@markitdownjs/image-ocr 0.1.0

package/dist/index.d.ts

@@ -0,0 +1,2 @@
+export { ImageOcrConverter } from "./ocr-converter.js";
+//# sourceMappingURL=index.d.ts.map

package/dist/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC"}

package/dist/index.js

@@ -0,0 +1,2 @@
+export { ImageOcrConverter } from "./ocr-converter.js";
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC"}

package/dist/ocr-converter.d.ts

@@ -0,0 +1,45 @@
+import type { Converter, ConversionInput, ConversionResult } from "@markitdownjs/shared";
+/**
+ * Converts images to text using OCR (Optical Character Recognition).
+ *
+ * Features:
+ * - Chunked base64 encoding to prevent stack overflow on large images
+ * - Language selection (defaults to English, configurable)
+ * - Image dimension extraction from PNG, JPEG, GIF, and BMP headers
+ * - OCR confidence score from tesseract.js (when available)
+ * - Magic byte detection for additional image formats
+ * - Comprehensive error messages when tesseract.js is not installed
+ *
+ * @example
+ * ```ts
+ * const converter = new ImageOcrConverter();
+ * const result = await converter.convert({
+ *   data: pngBuffer,
+ *   options: { customProperties: { ocrLanguage: "fra" } }
+ * });
+ * console.log(result.ast);  // DocumentNode with image + extracted text
+ * console.log(result.markdown);
+ * ```
+ */
+export declare class ImageOcrConverter implements Converter {
+    readonly id = "image-ocr";
+    readonly supportedMimeTypes: string[];
+    readonly supportedExtensions: string[];
+    /**
+     * Checks whether this converter can handle the given input.
+     * Validates MIME type, file extension, and magic byte signatures.
+     *
+     * @param input - The conversion input to check
+     * @returns True if this converter can process the input
+     */
+    canConvert(input: ConversionInput): Promise<boolean>;
+    /**
+     * Converts an image file into text via OCR, producing an AST with the image
+     * node and the extracted text as a paragraph.
+     *
+     * @param input - The image data as Uint8Array, ArrayBuffer, Blob, or string
+     * @returns A ConversionResult with the AST, Markdown, and image metadata
+     */
+    convert(input: ConversionInput): Promise<ConversionResult>;
+}
+//# sourceMappingURL=ocr-converter.d.ts.map

package/dist/ocr-converter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ocr-converter.d.ts","sourceRoot":"","sources":["../src/ocr-converter.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EACV,SAAS,EACT,eAAe,EACf,gBAAgB,EAGjB,MAAM,sBAAsB,CAAC;AAwL9B;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,qBAAa,iBAAkB,YAAW,SAAS;IACjD,QAAQ,CAAC,EAAE,eAAe;IAC1B,QAAQ,CAAC,kBAAkB,WAOzB;IACF,QAAQ,CAAC,mBAAmB,WAS1B;IAEF;;;;;;OAMG;IACG,UAAU,CAAC,KAAK,EAAE,eAAe,GAAG,OAAO,CAAC,OAAO,CAAC;IA2B1D;;;;;;OAMG;IACG,OAAO,CAAC,KAAK,EAAE,eAAe,GAAG,OAAO,CAAC,gBAAgB,CAAC;CAsIjE"}

package/dist/ocr-converter.js

@@ -0,0 +1,359 @@
+import { createNode, readInputData, detectMimeTypeFromData, generateId, } from "@markitdownjs/shared";
+/** Magic byte signatures for additional image format detection. */
+const IMAGE_SIGNATURES = [
+    { bytes: [0x89, 0x50, 0x4e, 0x47], mime: "image/png" }, // PNG
+    { bytes: [0xff, 0xd8, 0xff], mime: "image/jpeg" }, // JPEG
+    { bytes: [0x52, 0x49, 0x46, 0x46], mime: "image/webp" }, // RIFF (WebP)
+    { bytes: [0x47, 0x49, 0x46, 0x38], mime: "image/gif" }, // GIF8
+    { bytes: [0x42, 0x4d], mime: "image/bmp" }, // BMP
+    { bytes: [0x49, 0x49, 0x2a, 0x00], mime: "image/tiff" }, // TIFF (little-endian)
+    { bytes: [0x4d, 0x4d, 0x00, 0x2a], mime: "image/tiff" }, // TIFF (big-endian)
+    { bytes: [0x00, 0x00, 0x01, 0x00], mime: "image/x-icon" }, // ICO
+    { bytes: [0x38, 0x50, 0x53, 0x00], mime: "image/vnd.adobe.photoshop" }, // PSD
+];
+/**
+ * Converts a Uint8Array to a base64 string safely, without stack overflow.
