Enhance MobileCallPage with chat transcript feature

- Introduce a new `MobileCallTranscript` component to display chat messages during calls, improving user interaction.
- Implement message timestamp formatting and automatic scrolling for better usability.
- Add a toggle button to switch between camera view and chat view, enhancing the user experience during calls.
- Refactor the layout of the `MobileCallPage` to accommodate the new chat functionality while maintaining existing video call features.

frontend/src/components/pages/MobileCallPage.tsx

@@ -1,7 +1,16 @@
 "use client";
 
-import { useCallback, useEffect, useRef } from "react";
+import { useCallback, useEffect, useRef, useState } from "react";
-import { Camera, Headphones, Loader2, Mic, Phone, PhoneOff, Video } from "lucide-react";
+import {
+  Camera,
+  Headphones,
+  Loader2,
+  MessageSquareText,
+  Mic,
+  Phone,
+  PhoneOff,
+  Video,
+} from "lucide-react";
 import { Button } from "@/components/ui/button";
 import {
   DropdownMenu,
@@ -20,7 +29,81 @@ import {
   SelectValue,
 } from "@/components/ui/select";
 import { useCameraPreview } from "@/hooks/use-camera-preview";
-import { useVoicePreview } from "@/hooks/use-voice-preview";
+import {
+  useVoicePreview,
+  type ChatMessage,
+} from "@/hooks/use-voice-preview";
+
+type CallView = "camera" | "chat";
+
+function messageTime(timestamp: string): string {
+  const date = new Date(timestamp);
+  if (Number.isNaN(date.getTime())) return "";
+  return date.toLocaleTimeString("zh-CN", {
+    hour: "2-digit",
+    minute: "2-digit",
+    hour12: false,
+  });
+}
+
+function MobileCallTranscript({ messages }: { messages: ChatMessage[] }) {
+  const scrollRef = useRef<HTMLDivElement>(null);
+
+  useEffect(() => {
+    const container = scrollRef.current;
+    if (container) container.scrollTop = container.scrollHeight;
+  }, [messages]);
+
+  return (
+    <div className="absolute inset-0 z-[1] bg-[#07101a]/95 backdrop-blur-sm">
+      <div
+        ref={scrollRef}
+        className="scrollbar-subtle h-full overflow-y-auto px-4 pb-28 pt-20"
+      >
+        {messages.length === 0 ? (
+          <div className="flex h-full flex-col items-center justify-center gap-2 pb-16 text-center">
+            <MessageSquareText className="size-7 text-white/40" />
+            <p className="text-sm font-medium text-white/80">暂无聊天记录</p>
+            <p className="max-w-64 text-xs leading-5 text-white/45">
+              通话中的语音转写和助手回复会实时显示在这里。
+            </p>
+          </div>
+        ) : (
+          <div className="flex flex-col gap-4">
+            {messages.map((message) => {
+              const time = messageTime(message.timestamp);
+              const isAssistant = message.role === "assistant";
+              return (
+                <div
+                  key={message.id}
+                  className={[
+                    "flex max-w-[88%] flex-col gap-1",
+                    isAssistant ? "self-start" : "self-end items-end",
+                  ].join(" ")}
+                >
+                  <span className="px-1 text-[11px] text-white/40">
+                    {isAssistant ? "助手" : "我"}
+                    {time ? ` · ${time}` : ""}
+                  </span>
+                  <div
+                    className={[
+                      "whitespace-pre-wrap rounded-2xl px-3.5 py-2.5 text-sm leading-6 shadow-sm",
+                      isAssistant
+                        ? "rounded-tl-sm bg-white/10 text-white/90"
+                        : "rounded-tr-sm bg-white text-[#07101a]",
+                    ].join(" ")}
+                  >
+                    {message.content || (message.streaming ? "…" : "")}
+                  </div>
+                </div>
+              );
+            })}
+          </div>
+        )}
+      </div>
+    </div>
+  );
+}
 
 function CallAudioDeviceSelect({
   micValue,
@@ -163,6 +246,7 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
     selectDevice,
     selectOutputDevice,
     supportsOutputSelection,
+    messages,
     connect,
     replaceVideoStream,
     disconnect,
@@ -179,6 +263,7 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
     selectCamera: changeCamera,
   } = camera;
   const videoRef = useRef<HTMLVideoElement>(null);
+  const [view, setView] = useState<CallView>("camera");
   const connecting = status === "connecting";
   const inCall = connecting || status === "connected";
 
