Update app.py

app.py

@@ -15,55 +15,52 @@ def predict(file):
     try:
         with open(file.name, "r") as f:
             data = json.load(f)
-        
-        # Process the data (example: return the file content length)
-        return f"File successfully processed. Content: {len(data)} items."
-    except Exception as e:
-        return f"Error processing file: {str(e)}"
-
-    # 确保 JSON 格式正确
-    channels = ["AccelX", "AccelY", "AccelZ", "GyroX", "GyroY", "GyroZ"]
-    # if not all(channel in data for channel in channels):
-    #     return {"error": "JSON 文件格式错误，缺少必要通道数据。"}
 
-    # 转换为 NumPy 数组，形状为 (6, sequence_length)
-    imu_data = np.array([data[channel] for channel in channels])
+        # 确保 JSON 格式正确
+        channels = ["AccelX", "AccelY", "AccelZ", "GyroX", "GyroY", "GyroZ"]
+        # if not all(channel in data for channel in channels):
+        #     return {"error": "JSON 文件格式错误，缺少必要通道数据。"}
 
-    #Process data
-    len_data = imu_data.shape[1]
-    window_size = 30
-    overlap = 15
-    step = window_size - overlap
-    predictions = [0,0,0,0]
-    
+        # 转换为 NumPy 数组，形状为 (6, sequence_length)
+        imu_data = np.array([data[channel] for channel in channels])
 
-    for start in range(0,len_data,step):
-        end = start + window_size
-        if end > len_data:
-            # Zero padding to last window
-            window = np.pad(imu_data[:, start:], ((0, 0), (0, end - len_data)), mode='constant')
-        else:
-            window = imu_data[:, start:end]
+        #Process data
+        len_data = imu_data.shape[1]
+        window_size = 30
+        overlap = 15
+        step = window_size - overlap
+        predictions = [0,0,0,0]
         
-        input_tensor = torch.tensor(window, dtype=torch.float32).unsqueeze(0)
 
-        with torch.no_grad():
-            output = model(input_tensor) 
-            probabilities = torch.softmax(output, dim=-1).cpu().numpy()
-            predicted_class = np.argmax(probabilities)
-            #print(probabilities)
-            if np.max(probabilities) > 0.70:    #Threshold
-                predictions[predicted_class] += 1
-    #取眾數
-    max_count = max(predictions)
-    predictions = [x if x == max_count else 0 for x in predictions]
+        for start in range(0,len_data,step):
+            end = start + window_size
+            if end > len_data:
+                # Zero padding to last window
+                window = np.pad(imu_data[:, start:], ((0, 0), (0, end - len_data)), mode='constant')
+            else:
+                window = imu_data[:, start:end]
+            
+            input_tensor = torch.tensor(window, dtype=torch.float32).unsqueeze(0)
 
-    #Result
-    result = (
-        f"bicep: {predictions[0]} | abs: {predictions[1]} | "
-        f"chess: {predictions[2]} | legs: {predictions[3]}"
-    )
-    return result
+            with torch.no_grad():
+                output = model(input_tensor) 
+                probabilities = torch.softmax(output, dim=-1).cpu().numpy()
+                predicted_class = np.argmax(probabilities)
+                #print(probabilities)
+                if np.max(probabilities) > 0.70:    #Threshold
+                    predictions[predicted_class] += 1
+        #取眾數
+        max_count = max(predictions)
+        predictions = [x if x == max_count else 0 for x in predictions]
+
+        #Result
+        result = (
+            f"bicep: {predictions[0]} | abs: {predictions[1]} | "
+            f"chess: {predictions[2]} | legs: {predictions[3]}"
+        )
+        return result
+    except Exception as e:
+        return f"Error processing file: {str(e)}"
 
 # 定义 Gradio 界面
 iface = gr.Interface(