+ *
+ * The naive approach `btoa(String.fromCharCode(...data))` fails for arrays
+ * larger than ~64K elements due to JavaScript's maximum call stack size.
+ * This function processes data in 8KB chunks to avoid the issue.
+ *
+ * @param data - The byte array to encode
+ * @returns A base64-encoded string
+ */
+function uint8ArrayToBase64(data) {
+    let binary = "";
+    const chunkSize = 8192;
+    for (let i = 0; i < data.length; i += chunkSize) {
+        const chunk = data.subarray(i, i + chunkSize);
+        binary += String.fromCharCode(...chunk);
+    }
+    return btoa(binary);
+}
+/**
+ * Detects the MIME type of an image from its binary data using magic byte signatures.
+ * Falls back to the shared `detectMimeTypeFromData` utility.
+ *
+ * @param data - The raw image bytes
+ * @returns The detected MIME type, or undefined if unrecognized
+ */
+function detectImageMimeType(data) {
+    if (data.length < 2)
+        return undefined;
+    for (const sig of IMAGE_SIGNATURES) {
+        if (data.length >= sig.bytes.length) {
+            const match = sig.bytes.every((byte, idx) => data[idx] === byte);
+            if (match)
+                return sig.mime;
+        }
+    }
+    return detectMimeTypeFromData(data);
+}
+/**
+ * Attempts to extract image dimensions from the raw bytes.
+ * Supports PNG (IHDR chunk) and JPEG (SOF marker) formats.
+ *
+ * @param data - The raw image bytes
+ * @param mimeType - The detected MIME type
+ * @returns An object with width and height, or undefined if unable to parse
+ */
+function extractImageDimensions(data, mimeType) {
+    try {
+        if (mimeType === "image/png" && data.length >= 24) {
+            // PNG: width at offset 16 (4 bytes BE), height at offset 20 (4 bytes BE)
+            const width = (data[16] << 24) | (data[17] << 16) | (data[18] << 8) | data[19];
+            const height = (data[20] << 24) | (data[21] << 16) | (data[22] << 8) | data[23];
+            if (width > 0 && height > 0)
+                return { width, height };
+        }
+        if (mimeType === "image/jpeg" && data.length >= 4) {
+            // JPEG: scan for SOF0 (0xC0) or SOF2 (0xC2) marker
+            let offset = 2;
+            while (offset < data.length - 9) {
+                if (data[offset] !== 0xff)
+                    break;
+                const marker = data[offset + 1];
+                if (marker === 0xc0 || marker === 0xc2) {
+                    const height = (data[offset + 5] << 8) | data[offset + 6];
+                    const width = (data[offset + 7] << 8) | data[offset + 8];
+                    if (width > 0 && height > 0)
+                        return { width, height };
+                }
+                // Skip to next marker
+                const segmentLength = (data[offset + 2] << 8) | data[offset + 3];
+                offset += 2 + segmentLength;
+            }
+        }
+        if (mimeType === "image/gif" && data.length >= 10) {
+            // GIF: width at offset 6 (2 bytes LE), height at offset 8 (2 bytes LE)
+            const width = data[6] | (data[7] << 8);
+            const height = data[8] | (data[9] << 8);
+            if (width > 0 && height > 0)
+                return { width, height };
+        }
+        if (mimeType === "image/bmp" && data.length >= 26) {
+            // BMP: width at offset 18 (4 bytes LE), height at offset 22 (4 bytes LE)
+            const width = data[18] | (data[19] << 8) | (data[20] << 16) | (data[21] << 24);
+            const height = data[22] | (data[23] << 8) | (data[24] << 16) | (data[25] << 24);
+            if (width > 0 && height > 0)
+                return { width, height: Math.abs(height) };
+        }
+    }
+    catch {
+        // Dimension extraction is best-effort; don't fail conversion
+    }
+    return undefined;
+}
+/**
+ * Renders an array of AST nodes into a Markdown string.
