H1-H10 优化OCR实时体验闭环

这个提交包含在:
2026-05-09 11:48:57 +08:00
父节点 549bd81e30
当前提交 c26cf9b954
共修改 34 个文件,包含 1217 行新增和 184 行删除
+66 -10
查看文件
@@ -1,6 +1,7 @@
const cron = require("node-cron");
const { buildMissingRequiredOcrError, runOcrForInvoice } = require("../services/ocr");
const { publishInvoiceUpdated } = require("../services/events");
const { resolveInvoiceDedupBlocker } = require("../services/invoiceGovernance");
const { readFileBuffer } = require("../services/storage");
@@ -70,7 +71,30 @@ async function claimPendingInvoices(prisma, config, workerId) {
ocr_lock_owner = ${workerId}
FROM candidates
WHERE i.id = candidates.id
RETURNING i.id, i.owner_employee_id, i.file_sha256, i.ocr_provider, i.ocr_type, i.storage_path, i.mime_type, i.ocr_attempts
RETURNING i.id, i.owner_employee_id, i.uploaded_by_employee_id, i.file_sha256, i.ocr_provider, i.ocr_type, i.storage_path, i.mime_type, i.ocr_attempts
`;
}
async function claimPendingInvoiceById(prisma, config, workerId, invoiceId) {
const lockTtlSeconds = Math.max(60, Number(config.ocrLockTtlSeconds || 300));
return prisma.$queryRaw`
WITH candidates AS (
SELECT id
FROM "invoice_pool"
WHERE id = ${invoiceId}::uuid
AND status = 'pending_ocr'
AND (ocr_locked_at IS NULL OR ocr_locked_at < NOW() - make_interval(secs => ${lockTtlSeconds}))
ORDER BY created_at ASC
LIMIT 1
FOR UPDATE SKIP LOCKED
)
UPDATE "invoice_pool" AS i
SET ocr_locked_at = NOW(),
ocr_lock_owner = ${workerId}
FROM candidates
WHERE i.id = candidates.id
RETURNING i.id, i.owner_employee_id, i.uploaded_by_employee_id, i.file_sha256, i.ocr_provider, i.ocr_type, i.storage_path, i.mime_type, i.ocr_attempts
`;
}
@@ -226,11 +250,12 @@ async function markInvoiceAsFailed(prisma, claimedInvoice, errorMessage, config)
const now = new Date();
const nextAttempts = Number(claimedInvoice.ocr_attempts || 0) + 1;
const isTerminal = nextAttempts >= config.ocrMaxAttempts;
const nextStatus = isTerminal ? "ocr_failed" : "pending_ocr";
await prisma.invoicePool.update({
where: { id: claimedInvoice.id },
data: {
status: isTerminal ? "ocr_failed" : "pending_ocr",
status: nextStatus,
ocr_provider: getEffectiveOcrProvider(config),
ocr_attempts: nextAttempts,
ocr_last_error: errorMessage,
@@ -239,6 +264,8 @@ async function markInvoiceAsFailed(prisma, claimedInvoice, errorMessage, config)
ocr_lock_owner: null,
},
});
return nextStatus;
}
async function processClaimedInvoice(prisma, config, claimedInvoice, workerId, logger) {
@@ -276,7 +303,7 @@ async function processClaimedInvoice(prisma, config, claimedInvoice, workerId, l
},
"OCR invoice marked as failed for missing required fields",
);
return { outcome: "failed" };
return { outcome: "failed", status: "ocr_failed" };
}
const duplicate = await resolveInvoiceDedupBlocker(
@@ -302,14 +329,14 @@ async function processClaimedInvoice(prisma, config, claimedInvoice, workerId, l
},
"OCR invoice marked as duplicate",
);
return { outcome: "duplicate" };
return { outcome: "duplicate", status: "ocr_failed" };
}
await markInvoiceAsAvailable(prisma, config, claimedInvoice.id, normalized, ocrResultId);
return { outcome: "available" };
return { outcome: "available", status: "available" };
} catch (error) {
const message = trimErrorMessage(error);
await markInvoiceAsFailed(prisma, claimedInvoice, message, config);
const status = await markInvoiceAsFailed(prisma, claimedInvoice, message, config);
logger.error(
{
err: error,
@@ -320,12 +347,11 @@ async function processClaimedInvoice(prisma, config, claimedInvoice, workerId, l
},
"OCR processing failed",
);
return { outcome: "failed" };
