dsh-livebench-panel 0.1.20 → 0.1.21

package/lib/index.js

@@ -554,7 +554,7 @@ function apply(ctx) {
 	/** All runs: runId -> record. Several evaluations may run concurrently. */
 	const runs = new Map();
 
-	const startRun = async (body) => {
+	const startRun = async (body, autoRetries = 0) => {
 		const layout = livebenchLayout();
 		if (!layout.available) {
 			return { status: 409, payload: { ok: false, error: `LiveBench venv not found under ${layout.root}` } };
@@ -651,6 +651,7 @@ function apply(ctx) {
 			args.push("--question-end", String(asInt(body.end, 0, 100000, 0)));
 		}
 		if (body.resume === true) args.push("--resume");
+		if (body.retryFailures === true) args.push("--retry-failures");
 
 		const env = {
 			...process.env,
@@ -698,6 +699,8 @@ function apply(ctx) {
 			id: `${Date.now()}-${Math.random().toString(36).slice(2, 6)}`,
 			proc: child,
 			displayName,
+			body,
+			autoRetries,
 			exitCode: null,
 			startedAt: new Date().toISOString(),
 			command: [layout.pythonExe, ...args].join(" "),
@@ -720,6 +723,20 @@ function apply(ctx) {
 		child.on("close", (code) => {
 			record.exitCode = code === null ? -1 : code;
 			record.log.push(`[exit ${record.exitCode}]`);
+			// 自动补跑：正常结束但存在 $ERROR$ 答案时，用 --resume --retry-failures
+			// 只重跑失败题（最多 2 轮）。网络瞬断/超时导致的失败会被自动修复。
+			if (record.exitCode === 0 && record.autoRetries < 2 && record.body) {
+				const errors = countErrorAnswers(layout.dataDir, record.displayName);
+				if (errors > 0) {
+					record.log.push(`[自动补跑] 检测到 ${errors} 条失败答案，自动重试（第 ${record.autoRetries + 1}/2 轮）`);
+					const retryBody = { ...record.body, resume: true, retryFailures: true };
+					setTimeout(() => {
+						startRun(retryBody, record.autoRetries + 1).catch((e) => {
+							appendLog(record, `[自动补跑启动失败] ${e.message}`);
+						});
+					}, 3000);
+				}
+			}
 		});
 
 		return { status: 200, payload: { ok: true, runId: record.id, displayName, command: record.command } };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-livebench-panel",
-  "version": "0.1.20",
+  "version": "0.1.21",
   "description": "DSH web plugin: a LiveBench tab in the Trajectory view (right of 对话/轨迹). Run LiveBench evaluations against every model configured in the DeepSeek Harness — pick provider/model, category, task, release and question range from dropdowns, watch progress, and read scores in place.",
   "license": "MIT",
   "type": "module",