+ * This is the local renderMarkdown helper for the image-ocr converter.
+ *
+ * @param nodes - The AST nodes to render
+ * @returns The rendered Markdown string
+ */
+function renderMarkdown(nodes) {
+    const parts = [];
+    for (const node of nodes) {
+        parts.push(renderNode(node));
+    }
+    return parts.join("\n\n").replace(/\n{3,}/g, "\n\n").trim();
+}
+/**
+ * Renders a single AST node into its Markdown representation.
+ *
+ * @param node - The AST node to render
+ * @returns The Markdown string for this node
+ */
+function renderNode(node) {
+    switch (node.type) {
+        case "heading": {
+            const n = node;
+            return `${"#".repeat(n.level)} ${renderNodes(n.children)}`;
+        }
+        case "paragraph": {
+            return renderNodes(node.children);
+        }
+        case "image": {
+            const n = node;
+            return `![${n.alt ?? ""}](${n.src}${n.title ? ` "${n.title}"` : ""})`;
+        }
+        case "code": {
+            const n = node;
+            return "```" + (n.language ?? "") + "\n" + n.value + "\n```";
+        }
+        case "text": {
+            return node.value;
+        }
+        default:
+            if ("children" in node && Array.isArray(node.children)) {
+                return renderNodes(node.children);
+            }
+            return "";
+    }
+}
+/**
+ * Recursively renders an array of AST nodes into a single string.
+ *
+ * @param nodes - The AST nodes to render
+ * @returns The concatenated rendered output
+ */
+function renderNodes(nodes) {
+    return nodes.map((n) => renderNode(n)).join("");
+}
+/**
+ * Converts images to text using OCR (Optical Character Recognition).
+ *
+ * Features:
+ * - Chunked base64 encoding to prevent stack overflow on large images
+ * - Language selection (defaults to English, configurable)
+ * - Image dimension extraction from PNG, JPEG, GIF, and BMP headers
+ * - OCR confidence score from tesseract.js (when available)
+ * - Magic byte detection for additional image formats
+ * - Comprehensive error messages when tesseract.js is not installed
+ *
+ * @example
+ * ```ts
+ * const converter = new ImageOcrConverter();
+ * const result = await converter.convert({
+ *   data: pngBuffer,
+ *   options: { customProperties: { ocrLanguage: "fra" } }
+ * });
+ * console.log(result.ast);  // DocumentNode with image + extracted text
+ * console.log(result.markdown);
+ * ```
+ */
+export class ImageOcrConverter {
+    id = "image-ocr";
+    supportedMimeTypes = [
+        "image/png",
+        "image/jpeg",
+        "image/webp",
+        "image/tiff",
+        "image/bmp",
+        "image/gif",
+    ];
+    supportedExtensions = [
+        ".png",
+        ".jpg",
+        ".jpeg",
+        ".webp",
+        ".tiff",
+        ".tif",
+        ".bmp",
+        ".gif",
+    ];
+    /**
+     * Checks whether this converter can handle the given input.
+     * Validates MIME type, file extension, and magic byte signatures.
+     *
+     * @param input - The conversion input to check
+     * @returns True if this converter can process the input
+     */
+    async canConvert(input) {
+        if (input.mimeType && this.supportedMimeTypes.includes(input.mimeType)) {
+            return true;
+        }
+        if (input.fileName) {
+            const ext = input.fileName
+                .slice(input.fileName.lastIndexOf("."))
+                .toLowerCase();
+            if (this.supportedExtensions.includes(ext)) {
+                return true;
+            }
+        }
+        try {
+            const data = await readInputData(input.data);
+            const detected = detectImageMimeType(data);
+            if (detected && this.supportedMimeTypes.includes(detected)) {
+                return true;
+            }
+        }
+        catch {
+            return false;
+        }
+        return false;
+    }
+    /**
+     * Converts an image file into text via OCR, producing an AST with the image
+     * node and the extracted text as a paragraph.
