Update app.py

app.py

@@ -1,22 +1,14 @@
 import cv2
-import os
 import numpy as np
 import gradio as gr
-from moviepy import VideoFileClip
+import subprocess
+import os
 
-# Robust MediaPipe Import
-try:
-    import mediapipe as mp
-    from mediapipe.python.solutions import pose as mp_pose
-    from mediapipe.python.solutions import drawing_utils as mp_drawing
-    print("MediaPipe Solutions loaded successfully!")
-except ImportError as e:
-    print(f"Direct import failed, trying legacy: {e}")
-    import mediapipe as mp
-    mp_pose = mp.solutions.pose
-    mp_drawing = mp.solutions.drawing_utils
+# 1. Direct Imports (Bypasses the 'solutions' AttributeError bug)
+from mediapipe.python.solutions import pose as mp_pose
+from mediapipe.python.solutions import drawing_utils as mp_drawing
 
-# Create "Meaty" drawing specs
+# 2. Meaty drawing specs for EbSynth tracking
 MEATY_CONNECTIONS = mp_drawing.DrawingSpec(color=(0, 255, 0), thickness=10, circle_radius=5)
 MEATY_JOINTS = mp_drawing.DrawingSpec(color=(255, 255, 255), thickness=12, circle_radius=15)
 
@@ -24,33 +16,31 @@ def extract_pose(video_path):
     if video_path is None:
         return None
     
-    # 1. Load Video
+    output_path = "final_output.mp4"
+    temp_video = "temp_silent.mp4"
+    
+    # Initialize Video Capture
     cap = cv2.VideoCapture(video_path)
     width  = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
     height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
     fps    = cap.get(cv2.CAP_PROP_FPS)
     
-    output_path = "stickman_output.mp4"
-    temp_output = "temp_no_audio.mp4"
-    
-    # Define Video Writer
     fourcc = cv2.VideoWriter_fourcc(*'mp4v')
-    out = cv2.VideoWriter(temp_output, fourcc, fps, (width, height))
+    out = cv2.VideoWriter(temp_video, fourcc, fps, (width, height))
     
-    with mp_pose.Pose(min_detection_confidence=0.5, min_tracking_confidence=0.5, model_complexity=1) as pose:
+    # Process Video
+    with mp_pose.Pose(min_detection_confidence=0.5, min_tracking_confidence=0.5, model_complexity=1) as pose_tracker:
         while cap.isOpened():
             ret, frame = cap.read()
             if not ret:
                 break
             
-            # Convert to RGB for MediaPipe
             rgb_frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
-            results = pose.process(rgb_frame)
+            results = pose_tracker.process(rgb_frame)
             
-            # Create Black Canvas (The "Clear" Background)
+            # Pure Black Canvas
             canvas = np.zeros((height, width, 3), dtype=np.uint8)
             
-            # Draw Meaty Stickman
             if results.pose_landmarks:
                 mp_drawing.draw_landmarks(
                     canvas, 
@@ -65,27 +55,36 @@ def extract_pose(video_path):
     cap.release()
     out.release()
     
-    # 2. Merge Original Audio back using MoviePy v2.x syntax
+    # 3. Native FFmpeg Audio Merge
     try:
-        original_clip = VideoFileClip(video_path)
-        stickman_clip = VideoFileClip(temp_output)
-        
-        # New MoviePy v2.x method: with_audio instead of set_audio
-        final_video = stickman_clip.with_audio(original_clip.audio)
-        final_video.write_videofile(output_path, codec="libx264", audio_codec="aac")
-        
+        # Command copies the silent stickman video and maps the original audio to it
+        command = [
+            "ffmpeg", "-y",
+            "-i", temp_video,
+            "-i", video_path,
+            "-c:v", "copy",
+            "-c:a", "aac",
+            "-map", "0:v:0",
+            "-map", "1:a:0?", # The '?' prevents a crash if the source video has no audio track
+            "-shortest",
+            output_path
+        ]
+        subprocess.run(command, check=True, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
         return output_path
+    except subprocess.CalledProcessError as e:
+        print("FFmpeg routing error, returning silent video.")
+        return temp_video
     except Exception as e:
-        print(f"Audio merge failed: {e}")
-        return temp_output # Fallback to silent video
+        print("General error:", str(e))
+        return temp_video
 
-# Gradio UI
+# 4. Gradio Interface
 interface = gr.Interface(
     fn=extract_pose,
     inputs=gr.Video(label="Upload Dancing Clip (15-30s)"),
     outputs=gr.Video(label="Meaty Stickman Output"),
     title="Pose-to-Stickman Extractor",
-    description="Extracts high-thickness pose data for EbSynth & Motion Transfer."
+    description="Extracts thick pose data on a solid black background for EbSynth motion tracking."
 )
 
 if __name__ == "__main__":