deckhearth/lib/ocr-worker.js

171 lines
4.9 KiB
JavaScript
Raw Permalink Normal View History

/**
* Browser-side Tesseract OCR for Layer-1 card name strip recognition.
* Uses tesseract.js workers (off main thread). Language data loads from CDN on first use.
*/
let workerPromise = null;
const NAME_STRIP_PSM = '7';
const NUMBER_STRIP_PSM = '7';
const NAME_STRIP_HEIGHT_RATIO = 0.35;
const NUMBER_STRIP_HEIGHT_RATIO = 0.18;
const OCR_JPEG_QUALITY = 0.92;
async function getWorker() {
if (!workerPromise) {
workerPromise = (async () => {
const { createWorker } = await import('tesseract.js');
const worker = await createWorker('eng', 1, {
logger: () => {},
});
return worker;
})();
}
return workerPromise;
}
function cropCardStrip(imageDataUrl, { fromTop = true, heightRatio }) {
return new Promise((resolve, reject) => {
const img = new Image();
img.onload = () => {
const stripHeight = Math.max(1, Math.floor(img.height * heightRatio));
const canvas = document.createElement('canvas');
canvas.width = img.width;
canvas.height = stripHeight;
const ctx = canvas.getContext('2d');
const sourceY = fromTop ? 0 : Math.max(0, img.height - stripHeight);
ctx.drawImage(
img,
0,
sourceY,
img.width,
stripHeight,
0,
0,
canvas.width,
canvas.height
);
resolve(canvas.toDataURL('image/jpeg', OCR_JPEG_QUALITY));
};
img.onerror = () => reject(new Error('Failed to load image for OCR'));
img.src = imageDataUrl;
});
}
/**
* Pick the most likely collector number token from OCR of the bottom strip.
* @param {string} rawText
* @returns {string}
*/
export function extractCollectorNumberCandidate(rawText) {
if (!rawText || typeof rawText !== 'string') return '';
const lines = rawText
.split(/\r?\n/)
.map((line) => line.replace(/\s+/g, ' ').trim())
.filter(Boolean);
const joined = lines.join(' ');
const slashMatch = joined.match(/\b(\d{1,4}\s*\/\s*\d{1,4}[A-Za-z]?)\b/);
if (slashMatch) {
return slashMatch[1].replace(/\s+/g, '');
}
for (let i = lines.length - 1; i >= 0; i -= 1) {
const line = lines[i];
if (/\d/.test(line) && line.length <= 12) {
return line;
}
}
return '';
}
/**
* OCR the top name strip of a card crop.
* @param {string} imageDataUrl - full card crop data URL
* @returns {Promise<{ text: string, confidence: number }>}
*/
export async function recognizeCardNameStrip(imageDataUrl) {
const fields = await recognizeCardFields(imageDataUrl);
return {
text: fields.nameText,
confidence: fields.nameConfidence,
};
}
/**
* OCR name strip + bottom collector-number strip.
* Runs both strips in parallel using two dedicated workers (~50% faster on mobile).
* @param {string} imageDataUrl
* @returns {Promise<{ nameText: string, nameConfidence: number, cardNumber: string, numberConfidence: number }>}
*/
let numberWorkerPromise = null;
async function getNumberWorker() {
if (!numberWorkerPromise) {
numberWorkerPromise = (async () => {
const { createWorker } = await import('tesseract.js');
const worker = await createWorker('eng', 1, {
logger: () => {},
});
await worker.setParameters({ tessedit_pageseg_mode: NUMBER_STRIP_PSM });
return worker;
})();
}
return numberWorkerPromise;
}
export async function recognizeCardFields(imageDataUrl) {
if (typeof window === 'undefined') {
return { nameText: '', nameConfidence: 0, cardNumber: '', numberConfidence: 0 };
}
try {
const [nameStripUrl, numberStripUrl] = await Promise.all([
cropCardStrip(imageDataUrl, { fromTop: true, heightRatio: NAME_STRIP_HEIGHT_RATIO }),
cropCardStrip(imageDataUrl, { fromTop: false, heightRatio: NUMBER_STRIP_HEIGHT_RATIO }),
]);
const [nameWorker, numberWorker] = await Promise.all([
getWorker(),
getNumberWorker(),
]);
await nameWorker.setParameters({ tessedit_pageseg_mode: NAME_STRIP_PSM });
const [nameResult, numberResult] = await Promise.all([
nameWorker.recognize(nameStripUrl),
numberWorker.recognize(numberStripUrl),
]);
const cardNumber = extractCollectorNumberCandidate(numberResult.data.text || '');
return {
nameText: (nameResult.data.text || '').trim(),
nameConfidence: Math.round(nameResult.data.confidence || 0),
cardNumber,
numberConfidence: Math.round(numberResult.data.confidence || 0),
};
} catch (error) {
console.warn('[ocr-worker] recognizeCardFields failed:', error);
return { nameText: '', nameConfidence: 0, cardNumber: '', numberConfidence: 0 };
}
}
export async function terminateOcrWorker() {
const workers = [workerPromise, numberWorkerPromise];
workerPromise = null;
numberWorkerPromise = null;
await Promise.all(
workers.filter(Boolean).map(async (promise) => {
try {
const worker = await promise;
await worker.terminate();
} catch {
// ignore teardown errors
}
})
);
}