+     *
+     * @param input - The image data as Uint8Array, ArrayBuffer, Blob, or string
+     * @returns A ConversionResult with the AST, Markdown, and image metadata
+     */
+    async convert(input) {
+        const startTime = performance.now();
+        const data = await readInputData(input.data);
+        const mimeType = input.mimeType ?? detectImageMimeType(data) ?? "image/png";
+        const imageId = generateId();
+        // Extract image dimensions (best-effort, from raw bytes)
+        const dimensions = extractImageDimensions(data, mimeType);
+        // Determine OCR language from options (default: English)
+        const ocrLanguage = input.options?.customProperties?.ocrLanguage ?? "eng";
+        // Perform OCR
+        let ocrText = "";
+        let ocrConfidence;
+        let ocrSuccess = false;
+        try {
+            const Tesseract = await import("tesseract.js");
+            const blob = new Blob([new Uint8Array(data)], { type: mimeType });
+            const result = await Tesseract.recognize(blob, ocrLanguage, {
+                logger: undefined, // suppress verbose logging
+            });
+            ocrText = result.data.text.trim();
+            ocrConfidence = result.data.confidence;
+            ocrSuccess = ocrText.length > 0;
+        }
+        catch (err) {
+            const errorMessage = err instanceof Error ? err.message : String(err);
+            if (errorMessage.includes("Cannot find module") ||
+                errorMessage.includes("MODULE_NOT_FOUND") ||
+                errorMessage.includes("tesseract")) {
+                ocrText =
+                    "[OCR unavailable: tesseract.js is not installed. " +
+                        "Install it with: npm install tesseract.js]";
+            }
+            else {
+                ocrText = `[OCR failed: ${errorMessage}]`;
+            }
+        }
+        // Build the base64 data URL (using safe chunked encoding)
+        const base64 = uint8ArrayToBase64(new Uint8Array(data));
+        const dataUrl = `data:${mimeType};base64,${base64}`;
+        // Build the AST nodes
+        const children = [];
+        // Image node with dimension metadata
+        const imageNode = createNode({
+            type: "image",
+            src: dataUrl,
+            alt: "OCR source image",
+            data: new Uint8Array(data),
+            mimeType,
+            width: dimensions?.width,
+            height: dimensions?.height,
+        });
+        children.push(imageNode);
+        // OCR result paragraph with confidence metadata
+        if (ocrText) {
+            const textNode = createNode({
+                type: "text",
+                value: ocrText,
+            });
+            const paragraphNode = createNode({
+                type: "paragraph",
+                children: [textNode],
+                // Attach confidence as custom property
+                confidence: ocrConfidence,
+                ocrSuccess,
+                ocrLanguage,
+            });
+            children.push(paragraphNode);
+        }
