feat(message): implement version-based incremental sync and WS compression

Implement a more efficient conversation synchronization mechanism using
version numbers instead of sequence numbers to reduce bandwidth usage.
This includes adding support for Gzip decompression on WebSocket
messages to optimize data transfer.

- Add `pako` for WebSocket message decompression
- Implement `getSyncByVersion` in `MessageService`
- Implement `syncByVersion` in `MessageSyncService` to handle incremental
  updates and conversation state changes
- Update `WebSocketService` to support binary frames and Gzip inflation
- Add `syncVersion` to `MessageStore` for tracking synchronization state

src/services/core/wsService.ts

@@ -1,3 +1,5 @@
+import pako from 'pako';
+
 import { AppState, AppStateStatus } from 'react-native';
 
 import { api, WS_URL } from './api';
@@ -329,7 +331,7 @@ class WebSocketService {
   private fallbackCheckTimer: ReturnType<typeof setTimeout> | null = null;
 
   private getWSUrl(token: string | null): string {
-    return `${WS_URL}?token=${encodeURIComponent(token || '')}`;
+    return `${WS_URL}?token=${encodeURIComponent(token || '')}&compress=1`;
   }
 
   async connect(): Promise<boolean> {
@@ -354,6 +356,7 @@ class WebSocketService {
       console.log('[WebSocket] Connecting to:', url.replace(token, '***'));
       
       this.ws = new WebSocket(url);
+      this.ws.binaryType = 'arraybuffer';
 
       this.ws.onopen = () => {
         console.log('[WebSocket] Connected successfully');
@@ -392,9 +395,17 @@ class WebSocketService {
     }
   }
 
-  private handleIncoming(data: string): void {
+  private handleIncoming(data: string | ArrayBuffer): void {
+    let textData: string;
+    if (typeof data === 'string') {
+      textData = data;
+    } else {
+      // Binary frame — gzip decompress
+      const decompressed = pako.inflate(new Uint8Array(data as ArrayBuffer));
+      textData = new TextDecoder('utf-8').decode(decompressed);
+    }
     // 后端 writePump 会将多条消息以 \n 分隔合并发送，需要逐条解析
-    const lines = data.split('\n');
+    const lines = textData.split('\n');
     for (const line of lines) {
       const trimmed = line.trim();
       if (!trimmed) continue;
@@ -402,7 +413,7 @@ class WebSocketService {
         const msg: WSServerMessage = JSON.parse(trimmed);
         this.processMessage(msg);
       } catch (error) {
-        console.error('Failed to parse WebSocket message:', error, 'raw:', trimmed);
+        console.error('Failed to parse WebSocket message:', error);
       }
     }
   }

src/services/message/messageService.ts

@@ -476,6 +476,37 @@ class MessageService {
     return response.data?.conversations || [];
   }
 
+  /**
+   * 增量同步（基于版本号）
+   * GET /api/v1/conversations/sync?version=N&limit=100
+   */
+  async getSyncByVersion(version: number, limit: number = 100): Promise<{
+    current_version: number;
+    changes: Array<{
+      conversation_id: string;
+      change_type: string;
+      max_seq?: number;
+      last_message_at?: number;
+      version: number;
+    }>;
+    full_sync: boolean;
+    has_more: boolean;
+  }> {
+    const response = await api.get<{
+      current_version: number;
+      changes: Array<{
+        conversation_id: string;
+        change_type: string;
+        max_seq?: number;
+        last_message_at?: number;
+        version: number;
+      }>;
+      full_sync: boolean;
+      has_more: boolean;
+    }>('/conversations/sync', { version, limit });
+    return response.data;
+  }
+
   /**
    * 获取系统消息未读数
    * GET /api/v1/messages/system/unread-count