Update app.py

app.py

@@ -15,9 +15,7 @@ import math
 from huggingface_hub import hf_hub_download
 from safetensors.torch import load_file
 
-from PIL import Image
 import os
-import gradio as gr
 from gradio_client import Client, handle_file
 import tempfile
 from typing import Optional, Tuple, Any
@@ -70,38 +68,25 @@ def _generate_video_segment(
     """
     Generate a single video segment between two frames by calling an external
     Wan 2.2 image-to-video service hosted on Hugging Face Spaces.
-
-    This helper function is used internally when the user asks to create
-    a video between the input and output images.
-
-    Args:
-        input_image_path (str):
-            Path to the starting frame image on disk.
-        output_image_path (str):
-            Path to the ending frame image on disk.
-        prompt (str):
-            Text prompt describing the camera movement / transition.
-        request (gr.Request):
-            Gradio request object, used here to forward the `x-ip-token`
-            header to the downstream Space for authentication/rate limiting.
-
-    Returns:
-        str:
-            A string returned by the external service, usually a URL or path
-            to the generated video.
     """
     x_ip_token = request.headers['x-ip-token']
     video_client = Client(
         "multimodalart/wan-2-2-first-last-frame",
         headers={"x-ip-token": x_ip_token}
     )
+    
+    # استدعاء التوقع
     result = video_client.predict(
         start_image_pil=handle_file(input_image_path),
         end_image_pil=handle_file(output_image_path),
         prompt=prompt,
         api_name="/generate_video",
     )
-    return result[0]["video"]
+    
+    # --- FIX START ---
+    # الخطأ كان هنا: result[0] هو مسار الفيديو مباشرة كنص، وليس قاموساً
+    return result[0]
+    # --- FIX END ---
 
 
 def build_camera_prompt(
@@ -112,28 +97,6 @@ def build_camera_prompt(
 ) -> str:
     """
     Build a camera movement prompt based on the chosen controls.
-
-    This converts the provided control values into a prompt instruction with the corresponding trigger words for the multiple-angles LoRA.
-
-    Args:
-        rotate_deg (float, optional):
-            Horizontal rotation in degrees. Positive values rotate left,
-            negative values rotate right. Defaults to 0.0.
-        move_forward (float, optional):
-            Forward movement / zoom factor. Larger values imply moving the
-            camera closer or into a close-up. Defaults to 0.0.
-        vertical_tilt (float, optional):
-            Vertical angle of the camera:
-            - Negative ≈ bird's-eye view
-            - Positive ≈ worm's-eye view
-            Defaults to 0.0.
-        wideangle (bool, optional):
-            Whether to switch to a wide-angle lens style. Defaults to False.
-
-    Returns:
-        str:
-            A text prompt describing the camera motion. If no controls are
-            active, returns `"no camera movement"`.
     """
     prompt_parts = []
 
@@ -185,56 +148,10 @@ def infer_camera_edit(
     prev_output: Optional[Image.Image] = None,
 ) -> Tuple[Image.Image, int, str]:
     """
-    Edit the camera angles/view of an image with Qwen Image Edit 2509 and dx8152's Qwen-Edit-2509-Multiple-angles LoRA.
-
-    Applies a camera-style transformation (rotation, zoom, tilt, lens)
-    to an input image.
-
-    Args:
-        image (PIL.Image.Image | None, optional):
-            Input image to edit. If `None`, the function will instead try to
-            use `prev_output`. At least one of `image` or `prev_output` must
-            be available. Defaults to None.
-        rotate_deg (float, optional):
-            Horizontal rotation in degrees (-90, -45, 0, 45, 90). Positive values rotate
-            to the left, negative to the right. Defaults to 0.0.
-        move_forward (float, optional):
-            Forward movement / zoom factor (0, 5, 10). Higher values move the
-            camera closer; values >5 switch to a close-up style. Defaults to 0.0.
-        vertical_tilt (float, optional):
-            Vertical tilt (-1 to 1). -1 ≈ bird's-eye view, +1 ≈ worm's-eye view.
-            Defaults to 0.0.
-        wideangle (bool, optional):
-            Whether to use a wide-angle lens style. Defaults to False.
-        seed (int, optional):
-            Random seed for the generation. Ignored if `randomize_seed=True`.
-            Defaults to 0.
-        randomize_seed (bool, optional):
-            If True, a random seed (0..MAX_SEED) is chosen per call.
-            Defaults to True.
-        true_guidance_scale (float, optional):
-            CFG / guidance scale controlling prompt adherence.
-            Defaults to 1.0 since the demo is using a distilled transformer for faster inference.
-        num_inference_steps (int, optional):
-            Number of inference steps. Defaults to 4.
-        height (int, optional):
-            Output image height. Must typically be a multiple of 8.
-            If set to 0, the model will infer a size. Defaults to 1024 if none is provided.
-        width (int, optional):
-            Output image width. Must typically be a multiple of 8.
-            If set to 0, the model will infer a size. Defaults to 1024 if none is provided.
-        prev_output (PIL.Image.Image | None, optional):
-            Previous output image to use as input when no new image is uploaded.
-            Defaults to None.
-
-    Returns:
-        Tuple[PIL.Image.Image, int, str]:
-            - The edited output image.
-            - The actual seed used for generation.
-            - The constructed camera prompt string.
+    Edit the camera angles/view of an image with Qwen Image Edit 2509.
     """
     progress = gr.Progress(track_tqdm=True)
-    
+     
     prompt = build_camera_prompt(rotate_deg, move_forward, vertical_tilt, wideangle)
     print(f"Generated Prompt: {prompt}")
 