+        const documentNode = createNode({
+            type: "document",
+            children,
+        });
+        const markdown = renderMarkdown(children);
+        const endTime = performance.now();
+        const stats = {
+            startTime,
+            endTime,
+            duration: endTime - startTime,
+            inputSize: data.length,
+            outputSize: new TextEncoder().encode(markdown).byteLength,
+        };
+        const metadata = {
+            wordCount: ocrText.split(/\s+/).filter(Boolean).length,
+            format: mimeType,
+            customProperties: {
+                imageId,
+                width: dimensions?.width,
+                height: dimensions?.height,
+                ocrLanguage,
+                ocrConfidence,
+                ocrSuccess,
+            },
+        };
+        return {
+            markdown,
+            metadata,
+            assets: [],
+            tables: [],
+            images: [
+                {
+                    id: imageId,
+                    data: new Uint8Array(data),
+                    mimeType,
+                    alt: "OCR source image",
+                    width: dimensions?.width,
+                    height: dimensions?.height,
+                },
+            ],
+            headings: [],
+            ast: documentNode,
+            format: "markdown",
+            converterId: this.id,
+            stats,
+        };
+    }
+}
+//# sourceMappingURL=ocr-converter.js.map

package/dist/ocr-converter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ocr-converter.js","sourceRoot":"","sources":["../src/ocr-converter.ts"],"names":[],"mappings":"AAOA,OAAO,EACL,UAAU,EACV,aAAa,EACb,sBAAsB,EACtB,UAAU,GAQX,MAAM,sBAAsB,CAAC;AAE9B,mEAAmE;AACnE,MAAM,gBAAgB,GAAwC;IAC5D,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,EAAS,MAAM;IACrE,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,EAAe,OAAO;IACvE,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,EAAQ,cAAc;IAC7E,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,EAAS,OAAO;IACtE,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,WAAW,EAAE,EAAsB,MAAM;IACtE,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,EAAQ,uBAAuB;IACtF,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,EAAQ,oBAAoB;IACnF,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,cAAc,EAAE,EAAM,MAAM;IACrE,EAAE,KAAK,EAAE,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,EAAE,IAAI,EAAE,2BAA2B,EAAE,EAAE,MAAM;CAC/E,CAAC;AAEF;;;;;;;;;GASG;AACH,SAAS,kBAAkB,CAAC,IAAgB;IAC1C,IAAI,MAAM,GAAG,EAAE,CAAC;IAChB,MAAM,SAAS,GAAG,IAAI,CAAC;IACvB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC,IAAI,SAAS,EAAE,CAAC;QAChD,MAAM,KAAK,GAAG,IAAI,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,GAAG,SAAS,CAAC,CAAC;QAC9C,MAAM,IAAI,MAAM,CAAC,YAAY,CAAC,GAAG,KAAK,CAAC,CAAC;IAC1C,CAAC;IACD,OAAO,IAAI,CAAC,MAAM,CAAC,CAAC;AACtB,CAAC;AAED;;;;;;GAMG;AACH,SAAS,mBAAmB,CAAC,IAAgB;IAC3C,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,SAAS,CAAC;IAEtC,KAAK,MAAM,GAAG,IAAI,gBAAgB,EAAE,CAAC;QACnC,IAAI,IAAI,CAAC