fix: web-probe 输出 DOM 诊断分组 (#584)

Co-authored-by: Codex <codex@noreply.local>

scripts/src/hwlab-node-web-observe-runner-source.ts

@@ -1075,7 +1075,7 @@ const report = {
 await writeFile(reportJsonPath, JSON.stringify(report, null, 2) + "\n", { mode: 0o600 });
 await writeFile(reportMdPath, renderMarkdown(report), { mode: 0o600 });
 const [jsonMeta, mdMeta] = await Promise.all([fileMeta(reportJsonPath), fileMeta(reportMdPath)]);
-console.log(JSON.stringify({ ok: true, command: "web-probe-observe analyze", stateDir, reportJsonPath, reportJsonSha256: jsonMeta.sha256, reportMdPath, reportMdSha256: mdMeta.sha256, counts: report.counts, sampleMetrics: sampleMetrics.summary, pageProvenance: pageProvenance.summary, pagePerformance: pagePerformance.summary, promptNetwork: promptNetwork.summary, runtimeAlerts: runtimeAlerts.summary, turnTimingRecentUpdateSawtoothJumps: sampleMetrics.turnTimingRecentUpdateSawtoothJumps.slice(0, 20), turnTimingRecentUpdateLargestSteps: sampleMetrics.turnTimingRecentUpdateLargestSteps.slice(0, 20), pagePerformanceSlowApi: pagePerformance.sameOriginApiByPath.filter((item) => item.overFiveSecondCount > 0).slice(0, 20), pagePerformanceLongLivedStreams: pagePerformance.sameOriginApiByPath.filter((item) => item.isLongLivedStream).slice(0, 20), findings: findings.slice(0, 20), valuesRedacted: true }, null, 2));
+console.log(JSON.stringify({ ok: true, command: "web-probe-observe analyze", stateDir, reportJsonPath, reportJsonSha256: jsonMeta.sha256, reportMdPath, reportMdSha256: mdMeta.sha256, counts: report.counts, sampleMetrics: sampleMetrics.summary, pageProvenance: pageProvenance.summary, pagePerformance: pagePerformance.summary, promptNetwork: promptNetwork.summary, runtimeAlerts: runtimeAlerts.summary, domDiagnosticGroups: runtimeAlerts.domDiagnosticsByText.slice(0, 20), turnTimingRecentUpdateSawtoothJumps: sampleMetrics.turnTimingRecentUpdateSawtoothJumps.slice(0, 20), turnTimingRecentUpdateLargestSteps: sampleMetrics.turnTimingRecentUpdateLargestSteps.slice(0, 20), pagePerformanceSlowApi: pagePerformance.sameOriginApiByPath.filter((item) => item.overFiveSecondCount > 0).slice(0, 20), pagePerformanceLongLivedStreams: pagePerformance.sameOriginApiByPath.filter((item) => item.isLongLivedStream).slice(0, 20), findings: findings.slice(0, 20), valuesRedacted: true }, null, 2));
 
 async function readJson(file) {
   try { return JSON.parse(await readFile(file, "utf8")); } catch { return null; }
@@ -1129,7 +1129,7 @@ function buildFindings(samples, control, network, errors, sampleMetrics, promptN
   if (recentUpdateSawtoothJumps.length > 0) findings.push({ id: "turn-timing-recent-update-sawtooth-jump", severity: "amber", summary: "最近更新 value jumped faster than sample interval; expected sawtooth increase-or-reset", count: recentUpdateSawtoothJumps.length, samples: recentUpdateSawtoothJumps.slice(0, 20) });
   if ((runtimeAlerts?.summary?.httpErrorCount ?? 0) > 0) findings.push({ id: "runtime-http-errors", severity: "amber", summary: "natural page requests returned HTTP error status during observation", count: runtimeAlerts.summary.httpErrorCount, groups: runtimeAlerts.networkHttpErrorsByPath.slice(0, 12) });
   if ((runtimeAlerts?.summary?.requestFailedCount ?? 0) > 0) findings.push({ id: "runtime-requestfailed", severity: "amber", summary: "browser requestfailed events were captured during observation", count: runtimeAlerts.summary.requestFailedCount, groups: runtimeAlerts.networkRequestFailedByPath.slice(0, 12) });
-  if ((runtimeAlerts?.summary?.domDiagnosticSampleCount ?? 0) > 0) findings.push({ id: "runtime-dom-diagnostics", severity: "amber", summary: "diagnostic/error/warning-like text was visible in sampled DOM", count: runtimeAlerts.summary.domDiagnosticSampleCount, samples: runtimeAlerts.domDiagnostics.slice(0, 12) });
+  if ((runtimeAlerts?.summary?.domDiagnosticSampleCount ?? 0) > 0) findings.push({ id: "runtime-dom-diagnostics", severity: "amber", summary: "diagnostic/error/warning-like text was visible in sampled DOM", count: runtimeAlerts.summary.domDiagnosticSampleCount, groupCount: runtimeAlerts.summary.domDiagnosticGroupCount ?? 0, groups: runtimeAlerts.domDiagnosticsByText.slice(0, 12), samples: runtimeAlerts.domDiagnostics.slice(0, 12) });
   if ((runtimeAlerts?.summary?.executionErrorCount ?? 0) > 0) findings.push({ id: "runtime-execution-errors", severity: "red", summary: "Workbench rendered execution failure/error rows during observation", count: runtimeAlerts.summary.executionErrorCount, groups: runtimeAlerts.runtimeExecutionErrorsByCode.slice(0, 12) });
   if ((runtimeAlerts?.summary?.consoleAlertCount ?? 0) > 0) findings.push({ id: "runtime-console-alerts", severity: "amber", summary: "browser console warning/error entries were captured during observation", count: runtimeAlerts.summary.consoleAlertCount, groups: runtimeAlerts.consoleAlertsByPath.slice(0, 12) });
   const slowApi = Array.isArray(pagePerformance?.sameOriginApiByPath) ? pagePerformance.sameOriginApiByPath.filter((item) => item.overFiveSecondCount > 0) : [];
@@ -1604,6 +1604,7 @@ function buildRuntimeAlerts(samples, control, network, consoleEvents, errors) {
       httpErrorCount: httpErrors.length,
       requestFailedCount: requestFailed.length,
       domDiagnosticSampleCount: domDiagnostics.length,
+      domDiagnosticGroupCount: groupDomDiagnostics(domDiagnostics).length,
       executionErrorCount: executionErrors.length,
       baselineExecutionErrorCount: baselineExecutionErrors.length,
       consoleAlertCount: consoleAlerts.length,
@@ -1617,6 +1618,7 @@ function buildRuntimeAlerts(samples, control, network, consoleEvents, errors) {
     networkHttpErrorsByPath: groupNetworkAlerts(httpErrors),
     networkRequestFailedByPath: groupNetworkAlerts(requestFailed),
     domDiagnostics: domDiagnostics.slice(0, 80),
+    domDiagnosticsByText: groupDomDiagnostics(domDiagnostics),
     runtimeExecutionErrors: executionErrors.slice(0, 120),
     runtimeExecutionErrorsByCode: groupExecutionErrors(executionErrors),
     baselineRuntimeExecutionErrors: baselineExecutionErrors.slice(0, 80),
@@ -1627,6 +1629,86 @@ function buildRuntimeAlerts(samples, control, network, consoleEvents, errors) {
   };
 }
 
