engImageBytes, 'image/png', {\n *   ocr: { backend: 'tesseract-wasm', language: 'eng' }\n * });\n *\n * // Extract German text - model is cached after first use\n * const germanResult = await extractBytes(deImageBytes, 'image/png', {\n *   ocr: { backend: 'tesseract-wasm', language: 'deu' }\n * });\n * ```\n */\nexport async function enableOcr(): Promise<void> {\n\t// Check if WASM module is initialized\n\tif (!initialized) {\n\t\tthrow new Error(\"WASM module not initialized. Call initWasm() first.\");\n\t}\n\n\t// Check if in browser environment\n\tif (!isBrowser()) {\n\t\tthrow new Error(\n\t\t\t\"OCR is only available in browser environments. TesseractWasmBackend requires Web Workers and createImageBitmap.\",\n\t\t);\n\t}\n\n\ttry {\n\t\t// Create and initialize backend\n\t\tconst backend = new TesseractWasmBackend();\n\t\tawait backend.initialize();\n\n\t\t// Register the backend\n\t\tregisterOcrBackend(backend);\n\t} catch (error) {\n\t\tconst message = error instanceof Error ? error.message : String(error);\n\t\tthrow new Error(`Failed to enable OCR: ${message}`);\n\t}\n}\n"],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAkGA,0BAAkF;AAClF,sBAAmC;AACnC,oCAAqC;AACrC,qBAAkD;AAclD,IAAAA,uBAMO;AAGP,IAAAC,mBAMO;AAGP,IAAAC,iCAAqC;AAGrC,IAAAC,kBAqBO;AAqEP,IAAI,OAA0B;AAG9B,IAAI,cAAc;AAGlB,IAAI,sBAAoC;AAGxC,IAAI,wBAA8C;AA6ClD,eAAsB,WAA0B;AAC/C,MAAI,aAAa;AAChB;AAAA,EACD;AAGA,MAAI,uBAAuB;AAC1B,WAAO;AAAA,EACR;AAEA,2BAAyB,YAAY;AACpC,QAAI;AAEH,UAAI,KAAC,wBAAQ,GAAG;AACf,cAAM,IAAI,MAAM,kDAAkD;AAAA,MACnE;AAGA,UAAI;AACJ,UAAI;AAGH,qBAAa,MAAM,OAAO,0BAA0B;AAAA,MACrD,QAAQ;AAGP,qBAAa,MAAM,OAAO,kBAAkB;AAAA,MAC7C;AACA,aAAO;AAGP,UAAI,QAAQ,OAAO,KAAK,YAAY,YAAY;AAC/C,cAAM,KAAK,QAAQ;AAAA,MACpB;AAEA,oBAAc;AACd,4BAAsB;AAAA,IACvB,SAAS,OAAO;AACf,4BAAsB,iBAAiB,QAAQ,QAAQ,IAAI,MAAM,OAAO,KAAK,CAAC;AAC9E,gBAAM,mCAAc,OAAO,oCAAoC;AAAA,IAChE;AAAA,EACD,GAAG;AAEH,SAAO;AACR;AAcO,SAAS,gBAAyB;AACxC,SAAO;AACR;AAcO,SAAS,aAAqB;AACpC,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,SAAO,KAAK,QAAQ;AACrB;AASO,SAAS,yBAAuC;AACtD,SAAO;AACR;AA2CA,eAAsB,aACrB,MACA,UACA,QAC4B;AAC5B,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AAEH,QAAI,CAAC,QAAQ,KAAK,WAAW,GAAG;AAC/B,YAAM,IAAI,MAAM,+BAA+B;AAAA,IAChD;AAEA,QAAI,CAAC,UAAU;AACd,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACxC;AAGA,UAAM,uBAAmB,gCAAW,UAAU,IAAI;AAGlD,UAAM,SAAS,MAAM,KAAK,aAAa,MAAM,UAAU,gBAAgB;AAGvE,QAAI,CAAC,QAAQ;AACZ,YAAM,IAAI,MAAM,uDAAuD;AAAA,IACxE;AAGA,eAAO,0CAAqB,MAAM;AAAA,EACnC,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,uBAAuB;AAAA,EACnD;AACD;AAqCA,eAAsB,YACrB,MACA,UACA,QAC4B;AAC5B,