return { outcome: "failed", status };
}
}
async function runOcrOnce(prisma, config, logger, workerId = `ocr:${process.pid}`) {
const claimed = await claimPendingInvoices(prisma, config, workerId);
async function processClaimedInvoices(prisma, config, logger, workerId, claimed, options = {}) {
if (!claimed.length) {
return { claimed: 0, available: 0, duplicate: 0, failed: 0 };
}
@@ -341,11 +367,31 @@ async function runOcrOnce(prisma, config, logger, workerId = `ocr:${process.pid}
} else if (result?.outcome === "failed") {
stats.failed += 1;
}
if (typeof options.onInvoiceProcessed === "function") {
await options.onInvoiceProcessed({
invoiceId: item.id,
status: result?.status ?? item.status ?? null,
outcome: result?.outcome ?? null,
ownerEmployeeId: item.owner_employee_id ?? null,
uploadedByEmployeeId: item.uploaded_by_employee_id ?? null,
});
}
}
return stats;
}
async function runOcrOnce(prisma, config, logger, workerId = `ocr:${process.pid}`, options = {}) {
const claimed = await claimPendingInvoices(prisma, config, workerId);
return processClaimedInvoices(prisma, config, logger, workerId, claimed, options);
}
async function runOcrForInvoiceId(prisma, config, logger, invoiceId, workerId = `ocr:${process.pid}`, options = {}) {
const claimed = await claimPendingInvoiceById(prisma, config, workerId, invoiceId);
return processClaimedInvoices(prisma, config, logger, workerId, claimed, options);
}
function startOcrJob(fastify) {
if (!["tencent", "mock"].includes(fastify.config.ocrProvider)) {
fastify.log.info({ provider: fastify.config.ocrProvider }, "OCR job disabled");
@@ -359,6 +405,12 @@ function startOcrJob(fastify) {
let running = false;
const workerId = `ocr-worker:${process.pid}`;
const onInvoiceProcessed = (event) =>
publishInvoiceUpdated(fastify, {
invoiceId: event.invoiceId,
status: event.status,
audienceEmployeeIds: [event.ownerEmployeeId, event.uploadedByEmployeeId],
});
const task = cron.schedule(fastify.config.ocrPollCron, async () => {
if (running) {
@@ -367,7 +419,9 @@ function startOcrJob(fastify) {
running = true;
try {
const result = await runOcrOnce(fastify.prisma, fastify.config, fastify.log, workerId);
const result = await runOcrOnce(fastify.prisma, fastify.config, fastify.log, workerId, {
onInvoiceProcessed,
});
fastify.metrics?.jobRunsTotal?.inc?.({
job: "ocr.tick",
driver: "cron",
@@ -423,9 +477,11 @@ function startOcrJob(fastify) {
module.exports = {
OCR_DUPLICATE_ERROR,
claimPendingInvoiceById,
claimPendingInvoices,
getEffectiveOcrProvider,
markInvoiceAsQualityFailed,
runOcrForInvoiceId,
runOcrOnce,
startOcrJob,
};
+91 -3
查看文件
@@ -2,16 +2,18 @@ const PgBoss = require("pg-boss");
const { revokeExpiredDelegations } = require("./delegationRevokeJob");
const { archiveUnusedInvoices } = require("./invoiceArchiveJob");
const { runOcrOnce } = require("./ocrJob");
const { runOcrForInvoiceId, runOcrOnce } = require("./ocrJob");
const { runVerificationForRecord } = require("./verificationJob");
const { runIntegrationOutboxOnce } = require("./integrationOutboxJob");
const { runErpMasterDataSyncOnce } = require("./erpMasterDataSyncJob");
const { publishInvoiceUpdated } = require("../services/events");
const {
FEATURE_FLAG_KEYS,
evaluateFeatureFlag,
} = require("../services/featureFlags");
const JOB_OCR_TICK = "ocr.tick";
const JOB_OCR_INVOICE = "ocr.invoice";
const JOB_DELEGATION_REVOKE = "delegation.revoke-expired";
const JOB_INVOICE_ARCHIVE = "invoice.archive-unused";
const JOB_INVOICE_VERIFY = "invoice.verify";
@@ -43,6 +45,7 @@ async function startPgBossJobs(fastify) {