+function groupDomDiagnostics(events) {
+  const groups = new Map();
+  for (const item of events || []) {
+    const preview = String(item?.preview || "").trim();
+    const normalizedPreview = normalizeDiagnosticPreview(preview);
+    const key = [
+      item?.source || "",
+      item?.diagnosticCode || "",
+      normalizedPreview
+    ].join("|");
+    const existing = groups.get(key) || {
+      source: item?.source || null,
+      diagnosticCode: item?.diagnosticCode || null,
+      textHash: item?.textHash || null,
+      normalizedPreview,
+      preview,
+      count: 0,
+      firstAt: item?.ts || null,
+      lastAt: item?.ts || null,
+      promptIndexes: new Set(),
+      traceIds: new Set(),
+      sampleSeqs: []
+    };
+    existing.count += 1;
+    existing.firstAt = minIso(existing.firstAt, item?.ts || null);
+    existing.lastAt = maxIso(existing.lastAt, item?.ts || null);
+    if (Number.isFinite(Number(item?.promptIndex))) existing.promptIndexes.add(Number(item.promptIndex));
+    for (const traceId of extractDiagnosticTraceIds(item, preview)) existing.traceIds.add(traceId);
+    if (existing.sampleSeqs.length < 12 && item?.seq !== undefined && item?.seq !== null) existing.sampleSeqs.push(item.seq);
+    groups.set(key, existing);
+  }
+  return Array.from(groups.values())
+    .map((item) => ({
+      source: item.source,
+      diagnosticCode: item.diagnosticCode,
+      textHash: item.textHash,
+      normalizedPreview: item.normalizedPreview,
+      preview: item.preview,
+      count: item.count,
+      firstAt: item.firstAt,
+      lastAt: item.lastAt,
+      promptIndexes: Array.from(item.promptIndexes).sort((a, b) => a - b),
+      traceIds: Array.from(item.traceIds).sort(),
+      sampleSeqs: item.sampleSeqs
+    }))
+    .sort((a, b) => (b.count - a.count) || String(a.firstAt || "").localeCompare(String(b.firstAt || "")));
+}
+
+function normalizeDiagnosticPreview(text) {
+  return String(text || "")
+    .replace(/trace_id=[A-Za-z0-9_-]+/gu, "trace_id=:traceId")
+    .replace(/\btrc_[A-Za-z0-9_-]+\b/gu, "trc_:traceId")
+    .replace(/\bses_[A-Za-z0-9_-]+\b/gu, "ses_:sessionId")
+    .replace(/\brun_[A-Za-z0-9_-]+\b/gu, "run_:runId")
+    .replace(/\bcmd_[A-Za-z0-9_-]+\b/gu, "cmd_:commandId")
+    .replace(/\s+/gu, " ")
+    .trim();
+}
+
+function extractDiagnosticTraceIds(item, preview) {
+  const ids = new Set();
+  if (item?.traceId) ids.add(String(item.traceId));
+  const text = String(preview || "");
+  for (const match of text.matchAll(/\btrc_[A-Za-z0-9_-]+\b/gu)) ids.add(match[0]);
+  for (const match of text.matchAll(/trace_id=([A-Za-z0-9_-]+)/gu)) ids.add(match[1]);
+  return ids;
+}
+
+function minIso(a, b) {
+  if (!a) return b || null;
+  if (!b) return a || null;
+  return Date.parse(a) <= Date.parse(b) ? a : b;
+}
+
+function maxIso(a, b) {
+  if (!a) return b || null;
+  if (!b) return a || null;
+  return Date.parse(a) >= Date.parse(b) ? a : b;
+}
+
 function sampleExecutionErrorCandidates(sample) {
   const candidates = [];
   const add = (source, items) => {