R73-A2 收敛 VAT OCR 归一化与质量门槛

server/src/jobs/ocrJob.js

@@ -1,6 +1,6 @@
 const cron = require("node-cron");
 
-const { runOcrForInvoice } = require("../services/ocr");
+const { buildMissingRequiredOcrError, runOcrForInvoice } = require("../services/ocr");
 const { readFileBuffer } = require("../services/storage");
 
 const OCR_DUPLICATE_ERROR = "duplicate_invoice_code_number";
@@ -201,6 +201,30 @@ async function markInvoiceAsAvailable(prisma, invoiceId, normalized, ocrResultId
   });
 }
 
+async function markInvoiceAsQualityFailed(prisma, invoiceId, normalized, ocrResultId, errorMessage) {
+  const now = new Date();
+  await prisma.invoicePool.update({
+    where: { id: invoiceId },
+    data: {
+      status: "ocr_failed",
+      ocr_result_id: ocrResultId,
+      invoice_code: normalized.invoiceCode ?? null,
+      invoice_number: normalized.invoiceNumber ?? null,
+      invoice_date: normalized.invoiceDate ? new Date(toDateOnly(normalized.invoiceDate)) : null,
+      amount: toMoney(normalized.amount),
+      tax_amount: toMoney(normalized.taxAmount),
+      total_amount: toMoney(normalized.totalAmount),
+      seller_name: normalized.sellerName ?? null,
+      buyer_name: normalized.buyerName ?? null,
+      ocr_last_error: errorMessage,
+      ocr_completed_at: now,
+      ocr_attempts: { increment: 1 },
+      ocr_locked_at: null,
+      ocr_lock_owner: null,
+    },
+  });
+}
+
 async function markInvoiceAsFailed(prisma, claimedInvoice, errorMessage, config) {
   const now = new Date();
   const nextAttempts = Number(claimedInvoice.ocr_attempts || 0) + 1;
@@ -234,6 +258,26 @@ async function processClaimedInvoice(prisma, config, claimedInvoice, workerId, l
       imageBase64,
     );
 
+    const qualityError = buildMissingRequiredOcrError(claimedInvoice.ocr_type, normalized);
+    if (qualityError) {
+      await markInvoiceAsQualityFailed(
+        prisma,
+        claimedInvoice.id,
+        normalized,
+        ocrResultId,
+        qualityError,
+      );
+      logger.warn(
+        {
+          invoiceId: claimedInvoice.id,
+          ocrType: claimedInvoice.ocr_type,
+          qualityError,
+        },
+        "OCR invoice marked as failed for missing required fields",
+      );
+      return { outcome: "failed" };
+    }
+
     const duplicate = await findDuplicateInvoice(
       prisma,
       claimedInvoice.id,
@@ -363,6 +407,7 @@ function startOcrJob(fastify) {
 module.exports = {
   OCR_DUPLICATE_ERROR,
   claimPendingInvoices,
+  markInvoiceAsQualityFailed,
   runOcrOnce,
   startOcrJob,
 };

server/src/routes/invoicePool.js

@@ -9,7 +9,7 @@ const { compactDateString } = require("../utils/date");
 const { sha256Hex } = require("../utils/crypto");
 const { toMoneyString } = require("../utils/workflow");
 const { findActiveDelegationForDelegate } = require("../utils/delegation");
-const { isOcrEnabledForType } = require("../services/ocr");
+const { getMissingRequiredOcrFields, isOcrEnabledForType } = require("../services/ocr");
 const { registerInvoiceVerificationRoutes } = require("./invoicePool/verification");
 const { registerInvoicePoolReadRoutes } = require("./invoicePool/read");
 const { normalizeInvoiceResponse } = require("./invoicePool/shared/mappers");
@@ -46,7 +46,7 @@ const invoiceEditBodySchema = z
     invoiceNumber: z.string().trim().min(1).max(64).nullable().optional(),
     invoiceDate: z
       .string()
-      .regex(/^\\d{4}-\\d{2}-\\d{2}$/)
+      .regex(/^\d{4}-\d{2}-\d{2}$/)
       .nullable()
       .optional(),
     sellerName: z.string().trim().min(1).max(256).nullable().optional(),
@@ -163,7 +163,7 @@ function normalizeMoneyInput(value) {
     return null;
   }
 
-  if (!/^\\d+(?:\\.\\d{1,2})?$/.test(input)) {
+  if (!/^\d+(?:\.\d{1,2})?$/.test(input)) {
     return undefined;
   }
 
@@ -186,7 +186,7 @@ function parseDateOnlyString(value) {
     return null;
   }
 
-  if (!/^\\d{4}-\\d{2}-\\d{2}$/.test(trimmed)) {
+  if (!/^\d{4}-\d{2}-\d{2}$/.test(trimmed)) {
     return undefined;
   }
 
