feat(paddleocr): add PaddleOCR OCR service (Docker images, server, tests, docs) and CI workflows

This commit is contained in:
2026-01-16 13:23:01 +00:00
parent 67c38eeb67
commit bec379e9ca
10 changed files with 624 additions and 71 deletions

View File

@@ -22,16 +22,11 @@ interface IInvoice {
* Extract OCR text from an image using PaddleOCR
*/
async function extractOcrText(imageBase64: string): Promise<string> {
const formData = new FormData();
const imageBuffer = Buffer.from(imageBase64, 'base64');
const blob = new Blob([imageBuffer], { type: 'image/png' });
formData.append('img', blob, 'image.png');
formData.append('outtype', 'json');
try {
const response = await fetch(`${PADDLEOCR_URL}/ocr`, {
method: 'POST',
body: formData,
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ image: imageBase64 }),
});
if (!response.ok) return '';
@@ -180,29 +175,64 @@ function hashInvoice(invoice: IInvoice): string {
/**
* Extract with majority voting - run until 2 passes match
* Optimization: Run Pass 1, OCR, and Pass 2 (after OCR) in parallel
*/
async function extractWithConsensus(images: string[], invoiceName: string, maxPasses: number = 5): Promise<IInvoice> {
const results: Array<{ invoice: IInvoice; hash: string }> = [];
const hashCounts: Map<string, number> = new Map();
// Extract OCR text from first page
const ocrText = await extractOcrText(images[0]);
if (ocrText) {
console.log(` [OCR] Extracted ${ocrText.split('\n').length} text lines`);
const addResult = (invoice: IInvoice, passLabel: string): number => {
const hash = hashInvoice(invoice);
results.push({ invoice, hash });
hashCounts.set(hash, (hashCounts.get(hash) || 0) + 1);
console.log(` [${passLabel}] ${invoice.invoice_number} | ${invoice.invoice_date} | ${invoice.total_amount} ${invoice.currency}`);
return hashCounts.get(hash)!;
};
// OPTIMIZATION: Run Pass 1 (no OCR) in parallel with OCR -> Pass 2 (with OCR)
let ocrText = '';
const pass1Promise = extractOnce(images, 1, '').catch((err) => ({ error: err }));
// OCR then immediately Pass 2
const ocrThenPass2Promise = (async () => {
ocrText = await extractOcrText(images[0]);
if (ocrText) {
console.log(` [OCR] Extracted ${ocrText.split('\n').length} text lines`);
}
return extractOnce(images, 2, ocrText).catch((err) => ({ error: err }));
})();
// Wait for both to complete
const [pass1Result, pass2Result] = await Promise.all([pass1Promise, ocrThenPass2Promise]);
// Process Pass 1 result
if ('error' in pass1Result) {
console.log(` [Pass 1] Error: ${(pass1Result as {error: unknown}).error}`);
} else {
const count = addResult(pass1Result as IInvoice, 'Pass 1');
if (count >= 2) {
console.log(` [Consensus] Reached after parallel passes`);
return pass1Result as IInvoice;
}
}
for (let pass = 1; pass <= maxPasses; pass++) {
// Process Pass 2 result
if ('error' in pass2Result) {
console.log(` [Pass 2+OCR] Error: ${(pass2Result as {error: unknown}).error}`);
} else {
const count = addResult(pass2Result as IInvoice, 'Pass 2+OCR');
if (count >= 2) {
console.log(` [Consensus] Reached after parallel passes`);
return pass2Result as IInvoice;
}
}
// Continue with passes 3+ using OCR text if no consensus yet
for (let pass = 3; pass <= maxPasses; pass++) {
try {
const invoice = await extractOnce(images, pass, ocrText);
const hash = hashInvoice(invoice);
const count = addResult(invoice, `Pass ${pass}+OCR`);
results.push({ invoice, hash });
hashCounts.set(hash, (hashCounts.get(hash) || 0) + 1);
console.log(` [Pass ${pass}] ${invoice.invoice_number} | ${invoice.invoice_date} | ${invoice.total_amount} ${invoice.currency}`);
// Check if we have consensus (2+ matching)
const count = hashCounts.get(hash)!;
if (count >= 2) {
console.log(` [Consensus] Reached after ${pass} passes`);
return invoice;
@@ -267,6 +297,7 @@ function compareInvoice(
/**
* Find all test cases (PDF + JSON pairs) in .nogit/invoices/
* Priority invoices (like vodafone) run first for quick feedback
*/
function findTestCases(): Array<{ name: string; pdfPath: string; jsonPath: string }> {
const testDir = path.join(process.cwd(), '.nogit/invoices');
@@ -290,6 +321,22 @@ function findTestCases(): Array<{ name: string; pdfPath: string; jsonPath: strin
}
}
// Sort with priority invoices first, then alphabetically
const priorityPrefixes = ['vodafone'];
testCases.sort((a, b) => {
const aPriority = priorityPrefixes.findIndex((p) => a.name.startsWith(p));
const bPriority = priorityPrefixes.findIndex((p) => b.name.startsWith(p));
// Both have priority - sort by priority order
if (aPriority >= 0 && bPriority >= 0) return aPriority - bPriority;
// Only a has priority - a comes first
if (aPriority >= 0) return -1;
// Only b has priority - b comes first
if (bPriority >= 0) return 1;
// Neither has priority - alphabetical
return a.name.localeCompare(b.name);
});
return testCases;
}