feat: add first valid response timing and related metrics

- Introduced `first_valid_response_ms` and `ttfr_ms` to track the timing of the first valid response in LLM calls.
- Updated relevant interfaces and components to display and utilize the new metrics, including CallDetails, CallTable, and LatencyChart.
- Enhanced the database schema to accommodate the new timing fields.
- Implemented logic in the service layer to record the first valid response during model interactions.

apps/desktop/src/api.ts

@@ -86,7 +86,7 @@ export interface LegacyModelImportResult {
 
 export interface ModelConnectivityResult {
   duration_ms: number;
-  first_text_ms: number | null;
+  first_valid_response_ms: number | null;
   output_tokens: number;
   tokens_per_second: number;
   tokens_estimated: boolean;
@@ -197,6 +197,7 @@ export interface LlmCall {
   created_at_ms: number;
   ttfb_ms: number | null;
   ttft_ms: number | null;
+  ttfr_ms: number | null;
   duration_ms: number | null;
   input_tokens: number | null;
   output_tokens: number | null;

apps/desktop/src/components/CallDetails.tsx

@@ -32,6 +32,7 @@ export function CallDetails({ detail }: { detail: CallDetail }) {
     ["Created At", `${call.created_at_ms} · ${new Date(call.created_at_ms).toLocaleString()}`],
     [t("耗时"), timing(call.duration_ms)],
     ["TTFB", timing(call.ttfb_ms)],
+    ["TTFR", timing(call.ttfr_ms)],
     ["TTFT", timing(call.ttft_ms)],
     ["Input Token", show(call.input_tokens)],
     ["Output Token", show(call.output_tokens)],

apps/desktop/src/components/CallTable.tsx

@@ -87,6 +87,11 @@ export function CallTable({ calls, onDetails }: { calls: LlmCall[]; onDetails: (
       header: "TTFB",
       render: (call) => milliseconds(call.ttfb_ms),
     },
+    {
+      key: "ttfr",
+      header: "TTFR",
+      render: (call) => milliseconds(call.ttfr_ms),
+    },
     {
       key: "ttft",
       header: "TTFT",

apps/desktop/src/components/charts/LatencyChart.tsx

@@ -6,13 +6,14 @@ export function LatencyChart({ calls }: { calls: LlmCall[] }) {
   const points = calls.slice(0, 20).reverse();
   const option: EChartsCoreOption = {
     animationDuration: 280,
-    color: ["#d79a62", "#ad84cf"],
+    color: ["#d79a62", "#ad84cf", "#72a8d8"],
     grid: { top: 22, right: 18, bottom: 30, left: 48 },
     legend: { top: 0, right: 8, textStyle: { color: "#999" } },
     tooltip: { trigger: "axis" },
     xAxis: { type: "category", data: points.map((call) => new Date(call.created_at_ms).toLocaleTimeString([], { hour: "2-digit", minute: "2-digit" })), axisLabel: { color: "#888" }, axisLine: { lineStyle: { color: "#5555" } } },
     yAxis: { type: "value", axisLabel: { color: "#888", formatter: "{value} ms" }, splitLine: { lineStyle: { color: "#8882" } } },
     series: [
+      { name: "TTFR", type: "line", smooth: true, symbol: "none", data: points.map((call) => call.ttfr_ms ?? 0) },
       { name: "TTFT", type: "line", smooth: true, symbol: "none", data: points.map((call) => call.ttft_ms ?? 0) },
       { name: t("总耗时"), type: "line", smooth: true, symbol: "none", data: points.map((call) => call.duration_ms ?? 0) },
     ],

apps/desktop/src/components/cursor/CursorModelTestResult.tsx

@@ -21,7 +21,7 @@ export function CursorModelTestResult({ state, testing = false }: { state?: Curs
   const detail = success
     ? t("速度 {speed} tokens/s · 首字 {firstText} ms · 总耗时 {duration} ms · 输出 {tokens} tokens{estimated} · 返回：{output}", {
       speed: formatSpeed(state.result.tokens_per_second),
-      firstText: state.result.first_text_ms ?? "--",
+      firstText: state.result.first_valid_response_ms ?? "--",
       duration: state.result.duration_ms,
       tokens: state.result.output_tokens,
       estimated: state.result.tokens_estimated ? t("（估算）") : "",

apps/desktop/src/demo/api.ts

@@ -55,6 +55,7 @@ const calls: LlmCall[] = Array.from({ length: 24 }, (_, index) => {
     finish_reason: failed ? null : "stop",
     created_at_ms: FIXED_NOW - index * 3 * 60_000,
     ttfb_ms: 210 + index * 13,
+    ttfr_ms: 290 + index * 15,
     ttft_ms: 370 + index * 17,
     duration_ms: failed ? 812 : 1_420 + index * 71,
     input_tokens: 4_800 + index * 337,
@@ -116,7 +117,7 @@ export function installDemoApi() {
     }
     if (path === "/models/import-v0049") return json({ imported: 0, skipped: 0, total: 0 });
     if (/^\/models\/[^/]+\/test\/[^/]+$/.test(path) && method === "POST") {
-      return json({ duration_ms: 1_284, first_text_ms: 418, output_tokens: 42, tokens_per_second: 38.6, tokens_estimated: false, output: "Mock connectivity test passed." });
+      return json({ duration_ms: 1_284, first_valid_response_ms: 418, output_tokens: 42, tokens_per_second: 38.6, tokens_estimated: false, output: "Mock connectivity test passed." });
     }
     if (/^\/models\/[^/]+\/test\/[^/]+$/.test(path) || /^\/models\/[^/]+$/.test(path)) {
       return method === "DELETE" ? empty() : json(models[0]);