,MAAM,IAAI,GAAG,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC;YACpC,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,IAAI,CAAC,CAAC;YACjE,IAAI,KAAK;gBAAE,OAAO,GAAG,CAAC,IAAI,CAAC;QAC7B,CAAC;IACH,CAAC;IAED,OAAO,sBAAsB,CAAC,IAAI,CAAC,CAAC;AACtC,CAAC;AAED;;;;;;;GAOG;AACH,SAAS,sBAAsB,CAC7B,IAAgB,EAChB,QAAgB;IAEhB,IAAI,CAAC;QACH,IAAI,QAAQ,KAAK,WAAW,IAAI,IAAI,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;YAClD,yEAAyE;YACzE,MAAM,KAAK,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,EAAE,CAAE,CAAC;YACnF,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,EAAE,CAAE,CAAC;YACpF,IAAI,KAAK,GAAG,CAAC,IAAI,MAAM,GAAG,CAAC;gBAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;QACxD,CAAC;QAED,IAAI,QAAQ,KAAK,YAAY,IAAI,IAAI,CAAC,MAAM,IAAI,CAAC,EAAE,CAAC;YAClD,mDAAmD;YACnD,IAAI,MAAM,GAAG,CAAC,CAAC;YACf,OAAO,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;gBAChC,IAAI,IAAI,CAAC,MAAM,CAAC,KAAK,IAAI;oBAAE,MAAM;gBACjC,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;gBACjC,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;oBACvC,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;oBAC5D,MAAM,KAAK,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;oBAC3D,IAAI,KAAK,GAAG,CAAC,IAAI,MAAM,GAAG,CAAC;wBAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;gBACxD,CAAC;gBACD,sBAAsB;gBACtB,MAAM,aAAa,GAAG,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,IAAI,CAAC,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,CAAC,CAAE,CAAC;gBACnE,MAAM,IAAI,CAAC,GAAG,aAAa,CAAC;YAC9B,CAAC;QACH,CAAC;QAED,IAAI,QAAQ,KAAK,WAAW,IAAI,IAAI,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;YAClD,uEAAuE;YACvE,MAAM,KAAK,GAAG,IAAI,CAAC,CAAC,CAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAE,IAAI,CAAC,CAAC,CAAC;YACzC,MAAM,MAAM,GAAG,IAAI,CAAC,CAAC,CAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAE,IAAI,CAAC,CAAC,CAAC;YAC1C,IAAI,KAAK,GAAG,CAAC,IAAI,MAAM,GAAG,CAAC;gBAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,CAAC;QACxD,CAAC;QAED,IAAI,QAAQ,KAAK,WAAW,IAAI,IAAI,CAAC,MAAM,IAAI,EAAE,EAAE,CAAC;YAClD,yEAAyE;YACzE,MAAM,KAAK,GAAG,IAAI,CAAC,EAAE,CAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,CAAC;YACnF,MAAM,MAAM,GAAG,IAAI,CAAC,EAAE,CAAE,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,CAAC,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE,CAAE,IAAI,EAAE,CAAC,CAAC;YACpF,IAAI,KAAK,GAAG,CAAC,IAAI,MAAM,GAAG,CAAC;gBAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC;QAC1E,CAAC;IACH,CAAC;IAAC,MAAM,CAAC;QACP,6DAA6D;IAC/D,CAAC;IAED,OAAO,SAAS,CAAC;AACnB,CAAC;AAED;;;;;;GAMG;AACH,SAAS,cAAc,CAAC,KAAgB;IACtC,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;IAC/B,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;AAC9D,CAAC;AAED;;;;;GAKG;AACH,SAAS,UAAU,CAAC,IAAa;IAC/B,QAAQ,IAAI,CAAC,IAAI,EAAE,CAAC;QAClB,KAAK,SAAS,CAAC,CAAC,CAAC;YACf,MAAM,CAAC,GAAG,IAAmB,CAAC;YAC9B,OAAO,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,CAAC,CAAC,QAAQ,CAAC,EAAE,CAAC;QAC7D,CAAC;QACD,KAAK,WAAW,CAAC,CAAC,CAAC;YACjB,OAAO,WAAW,CAAE,IAAsB,CAAC,QAAQ,CAAC,CAAC;QACvD,CAAC;QACD,KAAK,OAAO,CAAC,CAAC,CAAC;YACb,MAAM,CAAC,GAAG,IAAiB,CAAC;YAC5B,OAAO,KAAK,CAAC,CAAC,GAAG,IAAI,EAAE,KAAK,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC;QACxE,CAAC;QACD,KAAK,MAAM,CAAC,CAAC,CAAC;YACZ,MAAM,CAAC,GAAG,IAAgB,CAAC;YAC3B,OAAO,KAAK,GAAG,CAAC,CAAC,CAAC,QAAQ,IAAI,EAAE,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,KAAK,GAAG,OAAO,CAAC;QAC/D,CAAC;QACD,KAAK,MAAM,CAAC,CAAC,CAAC;YACZ,OAAQ,IAAiB,CAAC,KAAK,CAAC;QAClC,CAAC;QACD;YACE,IAAI,UAAU,IAAI,IAAI,IAAI,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,EAAE,CAAC;gBACvD,OAAO,WAAW,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;YACpC,CAAC;YACD,OAAO,EAAE,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;;;GAKG;AACH,SAAS,WAAW,CAAC,KAAgB;IACnC,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAClD,CAAC;AAED;;;;;;;;;;;;;;;;;;;;;GAqBG;AACH,MAAM,OAAO,iBAAiB;IACnB,EAAE,GAAG,WAAW,CAAC;IACjB,kBAAkB,GAAG;QAC5B,WAAW;QACX,YAAY;QACZ,YAAY;QACZ,YAAY;QACZ,WAAW;QACX,WAAW;KACZ,CAAC;IACO,mBAAmB,GAAG;QAC7B,MAAM;QACN,MAAM;QACN,OAAO;QACP,OAAO;QACP,OAAO;QACP,MAAM;QACN,MAAM;QACN,MAAM;KACP,CAAC;IAEF;;;;;;OAMG;IACH,KAAK,CAAC,UAAU,CAAC,KAAsB;QACrC,IAAI,KAAK,CAAC,QAAQ,IAAI,IAAI,CAAC,kBAAkB,CAAC,QAAQ,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,CAAC;YACvE,OAAO,IAAI,CAAC;QACd,CAAC;QAED,IAAI,KAAK,CAAC,QAAQ,EAAE,CAAC;YACnB,MAAM,GAAG,GAAG,KAAK,CAAC,QAAQ;iBACvB,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;iBACtC,WAAW,EAAE,CAAC;YACjB,IAAI,IAAI,CAAC,mBAAmB,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC3C,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAED,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC7C,MAAM,QAAQ,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC;YAC3C,IAAI,QAAQ,IAAI,IAAI,CAAC,kBAAkB,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,CAAC;gBAC3D,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,KAAK,CAAC;QACf,CAAC;QAED,OAAO,KAAK,CAAC;IACf,CAAC;IAED;;;;;;OAMG;IACH,KAAK,CAAC,OAAO,CAAC,KAAsB;QAClC,MAAM,SAAS,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QACpC,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;QAC7C,MAAM,QAAQ,GACZ,KAAK,CAAC,QAAQ,IAAI,mBAAmB,CAAC,IAAI,CAAC,IAAI,WAAW,CAAC;QAC7D,MAAM,OAAO,GAAG,UAAU,EAAE,CAAC;QAE7B,yDAAyD;QACzD,MAAM,UAAU,GAAG,sBAAsB,CAAC,IAAI,EAAE,QAAQ,CAAC,CAAC;QAE1D,yDAAyD;QACzD,MAAM,WAAW,GACf,KAAK,CAAC,OAAO,EAAE,gBAAgB,EAAE,WAAiC,IAAI,KAAK,CAAC;QAE9E,cAAc;QACd,IAAI,OAAO,GAAG,EAAE,CAAC;QACjB,IAAI,aAAiC,CAAC;QACtC,IAAI,UAAU,GAAG,KAAK,CAAC;QAEvB,IAAI,CAAC;YACH,MAAM,SAAS,GAAG,MAAM,MAAM,CAAC,cAAc,CAAC,CAAC;YAC/C,MAAM,IAAI,GAAG,IAAI,IAAI,CAAC,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;YAClE,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,SAAS,CAAC,IAAI,EAAE,WAAW,EAAE;gBAC1D,MAAM,EAAE,SAAS,EAAE,2BAA2B;aAC/C,CAAC,CAAC;YAEH,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;YAClC,aAAa,GAAG,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC;YACvC,UAAU,GAAG,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC;QAClC,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,YAAY,GAChB,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAEnD,IACE,YAAY,CAAC,QAAQ,CAAC,oBAAoB,CAAC;gBAC3C,YAAY,