Update app.py

app.py

@@ -12,6 +12,14 @@ import cv2
 import subprocess
 import tempfile
 
+# Папки для файлов
+UPLOADS_IMAGES_DIR = './uploads/images'
+UPLOADS_VIDEOS_DIR = './uploads/videos'
+OUTPUT_DIR = './output'
+os.makedirs(UPLOADS_IMAGES_DIR, exist_ok=True)
+os.makedirs(UPLOADS_VIDEOS_DIR, exist_ok=True)
+os.makedirs(OUTPUT_DIR, exist_ok=True)
+
 class WanAnimateApp:
     def __init__(self):
         model_name = "stabilityai/stable-video-diffusion-img2vid-xt"
@@ -22,24 +30,45 @@ class WanAnimateApp:
             device_map="cpu"
         )
 
+    def save_uploaded_image(self, img):
+        """Сохраняет загруженное изображение в папку, возвращает путь"""
+        if img is not None:
+            filename = f"image_{uuid.uuid4()}.png"
+            path = os.path.join(UPLOADS_IMAGES_DIR, filename)
+            if isinstance(img, Image.Image):
+                img.save(path)
+            else:
+                Image.open(img).save(path)
+            return path
+        return None
+
+    def save_uploaded_video(self, vid):
+        """Сохраняет загруженное видео в папку, возвращает путь"""
+        if vid is not None:
+            filename = f"video_{uuid.uuid4()}.mp4"
+            path = os.path.join(UPLOADS_VIDEOS_DIR, filename)
+            if isinstance(vid, str):
+                shutil.copy(vid, path)
+            else:
+                with open(path, 'wb') as f:
+                    f.write(vid)
+            return path
+        return None
+
     def predict(
         self, 
-        ref_img,
-        video,
+        ref_img_path,
+        video_path,
         model_id,
         model,
     ):
-        if ref_img is None or video is None:
+        if ref_img_path is None or video_path is None:
             return None, "Upload both image and video."
 
         try:
-            # Local processing
-            if isinstance(ref_img, Image.Image):
-                ref_image = ref_img.convert("RGB").resize((576, 320))
-            else:
-                ref_image = Image.open(ref_img).convert("RGB").resize((576, 320))
-
-            cap = cv2.VideoCapture(video)
+            # Local processing with saved paths
+            ref_image = Image.open(ref_img_path).convert("RGB").resize((576, 320))
+            cap = cv2.VideoCapture(video_path)
             frame_count = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
             cap.release()
             motion_hint = f" with dynamic motion from {frame_count} frames"
@@ -68,7 +97,7 @@ class WanAnimateApp:
             temp_dir = tempfile.mkdtemp()
             for i, frame in enumerate(output):
                 frame.save(f"{temp_dir}/frame_{i:04d}.png")
-            temp_video = f"/tmp/output_{uuid.uuid4()}.mp4"
+            temp_video = os.path.join(OUTPUT_DIR, f"output_{uuid.uuid4()}.mp4")
             subprocess.run([
                 'ffmpeg', '-y', '-framerate', '7', '-i', f"{temp_dir}/frame_%04d.png",
                 '-c:v', 'libx264', '-pix_fmt', 'yuv420p', temp_video
@@ -111,68 +140,49 @@ def start_app():
         
         gr.HTML("""
             <details>
-                <summary>‼️Usage (使用说明)</summary>
-                Wan-Animate supports two modes:
-                <ul>
-                    <li>Move Mode: animate the character in input image with movements from the input video</li>
-                    <li>Mix Mode: replace the character in input video with the character in input image</li>
-                </ul>
-                Wan-Animate supports two modes:
-                <ul>
-                    <li>Move Mode: Use the movements extracted from the input video to drive the character in the input image</li>
-                    <li>Mix Mode: Use the character in the input image to replace the character in the input video</li>
-                </ul>
-                Currently, the following restrictions apply to inputs:
-                <ul>
-                    <li>Video file size: Less than 200MB</li>
-                    <li>Video resolution: The shorter side must be greater than 200, and the longer side must be less than 2048</li>
-                    <li>Video duration: 2s to 30s</li>
-                    <li>Video aspect ratio: 1:3 to 3:1</li>
-                    <li>Video formats: mp4, avi, mov</li>
-                    <li>Image file size: Less than 5MB</li>
-                    <li>Image resolution: The shorter side must be greater than 200, and the longer side must be less than 4096</li>
-                    <li>Image formats: jpg, png, jpeg, webp, bmp</li>
-                </ul>
-                Current, the inference quality has two variants. You can use our open-source code for more flexible configuration.
+                <summary>‼️Usage (использования)</summary>
+                Wan-Animate поддерживает два режима:
                 <ul>
-                    <li>wan-pro: 25fps, 720p</li>
-                    <li>wan-std: 15fps, 720p</li>
+                    <li>Move Mode: анимирует перс��нажа в изображении движениями из видео</li>
+                    <li>Mix Mode: заменяет персонажа в видео на персонажа из изображения</li>
                 </ul>
+                Ограничения: Видео <200MB, 2–30s, 200–2048 res; Изображение <5MB, 200–4096 res.
+                Quality: wan-pro (25fps, 720p), wan-std (15fps, 720p).
             </details>
         """)
 
         with gr.Row():
             with gr.Column():    
                 ref_img = gr.Image(
-                    label="Reference Image(参考图像)",
-                    type="pil",
+                    label="Reference Image(изображение)",
+                    type="filepath",
                     sources=["upload"],
                 )
                 
                 video = gr.Video(
-                    label="Template Video(模版视频)",
+                    label="Template Video(шаблонное видео)",
                     sources=["upload"],
                 )
                 
                 with gr.Row():
                     model_id = gr.Dropdown(
-                        label="Mode(模式)",
+                        label="Mode(режим)",
                         choices=["wan2.2-animate-move", "wan2.2-animate-mix"],
                         value="wan2.2-animate-move",
                         info=""
                     )
 
                     model = gr.Dropdown(
-                        label="Inference Quality",
+                        label="Inference Quality(качество)",
                         choices=["wan-pro", "wan-std"],
                         value="wan-pro",
                     )
 
-                run_button = gr.Button("Generate Video(生成视频)")
+                run_button = gr.Button("Generate Video(генерировать видео)")
 
             with gr.Column():
-                output_video = gr.Video(label="Output Video(输出视频)")
-                output_status = gr.Textbox(label="Status(状态)")
+                output_video = gr.Video(label="Output Video(выходное видео)")
+                output_status = gr.Textbox(label="Status(статус)")
         
         run_button.click(
             fn=app.predict,