@@ -576,6 +576,32 @@ async function invoicePoolRoutes(fastify) {
 
         if (input.status) {
           if (input.status === "available" && current.status === "ocr_failed") {
+            const nextSnapshot = {
+              invoiceCode: "invoiceCode" in input ? input.invoiceCode : current.invoice_code,
+              invoiceNumber: "invoiceNumber" in input ? input.invoiceNumber : current.invoice_number,
+              invoiceDate:
+                "invoiceDate" in input
+                  ? input.invoiceDate
+                  : current.invoice_date?.toISOString?.()?.slice(0, 10) ?? null,
+              totalAmount:
+                "totalAmount" in input
+                  ? normalizeMoneyInput(input.totalAmount)
+                  : current.total_amount?.toString?.() ?? null,
+              sellerName: "sellerName" in input ? input.sellerName : current.seller_name,
+            };
+            const missingFields = getMissingRequiredOcrFields(current.ocr_type, nextSnapshot);
+            if (missingFields.length) {
+              throw new AppError({
+                code: "VALID_001",
+                message: "参数校验失败",
+                statusCode: 409,
+                details: {
+                  ocr: "missing_required_fields",
+                  missingFields,
+                },
+              });
+            }
+
             updateData.status = "available";
             updateData.ocr_last_error = null;
             updateData.ocr_completed_at = new Date();

server/src/services/ocr/index.js

@@ -1,6 +1,10 @@
 const { runTencentVatOcr } = require("./tencent");
 const { runMockOcr } = require("./mock");
 
+const REQUIRED_OCR_FIELDS_BY_TYPE = {
+  vat: ["invoiceCode", "invoiceNumber", "invoiceDate", "totalAmount", "sellerName"],
+};
+
 function isOcrEnabledForType(config, ocrType) {
   if (config.ocrProvider === "mock") {
     return ocrType === "vat";
@@ -9,6 +13,27 @@ function isOcrEnabledForType(config, ocrType) {
   return config.ocrProvider === "tencent" && ocrType === "vat";
 }
 
+function getMissingRequiredOcrFields(ocrType, normalized) {
+  const requiredFields = REQUIRED_OCR_FIELDS_BY_TYPE[ocrType] ?? [];
+  if (!requiredFields.length) {
+    return [];
+  }
+
+  return requiredFields.filter((field) => {
+    const value = normalized?.[field];
+    return value == null || value === "";
+  });
+}
+
+function buildMissingRequiredOcrError(ocrType, normalized) {
+  const missingFields = getMissingRequiredOcrFields(ocrType, normalized);
+  if (!missingFields.length) {
+    return null;
+  }
+
+  return `missing_required_fields:${missingFields.join(",")}`;
+}
+
 async function runOcrForInvoice({ config, ocrType, imageBase64 }) {
   if (config.ocrProvider === "mock") {
     return runMockOcr({ ocrType, imageBase64 });
@@ -26,6 +51,8 @@ async function runOcrForInvoice({ config, ocrType, imageBase64 }) {
 }
 
 module.exports = {
+  buildMissingRequiredOcrError,
+  getMissingRequiredOcrFields,
   isOcrEnabledForType,
   runOcrForInvoice,
 };

server/src/services/ocr/mock.js

@@ -18,6 +18,18 @@ function buildInvoiceKey(markerText) {
   return `hash:${hash.slice(0, 20)}`;
 }
 
+function parseMissingFields(markerText) {
+  const match = markerText.match(/MOCK_OCR_MISSING:([A-Za-z0-9_,-]+)/);
+  if (!match?.[1]) {
+    return [];
+  }
+
+  return match[1]
+    .split(",")
+    .map((field) => field.trim())
+    .filter(Boolean);
+}
+
 function toAmount(seedHex, fallback) {
   const value = Number.parseInt(seedHex.slice(0, 8), 16);
   if (!Number.isFinite(value)) {
@@ -38,7 +50,7 @@ async function runMockOcr({ ocrType, imageBase64 }) {
   const invoiceCode = `MOCK${hash.slice(0, 8).toUpperCase()}`;
   const invoiceNumber = hash.slice(8, 16).toUpperCase();
 
-  return {
+  const result = {
     invoiceCode,
     invoiceNumber,
     invoiceDate: "2026-02-08",
@@ -53,6 +65,14 @@ async function runMockOcr({ ocrType, imageBase64 }) {
       marker: markerText.slice(0, 128),
     },
   };
+
+  for (const field of parseMissingFields(markerText)) {
+    if (Object.prototype.hasOwnProperty.call(result, field)) {
+      result[field] = null;
+    }
+  }
+
+  return result;
 }
 
 module.exports = {

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,
 };