R73-A2 收敛 VAT OCR 归一化与质量门槛

这个提交包含在:
2026-04-28 21:13:21 +08:00
父节点 82fbd39114
当前提交 b7c018922c
共修改 9 个文件,包含 600 行新增和 14 行删除
+46 -1
查看文件
@@ -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,
};