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AACH,QAAI,CAAC,MAAM;AACV,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACxC;AAIA,UAAM,cAAU,8BAAc;AAC9B,QAAI,YAAY,WAAW;AAC1B,YAAM,IAAI,MAAM,iEAAiE;AAAA,IAClF;AAGA,QAAI;AAEJ,QAAI,YAAY,QAAQ;AAEvB,YAAM,EAAE,SAAS,IAAI,MAAM,OAAO,kBAAkB;AACpD,YAAM,SAAS,MAAM,SAAS,IAAI;AAClC,iBAAW,IAAI,WAAW,MAAM;AAAA,IACjC,WAAW,YAAY,QAAQ;AAE9B,YAAM,OAAQ,WAAuC;AAGrD,iBAAW,MAAM,KAAK,SAAS,IAAI;AAAA,IACpC,WAAW,YAAY,OAAO;AAE7B,YAAM,EAAE,SAAS,IAAI,MAAM,OAAO,kBAAkB;AACpD,YAAM,SAAS,MAAM,SAAS,IAAI;AAClC,iBAAW,IAAI,WAAW,MAAM;AAAA,IACjC,OAAO;AACN,YAAM,IAAI,MAAM,4CAA4C,OAAO,EAAE;AAAA,IACtE;AAGA,QAAI,mBAAmB;AACvB,QAAI,CAAC,kBAAkB;AACtB,yBAAmB,KAAK,oBAAoB,QAAQ;AAAA,IACrD;AAEA,QAAI,CAAC,kBAAkB;AACtB,YAAM,IAAI,MAAM,yEAAyE;AAAA,IAC1F;AAGA,uBAAmB,KAAK,kBAAkB,gBAAgB;AAG1D,WAAO,MAAM,aAAa,UAAU,kBAAkB,MAAM;AAAA,EAC7D,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,yBAAyB,IAAI,EAAE;AAAA,EAC3D;AACD;AAkCA,eAAsB,gBACrB,MACA,UACA,QAC4B;AAC5B,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AACH,UAAM,QAAQ,UAAM,sCAAiB,IAAI;AACzC,QAAI,OAAO,aAAa,gBAAgB,OAAO,KAAK,OAAO;AAG3D,WAAO,KAAK,kBAAkB,IAAI;AAElC,WAAO,MAAM,aAAa,OAAO,MAAM,MAAM;AAAA,EAC9C,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,mBAAmB,gBAAgB,OAAO,SAAS,MAAM,EAAE;AAAA,EACvF;AACD;AAqBO,SAAS,iBACf,MACA,UACA,QACmB;AACnB,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AAEH,QAAI,CAAC,QAAQ,KAAK,WAAW,GAAG;AAC/B,YAAM,IAAI,MAAM,+BAA+B;AAAA,IAChD;AAEA,QAAI,CAAC,UAAU;AACd,YAAM,IAAI,MAAM,uBAAuB;AAAA,IACxC;AAGA,UAAM,uBAAmB,gCAAW,UAAU,IAAI;AAGlD,UAAM,SAAS,KAAK,iBAAiB,MAAM,UAAU,gBAAgB;AAGrE,QAAI,CAAC,QAAQ;AACZ,YAAM,IAAI,MAAM,uDAAuD;AAAA,IACxE;AAGA,eAAO,0CAAqB,MAAM;AAAA,EACnC,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,8BAA8B;AAAA,EAC1D;AACD;AAuBA,eAAsB,kBACrB,OACA,QAC8B;AAC9B,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AAEH,QAAI,CAAC,MAAM,QAAQ,KAAK,GAAG;AAC1B,YAAM,IAAI,MAAM,kCAAkC;AAAA,IACnD;AAEA,QAAI,MAAM,WAAW,GAAG;AACvB,YAAM,IAAI,MAAM,6BAA6B;AAAA,IAC9C;AAGA,UAAM,WAAyB,CAAC;AAChC,UAAM,YAAsB,CAAC;AAE7B,aAAS,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK,GAAG;AACzC,YAAM,OAAO,MAAM,CAAC;AACpB,UAAI,CAAC,QAAQ,OAAO,SAAS,UAAU;AACtC,cAAM,IAAI,MAAM,yBAAyB,CAAC,4CAA4C;AAAA,MACvF;AAEA,YAAM,IAAI;AAEV,UAAI,EAAE,EAAE,gBAAgB,aAAa;AACpC,cAAM,IAAI,MAAM,yBAAyB,CAAC,2BAA2B;AAAA,MACtE;AAEA,UAAI,OAAO,EAAE,aAAa,UAAU;AACnC,cAAM,IAAI,MAAM,yBAAyB,CAAC,6BAA6B;AAAA,MACxE;AAEA,UAAI,EAAE,KAAK,WAAW,GAAG;AACxB,cAAM,IAAI,MAAM,yBAAyB,CAAC,wBAAwB;AAAA,MACnE;AAEA,eAAS,KAAK,EAAE,IAAI;AACpB,gBAAU,KAAK,EAAE,QAAQ;AAAA,IAC1B;AAGA,UAAM,uBAAmB,gCAAW,UAAU,IAAI;AAGlD,UAAM,UAAU,MAAM,KAAK,kBAAkB,UAAU,WAAW,gBAAgB;AAGlF,QAAI,CAAC,MAAM,QAAQ,OAAO,GAAG;AAC5B,YAAM,IAAI,MAAM,iDAAiD;AAAA,IAClE;AAGA,WAAO,QAAQ,IAAI,CAAC,QAAQ,UAAU;AACrC,UAAI,CAAC,QAAQ;AACZ,cAAM,IAAI,MAAM,sCAAsC,KAAK,8BAA8B;AAAA,MAC1F;AAEA,iBAAO,0CAAqB,MAAM;AAAA,IACnC,CAAC;AAAA,EACF,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,6BAA6B;AAAA,EACzD;AACD;AAuBO,SAAS,sBACf,OACA,QACqB;AACrB,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI,CAAC,MAAM;AACV,UAAM,IAAI,MAAM,gDAAgD;AAAA,EACjE;AAEA,MAAI;AAEH,QAAI,CAAC,MAAM,QAAQ,KAAK,GAAG;AAC1B,YAAM,IAAI,MAAM,kCAAkC;AAAA,IACnD;AAEA,QAAI,MAAM,WAAW,GAAG;AACvB,YAAM,IAAI,MAAM,6BAA6B;AAAA,IAC9C;AAGA,UAAM,WAAyB,CAAC;AAChC,UAAM,YAAsB,CAAC;AAE7B,aAAS,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK,GAAG;AACzC,YAAM,OAAO,MAAM,CAAC;AACpB,UAAI,CAAC,QAAQ,OAAO,SAAS,UAAU;AACtC,cAAM,IAAI,MAAM,yBAAyB,CAAC,4CAA4C;AAAA,MACvF;AAEA,YAAM,IAAI;AAEV,UAAI,EAAE,EAAE,gBAAgB,aAAa;AACpC,cAAM,IAAI,MAAM,yBAAyB,CAAC,2BAA2B;AAAA,MACtE;AAEA,UAAI,OAAO,EAAE,aAAa,UAAU;AACnC,cAAM,IAAI,MAAM,yBAAyB,CAAC,6BAA6B;AAAA,MACxE;AAEA,UAAI,EAAE,KAAK,WAAW,GAAG;AACxB,cAAM,IAAI,MAAM,yBAAyB,CAAC,wBAAwB;AAAA,MACnE;AAEA,eAAS,KAAK,EAAE,IAAI;AACpB,gBAAU,KAAK,EAAE,QAAQ;AAAA,IAC1B;AAGA,UAAM,uBAAmB,gCAAW,UAAU,IAAI;AAGlD,UAAM,UAAU,KAAK,sBAAsB,UAAU,WAAW,gBAAgB;AAGhF,QAAI,CAAC,MAAM,QAAQ,OAAO,GAAG;AAC5B,YAAM,IAAI,MAAM,iDAAiD;AAAA,IAClE;AAGA,WAAO,QAAQ,IAAI,CAAC,QAAQ,UAAU;AACrC,UAAI,CAAC,QAAQ;AACZ,cAAM,IAAI,MAAM,sCAAsC,KAAK,8BAA8B;AAAA,MAC1F;AAEA,iBAAO,0CAAqB,MAAM;AAAA,IACnC,CAAC;AAAA,EACF,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,oCAAoC;AAAA,EAChE;AACD;AAuBA,eAAsB,kBACrB,OACA,QAC8B;AAC9B,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAEA,MAAI;AAEH,QAAI,CAAC,MAAM,QAAQ,KAAK,GAAG;AAC1B,YAAM,IAAI,MAAM,kCAAkC;AAAA,IACnD;AAEA,QAAI,MAAM,WAAW,GAAG;AACvB,YAAM,IAAI,MAAM,6BAA6B;AAAA,IAC9C;AAGA,UAAM,YAA2D,CAAC;AAElE,aAAS,IAAI,GAAG,IAAI,MAAM,QAAQ,KAAK,GAAG;AACzC,YAAM,OAAO,MAAM,CAAC;AACpB,UAAI,EAAE,gBAAgB,OAAO;AAC5B,cAAM,IAAI,MAAM,yBAAyB,CAAC,yBAAyB;AAAA,MACpE;AAEA,YAAM,QAAQ,UAAM,sCAAiB,IAAI;AACzC,gBAAU,KAAK;AAAA,QACd,MAAM;AAAA,QACN,UAAU,KAAK,QAAQ;AAAA,MACxB,CAAC;AAAA,IACF;AAGA,WAAO,MAAM,kBAAkB,WAAW,MAAM;AAAA,EACjD,SAAS,OAAO;AACf,cAAM,mCAAc,OAAO,6BAA6B;AAAA,EACzD;AACD;AAiFA,eAAsB,YAA2B;AAEhD,MAAI,CAAC,aAAa;AACjB,UAAM,IAAI,MAAM,qDAAqD;AAAA,EACtE;AAGA,MAAI,KAAC,0BAAU,GAAG;AACjB,UAAM,IAAI;AAAA,MACT;AAAA,IACD;AAAA,EACD;AAEA,MAAI;AAEH,UAAM,UAAU,IAAI,mDAAqB;AACzC,UAAM,QAAQ,WAAW;AAGzB,4CAAmB,OAAO;AAAA,EAC3B,SAAS,OAAO;AACf,UAAM,UAAU,iBAAiB,QAAQ,MAAM,UAAU,OAAO,KAAK;AACrE,UAAM,IAAI,MAAM,yBAAyB,OAAO,EAAE;AAAA,EACnD;AACD;","names":["import_wasm_adapter","import_registry","import_tesseract_wasm_backend","import_runtime"]}