@@ -219,22 +304,28 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
   const error = previewError || cameraError;
 
   return (
-    <main className="relative isolate h-dvh min-h-80 w-full overflow-hidden bg-[#07101a] text-white">
+    <main className="flex h-dvh w-full items-center justify-center overflow-hidden bg-black lg:p-4">
       <audio ref={audioRef} autoPlay playsInline className="hidden" />
+      <div
+        data-testid="mobile-call-viewport"
+        className="relative isolate h-dvh min-h-80 w-full overflow-hidden bg-[#07101a] text-white lg:h-[calc(100dvh-2rem)] lg:min-h-0 lg:w-auto lg:aspect-[9/16] lg:rounded-[2rem] lg:ring-1 lg:ring-white/15 lg:shadow-2xl"
+      >
       <video
         ref={videoRef}
         autoPlay
         playsInline
         muted
         className={[
-          "absolute inset-0 h-full w-full -scale-x-100 object-cover transition-opacity duration-300",
+          "absolute inset-0 h-full w-full -scale-x-100 object-cover transition-opacity duration-300 lg:object-contain",
-          cameraStream ? "opacity-100" : "opacity-0",
+          cameraStream && view === "camera" ? "opacity-100" : "opacity-0",
         ].join(" ")}
       />
 
       <div className="pointer-events-none absolute inset-0 bg-gradient-to-b from-black/35 via-transparent to-black/65" />
 
-      <div className="absolute left-1/2 top-[max(1rem,env(safe-area-inset-top))] flex -translate-x-1/2 items-center gap-2 rounded-full border border-white/10 bg-black/30 px-3 py-1.5 text-xs text-white/85 backdrop-blur-md">
+      {view === "chat" && <MobileCallTranscript messages={messages} />}
+
+      <div className="absolute left-1/2 top-[max(1rem,env(safe-area-inset-top))] z-10 flex -translate-x-1/2 items-center gap-2 rounded-full border border-white/10 bg-black/30 px-3 py-1.5 text-xs text-white/85 backdrop-blur-md">
         <span
           className={[
             "h-2 w-2 rounded-full",
@@ -252,8 +343,22 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
             : "通话已结束"}
       </div>
 
-      {!cameraStream && inCall && (
+      <button
-        <div className="absolute inset-0 flex items-center justify-center px-8 text-center">
+        type="button"
+        onClick={() => setView((current) => current === "camera" ? "chat" : "camera")}
+        aria-label={view === "camera" ? "显示聊天记录" : "显示摄像头画面"}
+        title={view === "camera" ? "聊天记录" : "摄像头画面"}
+        className="absolute right-4 top-[max(1rem,env(safe-area-inset-top))] z-10 flex size-9 items-center justify-center rounded-full border border-white/15 bg-black/40 text-white shadow-lg backdrop-blur-md transition-colors hover:bg-black/60 focus-visible:outline-none focus-visible:ring-2 focus-visible:ring-white/40"
+      >
+        {view === "camera" ? (
+          <MessageSquareText className="size-[18px]" />
+        ) : (
+          <Video className="size-[18px]" />
+        )}
+      </button>
+
+      {!cameraStream && inCall && view === "camera" && (
+        <div className="absolute inset-0 z-[2] flex items-center justify-center px-8 text-center">
           <div className="flex max-w-sm flex-col items-center gap-3">
             {cameraStarting ? (
               <Loader2 size={34} className="animate-spin text-white/70" />
@@ -270,7 +375,7 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
       )}
 
       {!inCall && (
-        <div className="absolute inset-0 flex items-center justify-center px-8">
+        <div className="absolute inset-0 z-[2] flex items-center justify-center px-8">
           <div className="flex max-w-sm flex-col items-center gap-4 text-center">
             <Button
               type="button"
@@ -286,7 +391,7 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
       )}
 
       {inCall && (
-        <div className="absolute bottom-[max(1.25rem,env(safe-area-inset-bottom))] left-1/2 flex -translate-x-1/2 items-center gap-3 min-[380px]:gap-5 landscape:bottom-[max(1rem,env(safe-area-inset-bottom))]">
+        <div className="absolute bottom-[max(1.25rem,env(safe-area-inset-bottom))] left-1/2 z-10 flex -translate-x-1/2 items-center gap-3 min-[380px]:gap-5 landscape:bottom-[max(1rem,env(safe-area-inset-bottom))]">
           <CallAudioDeviceSelect
             micValue={selectedDeviceId}
             micDevices={audioInputs}
@@ -314,6 +419,7 @@ export function MobileCallPage({ assistantId }: { assistantId: string }) {
           />
         </div>
       )}
+      </div>
     </main>
   );
 }