R73-A2 收敛 VAT OCR 归一化与质量门槛

server/src/services/ocr/tencent.js

@@ -46,12 +46,41 @@ function toNumberString(value) {
   return num.toFixed(2);
 }
 
+function addNumberStrings(...values) {
+  const normalized = values.map((value) => toNumberString(value));
+  if (normalized.some((value) => value == null)) {
+    return null;
+  }
+
+  const total = normalized.reduce((sum, value) => sum + Number(value), 0);
+  return total.toFixed(2);
+}
+
+function subtractNumberStrings(minuend, subtrahend) {
+  const normalizedMinuend = toNumberString(minuend);
+  const normalizedSubtrahend = toNumberString(subtrahend);
+  if (normalizedMinuend == null || normalizedSubtrahend == null) {
+    return null;
+  }
+
+  return (Number(normalizedMinuend) - Number(normalizedSubtrahend)).toFixed(2);
+}
+
+function normalizeIdentifier(value) {
+  if (typeof value !== "string") {
+    return null;
+  }
+
+  const normalized = value.replace(/\s+/g, "").trim();
+  return normalized || null;
+}
+
 function normalizeString(value) {
   if (typeof value !== "string") {
     return null;
   }
 
-  const normalized = value.trim();
+  const normalized = value.replace(/\s+/g, " ").trim();
   return normalized || null;
 }
 
@@ -76,13 +105,19 @@ function createTencentOcrClient(config) {
 }
 
 function normalizeVatInvoiceResult(result) {
+  const amount = toNumberString(result?.AmountWithoutTax);
+  const taxAmount = toNumberString(result?.TaxAmount);
+  const totalAmountFromSource = toNumberString(result?.TotalAmount ?? result?.AmountInFiguers);
+  const derivedAmount = amount ?? subtractNumberStrings(totalAmountFromSource, taxAmount);
+  const derivedTotalAmount = totalAmountFromSource ?? addNumberStrings(derivedAmount, taxAmount);
+
   return {
-    invoiceCode: normalizeString(result?.InvoiceCode),
-    invoiceNumber: normalizeString(result?.InvoiceNum),
+    invoiceCode: normalizeIdentifier(result?.InvoiceCode),
+    invoiceNumber: normalizeIdentifier(result?.InvoiceNum),
     invoiceDate: normalizeDate(result?.InvoiceDate),
-    amount: toNumberString(result?.AmountWithoutTax),
-    taxAmount: toNumberString(result?.TaxAmount),
-    totalAmount: toNumberString(result?.TotalAmount ?? result?.AmountInFiguers),
+    amount: derivedAmount,
+    taxAmount,
+    totalAmount: derivedTotalAmount,
     sellerName: normalizeString(result?.SellerName),
     buyerName: normalizeString(result?.BuyerName),
     raw: result ?? {},
@@ -99,5 +134,6 @@ async function runTencentVatOcr(config, imageBase64) {
 }
 
 module.exports = {
+  normalizeVatInvoiceResult,
   runTencentVatOcr,
 };