fix: WebSocket connectivity and DensePose initialization in dev mode

ui/config/api.config.js

@@ -1,13 +1,15 @@
 // API Configuration for WiFi-DensePose UI
 
-// Auto-detect the backend URL from the page origin so the UI works whether
-// served from Docker (:3000), local dev (:8080), or any other port.
-const _origin = (typeof window !== 'undefined' && window.location && window.location.origin)
-  ? window.location.origin
-  : 'http://localhost:3000';
+// Backend URL configuration
+// When running locally, frontend is on :3000 and backend is on :8000
+const _backendUrl = (typeof window !== 'undefined' && window.location) 
+  ? (window.location.port === '3000' 
+      ? 'http://localhost:8000'  // Local dev: frontend on :3000, backend on :8000
+      : window.location.origin)   // Production/Docker: same origin
+  : 'http://localhost:8000';
 
 export const API_CONFIG = {
-  BASE_URL: _origin,
+  BASE_URL: _backendUrl,
   API_VERSION: '/api/v1',
   WS_PREFIX: 'ws://',
   WSS_PREFIX: 'wss://',
@@ -121,9 +123,10 @@ export function buildWsUrl(endpoint, params = {}) {
     ? API_CONFIG.WSS_PREFIX
     : API_CONFIG.WS_PREFIX;
 
-  // Derive host from the page origin so it works on any port (Docker :3000, dev :8080, etc.)
-  const host = window.location.host;
-  let url = `${protocol}${host}${endpoint}`;
+  // Extract host from BASE_URL (e.g., "http://localhost:8000" → "localhost:8000")
+  // This ensures WebSocket connects to backend port, not frontend port
+  const backendHost = API_CONFIG.BASE_URL.replace(/^https?:\/\//, '');
+  let url = `${protocol}${backendHost}${endpoint}`;
 
   // Add query parameters
   const queryParams = new URLSearchParams(params);

ui/services/sensing.service.js

@@ -1,19 +1,19 @@
 /**
  * Sensing WebSocket Service
  *
- * Manages the connection to the Python sensing WebSocket server
- * (ws://localhost:8765) and provides a callback-based API for the UI.
+ * Manages the connection to the Python backend WebSocket server
+ * (ws://localhost:8000/api/v1/stream/pose) and provides a callback-based API for the UI.
  *
  * Falls back to simulated data only after MAX_RECONNECT_ATTEMPTS exhausted.
  * While reconnecting the service stays in "reconnecting" state and does NOT
  * emit simulated frames so the UI can clearly distinguish live vs. fallback data.
  */
 
-// Derive WebSocket URL from the page origin so it works on any port.
-// The /ws/sensing endpoint is available on the same HTTP port (3000).
+// Derive WebSocket URL - connect to backend on port 8000 for local dev
+const _isLocalDev = (typeof window !== 'undefined' && window.location.port === '3000');
 const _wsProto = (typeof window !== 'undefined' && window.location.protocol === 'https:') ? 'wss:' : 'ws:';
-const _wsHost  = (typeof window !== 'undefined' && window.location.host) ? window.location.host : 'localhost:3000';
-const SENSING_WS_URL = `${_wsProto}//${_wsHost}/ws/sensing`;
+const _wsHost  = _isLocalDev ? 'localhost:8000' : window.location.host;
+const SENSING_WS_URL = `${_wsProto}//${_wsHost}/api/v1/stream/pose`;
 const RECONNECT_DELAYS = [1000, 2000, 4000, 8000, 16000];
 const MAX_RECONNECT_ATTEMPTS = 20;
 // Number of failed attempts that must occur before simulation starts.
@@ -308,20 +308,65 @@ class SensingService {
   // ---- Data handling -----------------------------------------------------
 
