Update app.py

app.py

@@ -2,7 +2,6 @@
 import os
 import json
 from typing import Dict, List, Optional, Tuple, Union, Any
-import io
 
 import gradio as gr
 import spaces
@@ -138,10 +137,11 @@ def face_image_embedding(image: Image.Image) -> List[float] | None:
         return None
 
 @spaces.GPU
-def scenes_extraction(video_file: io.IOBase, threshold: float, offset_frames: int, crop_ratio: float) -> Tuple[List[Image.Image], List[Dict]] | None:
+def scenes_extraction(video_file: str, threshold: float, offset_frames: int, crop_ratio: float) -> Tuple[List[Image.Image], List[Dict]] | None:
+    # video_file es un str ya que aunque realmente el usuario subió un archivo desde la UI, Gradio lo guarda temporalmente como ruta
+    
     # Detectamos las escenas
-    video_path = video_file.name
-    video_manager = VideoManager([video_path])
+    video_manager = VideoManager([video_file])
     scene_manager = SceneManager()
     scene_manager.add_detector(ContentDetector(threshold=threshold))
     video_manager.start()
@@ -226,7 +226,7 @@ with gr.Blocks(title="Salamandra Vision 7B · ZeroGPU") as demo:
     face_btn.click(face_image_embedding, [face_img], face_out, api_name="face_image_embedding", concurrency_limit=1)
 
     with gr.Row():
-        video_file = gr.Video(label="Sube un vídeo", type="file")
+        video_file = gr.Video(label="Sube un vídeo")
         threshold = gr.Slider(0.0, 100.0, value=30.0, step=1.0, label="Threshold")
         offset_frames = gr.Slider(0, 30, value=5, step=1, label="Offset frames")
         crop_ratio = gr.Slider(0.0, 1.0, value=1.0, step=0.05, label="Crop ratio")