});
await boss.createQueue(JOB_OCR_TICK);
await boss.createQueue(JOB_OCR_INVOICE);
await boss.createQueue(JOB_DELEGATION_REVOKE);
await boss.createQueue(JOB_INVOICE_ARCHIVE);
await boss.createQueue(JOB_INVOICE_VERIFY);
@@ -51,6 +54,13 @@ async function startPgBossJobs(fastify) {
await boss.createQueue(JOB_ERP_MASTER_DATA_SYNC);
}
const onOcrInvoiceProcessed = (event) =>
publishInvoiceUpdated(fastify, {
invoiceId: event.invoiceId,
status: event.status,
audienceEmployeeIds: [event.ownerEmployeeId, event.uploadedByEmployeeId],
});
await boss.work(JOB_DELEGATION_REVOKE, { batchSize: 1 }, async (jobs) => {
const job = jobs?.[0];
try {
@@ -122,7 +132,9 @@ async function startPgBossJobs(fastify) {
const job = jobs?.[0];
const workerId = `pgboss:ocr:${process.pid}`;
try {
const stats = await runOcrOnce(fastify.prisma, fastify.config, fastify.log, workerId);
const stats = await runOcrOnce(fastify.prisma, fastify.config, fastify.log, workerId, {
onInvoiceProcessed: onOcrInvoiceProcessed,
});
fastify.metrics?.jobRunsTotal?.inc?.({
job: JOB_OCR_TICK,
driver: "pgboss",
@@ -164,6 +176,67 @@ async function startPgBossJobs(fastify) {
}
});
await boss.work(JOB_OCR_INVOICE, { batchSize: 1 }, async (jobs) => {
if (!["tencent", "mock"].includes(fastify.config.ocrProvider)) {
return;
}
const job = jobs?.[0];
const invoiceId = job?.data?.invoiceId;
if (!invoiceId) {
return;
}
const workerId = `pgboss:ocr-invoice:${process.pid}`;
try {
const stats = await runOcrForInvoiceId(
fastify.prisma,
fastify.config,
fastify.log,
invoiceId,
workerId,
{ onInvoiceProcessed: onOcrInvoiceProcessed },
);
fastify.metrics?.jobRunsTotal?.inc?.({
job: JOB_OCR_INVOICE,
driver: "pgboss",
outcome: "success",
});
if (stats.available > 0) {
fastify.metrics?.jobProcessedTotal?.inc?.(
{ job: JOB_OCR_INVOICE, driver: "pgboss", kind: "available" },
stats.available,
);
}
if (stats.duplicate > 0) {
fastify.metrics?.jobProcessedTotal?.inc?.(
{ job: JOB_OCR_INVOICE, driver: "pgboss", kind: "duplicate" },
stats.duplicate,
);
}
if (stats.failed > 0) {
fastify.metrics?.jobProcessedTotal?.inc?.(
{ job: JOB_OCR_INVOICE, driver: "pgboss", kind: "failed" },
stats.failed,
);
}
if (stats.claimed > 0) {
fastify.log.info(
{ jobId: job?.id, invoiceId, workerId, ...stats },
"pg-boss OCR invoice job processed",
);
}
} catch (error) {
fastify.metrics?.jobRunsTotal?.inc?.({
job: JOB_OCR_INVOICE,
driver: "pgboss",
outcome: "failure",
});
fastify.log.error({ err: error, jobId: job?.id, invoiceId }, "pg-boss OCR invoice job failed");
throw error;
}
});
await boss.work(JOB_INVOICE_VERIFY, { batchSize: 1 }, async (jobs) => {
const job = jobs?.[0];
const recordId = job?.data?.recordId;
@@ -301,7 +374,7 @@ async function startPgBossJobs(fastify) {
await boss.schedule(JOB_OCR_TICK, fastify.config.ocrPollCron, null, {
tz: fastify.config.appTimezone,
singletonKey: JOB_OCR_TICK,
singletonSeconds: 55,
singletonSeconds: 1,
});
} else {
await boss.unschedule(JOB_OCR_TICK).catch(() => {});
@@ -359,12 +432,27 @@ async function startPgBossJobs(fastify) {
return boss;
}
async function enqueueOcrInvoice(fastify, invoiceId) {
if (
fastify?.config?.jobDriver === "pgboss" &&
fastify.boss &&
["tencent", "mock"].includes(fastify.config.ocrProvider)
) {
await fastify.boss.send(JOB_OCR_INVOICE, { invoiceId });
return { mode: "queued" };
}
return { mode: "fallback" };
}
module.exports = {
JOB_OCR_INVOICE,
JOB_OCR_TICK,
JOB_DELEGATION_REVOKE,
JOB_INVOICE_ARCHIVE,
JOB_INVOICE_VERIFY,
JOB_INTEGRATION_OUTBOX,
JOB_ERP_MASTER_DATA_SYNC,
enqueueOcrInvoice,
startPgBossJobs,
};