   _handleData(data) {
-    this._lastMessage = data;
+    // Transform backend pose_data format to sensing format for compatibility
+    let transformedData = data;
+
+    if (data.type === 'pose_data') {
+      // Backend is sending pose data - transform to sensing format
+      transformedData = {
+        type: 'sensing_update',
+        timestamp: new Date(data.timestamp).getTime() / 1000,
+        source: 'server-simulated', // Backend is in mock mode
+        nodes: [{
+          node_id: 1,
+          rssi_dbm: -45,
+          position: [2, 0, 1.5],
+          amplitude: [],
+          subcarrier_count: 0,
+        }],
+        features: {
+          mean_rssi: -45,
+          variance: data.data?.confidence || 0.5,
+          std: Math.sqrt(data.data?.confidence || 0.5),
+          motion_band_power: data.data?.pose?.count > 0 ? 0.15 : 0.05,
+          breathing_band_power: 0.05,
+          dominant_freq_hz: 0.3,
+          change_points: data.data?.pose?.count || 0,
+          spectral_power: 0.2,
+          range: 1.5,
+          iqr: 1.0,
+          skewness: 0,
+          kurtosis: 1,
+        },
+        classification: {
+          motion_level: data.data?.pose?.count > 0 ? 'active' : 'present_still',
+          presence: data.data?.pose?.count > 0,
+          confidence: data.data?.confidence || 0.5,
+        },
+        // Store original pose data for components that need it
+        _pose_data: data.data
+      };
+    } else if (data.type === 'connection_established') {
+      // Connection confirmation - set data source to server-simulated
+      console.log('[Sensing] WebSocket connected:', data);
+      this._setDataSource('server-simulated');
+      return; // Don't emit connection message as data
+    }
+
+    this._lastMessage = transformedData;
 
     // Track the server's source field from each frame so the UI
     // can react if the server switches between esp32 ↔ simulated at runtime.
-    if (data.source && this._state === 'connected') {
-      const raw = data.source;
+    if (transformedData.source && this._state === 'connected') {
+      const raw = transformedData.source;
       if (raw !== this._serverSource) {
         this._applyServerSource(raw);
       }
     }
 
     // Update RSSI history for sparkline
-    if (data.features && data.features.mean_rssi != null) {
-      this._rssiHistory.push(data.features.mean_rssi);
+    if (transformedData.features && transformedData.features.mean_rssi != null) {
+      this._rssiHistory.push(transformedData.features.mean_rssi);
       if (this._rssiHistory.length > this._maxHistory) {
         this._rssiHistory.shift();
       }
@@ -330,7 +375,7 @@ class SensingService {
     // Notify all listeners
     for (const cb of this._listeners) {
       try {
-        cb(data);
+        cb(transformedData);
       } catch (e) {
         console.error('[Sensing] Listener error:', e);
       }

v1/src/services/pose_service.py

@@ -107,16 +107,29 @@ async def initialize(self):
     async def _initialize_models(self):
         """Initialize neural network models."""
         try:
+            # DensePose model configuration
+            densepose_config = {
+                'input_channels': 256,  # Feature channels from modality translator
+                'num_body_parts': 24,   # 24 body parts for DensePose
+                'num_uv_coordinates': 2,  # U and V coordinates
+                'hidden_channels': [256, 128, 64],
+                'kernel_size': 3,
+                'padding': 1,
+                'dropout_rate': 0.1,
+                'use_fpn': False,
+                'output_stride': 4
+            }
+
             # Initialize DensePose model
             if self.settings.pose_model_path:
-                self.densepose_model = DensePoseHead()
+                self.densepose_model = DensePoseHead(densepose_config)
                 # Load model weights if path is provided
                 # model_state = torch.load(self.settings.pose_model_path)
                 # self.densepose_model.load_state_dict(model_state)
                 self.logger.info("DensePose model loaded")
             else:
                 self.logger.warning("No pose model path provided, using default model")
-                self.densepose_model = DensePoseHead()
+                self.densepose_model = DensePoseHead(densepose_config)
 
             # Initialize modality translation
             config = {