CAAC,QAAQ,CAAC,kBAAkB,CAAC;gBACzC,YAAY,CAAC,QAAQ,CAAC,WAAW,CAAC,EAClC,CAAC;gBACD,OAAO;oBACL,mDAAmD;wBACnD,4CAA4C,CAAC;YACjD,CAAC;iBAAM,CAAC;gBACN,OAAO,GAAG,gBAAgB,YAAY,GAAG,CAAC;YAC5C,CAAC;QACH,CAAC;QAED,0DAA0D;QAC1D,MAAM,MAAM,GAAG,kBAAkB,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;QACxD,MAAM,OAAO,GAAG,QAAQ,QAAQ,WAAW,MAAM,EAAE,CAAC;QAEpD,sBAAsB;QACtB,MAAM,QAAQ,GAAc,EAAE,CAAC;QAE/B,qCAAqC;QACrC,MAAM,SAAS,GAAG,UAAU,CAAY;YACtC,IAAI,EAAE,OAAO;YACb,GAAG,EAAE,OAAO;YACZ,GAAG,EAAE,kBAAkB;YACvB,IAAI,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC;YAC1B,QAAQ;YACR,KAAK,EAAE,UAAU,EAAE,KAAK;YACxB,MAAM,EAAE,UAAU,EAAE,MAAM;SAC3B,CAAC,CAAC;QACH,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QAEzB,gDAAgD;QAChD,IAAI,OAAO,EAAE,CAAC;YACZ,MAAM,QAAQ,GAAG,UAAU,CAAW;gBACpC,IAAI,EAAE,MAAM;gBACZ,KAAK,EAAE,OAAO;aACf,CAAC,CAAC;YAEH,MAAM,aAAa,GAAG,UAAU,CAAgB;gBAC9C,IAAI,EAAE,WAAW;gBACjB,QAAQ,EAAE,CAAC,QAAQ,CAAC;gBACpB,uCAAuC;gBACvC,UAAU,EAAE,aAAa;gBACzB,UAAU;gBACV,WAAW;aACZ,CAAC,CAAC;YACH,QAAQ,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC;QAC/B,CAAC;QAED,MAAM,YAAY,GAAG,UAAU,CAAe;YAC5C,IAAI,EAAE,UAAU;YAChB,QAAQ;SACT,CAAC,CAAC;QAEH,MAAM,QAAQ,GAAG,cAAc,CAAC,QAAQ,CAAC,CAAC;QAE1C,MAAM,OAAO,GAAG,WAAW,CAAC,GAAG,EAAE,CAAC;QAClC,MAAM,KAAK,GAAoB;YAC7B,SAAS;YACT,OAAO;YACP,QAAQ,EAAE,OAAO,GAAG,SAAS;YAC7B,SAAS,EAAE,IAAI,CAAC,MAAM;YACtB,UAAU,EAAE,IAAI,WAAW,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,UAAU;SAC1D,CAAC;QAEF,MAAM,QAAQ,GAAqB;YACjC,SAAS,EAAE,OAAO,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,MAAM;YACtD,MAAM,EAAE,QAAQ;YAChB,gBAAgB,EAAE;gBAChB,OAAO;gBACP,KAAK,EAAE,UAAU,EAAE,KAAK;gBACxB,MAAM,EAAE,UAAU,EAAE,MAAM;gBAC1B,WAAW;gBACX,aAAa;gBACb,UAAU;aACX;SACF,CAAC;QAEF,OAAO;YACL,QAAQ;YACR,QAAQ;YACR,MAAM,EAAE,EAAE;YACV,MAAM,EAAE,EAAE;YACV,MAAM,EAAE;gBACN;oBACE,EAAE,EAAE,OAAO;oBACX,IAAI,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC;oBAC1B,QAAQ;oBACR,GAAG,EAAE,kBAAkB;oBACvB,KAAK,EAAE,UAAU,EAAE,KAAK;oBACxB,MAAM,EAAE,UAAU,EAAE,MAAM;iBAC3B;aACF;YACD,QAAQ,EAAE,EAAE;YACZ,GAAG,EAAE,YAAY;YACjB,MAAM,EAAE,UAAU;YAClB,WAAW,EAAE,IAAI,CAAC,EAAE;YACpB,KAAK;SACN,CAAC;IACJ,CAAC;CACF"}

package/package.json

@@ -0,0 +1,40 @@
+{
+  "name": "@markitdownjs/image-ocr",
+  "version": "0.1.0",
+  "description": "Image OCR converter for MarkItDownJS",
+  "license": "MIT",
+  "type": "module",
+  "main": "./dist/index.js",
+  "module": "./dist/index.js",
+  "types": "./dist/index.d.ts",
+  "exports": {
+    ".": {
+      "import": "./dist/index.js",
+      "types": "./dist/index.d.ts"
+    }
+  },
+  "files": [
+    "dist"
+  ],
+  "dependencies": {
+    "@markitdownjs/shared": "0.1.0"
+  },
+  "peerDependencies": {
+    "tesseract.js": ">=5.0.0"
+  },
+  "peerDependenciesMeta": {
+    "tesseract.js": {
+      "optional": true
+    }
+  },
+  "devDependencies": {
+    "tesseract.js": "^5.0.0",
+    "typescript": "^5.5.0"
+  },
+  "scripts": {
+    "build": "tsc --project tsconfig.json",
+    "dev": "tsc --watch --project tsconfig.json",
+    "typecheck": "tsc --noEmit",
+    "clean": "rm -rf dist tsconfig.tsbuildinfo"
+  }
+}