@@ -281,25 +198,6 @@ def create_video_between_images(
     """
     Create a short transition video between the input and output images via the 
     Wan 2.2 first-last-frame Space.
-
-    Args:
-        input_image (PIL.Image.Image | None):
-            Starting frame image (the original / previous view).
-        output_image (numpy.ndarray | None):
-            Ending frame image - the output image with the the edited camera angles.
-        prompt (str):
-            The camera movement prompt used to describe the transition.
-        request (gr.Request):
-            Gradio request object, used to forward the `x-ip-token` header
-            to the video generation app.
-
-    Returns:
-        str:
-            a path pointing to the generated video.
-
-    Raises:
-        gr.Error:
-            If either image is missing or if the video generation fails.
     """
     if input_image is None or output_image is None:
         raise gr.Error("Both input and output images are required to create a video.")
@@ -334,33 +232,11 @@ css = '''#col-container { max-width: 800px; margin: 0 auto; }
 def reset_all() -> list:
     """
     Reset all camera control knobs and flags to their default values.
-
-    This is used by the "Reset" button to set:
-    - rotate_deg = 0
-    - move_forward = 0
-    - vertical_tilt = 0
-    - wideangle = False
-    - is_reset = True
-
-    Returns:
-        list:
-            A list of values matching the order of the reset outputs:
-            [rotate_deg, move_forward, vertical_tilt, wideangle, is_reset, True]
     """
     return [0, 0, 0, 0, False, True]
 
 
 def end_reset() -> bool:
-    """
-    Mark the end of a reset cycle.
-
-    This helper is chained after `reset_all` to set the internal
-    `is_reset` flag back to False, so that live inference can resume.
-
-    Returns:
-        bool:
-            Always returns False.
-    """
     return False
 
 
@@ -370,14 +246,6 @@ def update_dimensions_on_upload(
     """
     Compute recommended (width, height) for the output resolution when an
     image is uploaded while preserveing the aspect ratio.
-
-    Args:
-        image (PIL.Image.Image | None):
-            The uploaded image. If `None`, defaults to (1024, 1024).
-
-    Returns:
-        Tuple[int, int]:
-            The new (width, height).
     """
     if image is None:
         return 1024, 1024
@@ -515,21 +383,6 @@ with gr.Blocks(theme=gr.themes.Citrus(), css=css) as demo:
 
     # Manual generation with video button visibility control
     def infer_and_show_video_button(*args: Any):
-        """
-        Wrapper around `infer_camera_edit` that also controls the visibility
-        of the 'Create Video Between Images' button.
-
-        The first argument in `args` is expected to be the input image; if both
-        input and output images are present, the video button is shown.
-
-        Args:
-            *args:
-                Positional arguments forwarded directly to `infer_camera_edit`.
-
-        Returns:
-            tuple:
-                (output_image, seed, prompt, video_button_visibility_update)
-        """
         result_img, result_seed, result_prompt = infer_camera_edit(*args)
         # Show video button if we have both input and output images
         show_button = args[0] is not None and result_img is not None
@@ -629,4 +482,4 @@ with gr.Blocks(theme=gr.themes.Citrus(), css=css) as demo:
     gr.api(infer_camera_edit, api_name="infer_edit_camera_angles")
     gr.api(create_video_between_images, api_name="create_video_between_images")
 
-demo.launch(mcp_server=True, show_api=True)
+demo.launch(mcp_server=True, show_api=True)