Update app.py

app.py

@@ -284,10 +284,20 @@ def generate_all_scenes(prompts_text):
     
     generated_images = []
     formatted_prompts = []
+    progress_messages = []
+    total_scenes = len([s for s in prompts_text.split('='*50) if s.strip()])
+    
+    def update_progress():
+        """Create a progress message showing completed/total scenes"""
+        completed = len(generated_images)
+        message = f"Generated {completed}/{total_scenes} scenes\n\n"
+        if progress_messages:
+            message += "\n".join(progress_messages[-3:])  # Show last 3 status messages
+        return message
     
     sections = prompts_text.split('='*50)
     
-    for section in sections:
+    for section_num, section in enumerate(sections, 1):
         if not section.strip():
             continue
         
@@ -304,7 +314,11 @@ def generate_all_scenes(prompts_text):
         if scene_prompt:
             try:
                 clear_memory()
-                print(f"Generating image for scene: {scene_prompt}")
+                status_msg = f"🎨 Creating scene {section_num}: '{scene_prompt[:50]}...'"
+                progress_messages.append(status_msg)
+                
+                # Yield progress update
+                yield generated_images, "\n\n".join(formatted_prompts), update_progress()
                 
                 image = generate_story_image(scene_prompt)
                 
@@ -317,78 +331,39 @@ def generate_all_scenes(prompts_text):
                     # Verify array shape and type
                     if len(img_array.shape) == 3 and img_array.shape[2] == 3:
                         generated_images.append(img_array)
-                        print(f"Successfully added image for scene {scene_num}")
+                        progress_messages.append(f"✅ Successfully completed scene {section_num}")
                     else:
-                        print(f"Invalid image array shape: {img_array.shape}")
+                        progress_messages.append(f"❌ Error: Invalid image format for scene {section_num}")
                 else:
-                    print(f"Failed to generate image for scene {scene_num}")
+                    progress_messages.append(f"❌ Failed to generate scene {section_num}")
                 
                 clear_memory()
                 
             except Exception as e:
-                print(f"Error generating image for scene {scene_num}: {str(e)}")
+                error_msg = f"❌ Error generating scene {section_num}: {str(e)}"
+                progress_messages.append(error_msg)
                 clear_memory()
                 continue
+            
+            # Yield progress update after each scene
+            yield generated_images, "\n\n".join(formatted_prompts), update_progress()
     
-    # Verify we have images before returning
+    # Final status update
     if not generated_images:
-        print("No images were successfully generated")
-        return [], "\n\n".join(formatted_prompts)
+        progress_messages.append("❌ No images were successfully generated")
+    else:
+        progress_messages.append(f"✅ Successfully completed all {len(generated_images)} scenes!")
     
-    print(f"Successfully generated {len(generated_images)} images")
-    return generated_images, "\n\n".join(formatted_prompts)
-
-def overlay_text_on_image(image, text):
-    if image is None:
-        return None
+    # Final yield
+    yield generated_images, "\n\n".join(formatted_prompts), update_progress()
     
-    try:
-        img = image.convert('RGB')
-        draw = ImageDraw.Draw(img)
-        
-        font_size = int(img.width * 0.025)
-        try:
-            font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", font_size)
-        except:
-            font = ImageFont.load_default()
-
-        y_position = int(img.height * 0.005)
-        x_margin = int(img.width * 0.005)
-        available_width = img.width - (2 * x_margin)
-        
-        wrapped_text = textwrap.fill(text, width=int(available_width / (font_size * 0.6)))
-        
-        outline_color = (255, 255, 255)
-        text_color = (0, 0, 0)
-        offsets = [-2, -1, 1, 2]
-
-        for dx in offsets:
-            for dy in offsets:
-                draw.multiline_text(
-                    (x_margin + dx, y_position + dy),
-                    wrapped_text,
-                    font=font,
-                    fill=outline_color
-                )
-
-        draw.multiline_text(
-            (x_margin, y_position),
-            wrapped_text,
-            font=font,
-            fill=text_color
-        )
-
-        return img
-        
-    except Exception as e:
-        print(f"Error overlaying text: {e}")
-        return None
-
 @spaces.GPU(duration=60)
 def add_text_to_scenes(gallery_images, prompts_text):
     if not isinstance(gallery_images, list):
         return [], []
     
+    clear_memory()
+    
     sections = prompts_text.split('='*50)
     overlaid_images = []
     output_files = []
@@ -410,10 +385,21 @@ def add_text_to_scenes(gallery_images, prompts_text):
         
         if paragraph and image_data is not None:
             try:
+                # Handle tuple case (image, label) from gallery
+                if isinstance(image_data, tuple):
+                    image_data = image_data[0]
+                
+                # Convert numpy array to PIL Image
                 if isinstance(image_data, np.ndarray):
                     image = Image.fromarray(image_data)
                 else:
                     image = image_data
+                
+                print(f"Processing image {i+1}, type: {type(image)}")
+                
+                # Ensure we have a PIL Image
+                if not isinstance(image, Image.Image):
+                    raise TypeError(f"Expected PIL Image, got {type(image)}")
                     
                 overlaid_img = overlay_text_on_image(image, paragraph)
                 if overlaid_img is not None:
@@ -423,13 +409,73 @@ def add_text_to_scenes(gallery_images, prompts_text):
                     output_path = os.path.join(temp_dir, f"panel_{i+1}.png")
                     overlaid_img.save(output_path)
                     output_files.append(output_path)
+                    print(f"Successfully processed image {i+1}")
             except Exception as e:
                 print(f"Error processing image {i+1}: {str(e)}")
                 continue
     
+    if not overlaid_images:
+        print("No images were successfully processed")
+    else:
+        print(f"Successfully processed {len(overlaid_images)} images")
+    
     clear_memory()
     return overlaid_images, output_files
 
+def overlay_text_on_image(image, text):
+    """Helper function to overlay text on an image"""
+    if image is None:
+        return None
+    
+    try:
+        # Ensure we're working with RGB mode
+        img = image.convert('RGB')
+        draw = ImageDraw.Draw(img)
+        
+        # Calculate font size based on image dimensions
+        font_size = int(img.width * 0.025)
+        try:
+            font = ImageFont.truetype("/usr/share/fonts/truetype/dejavu/DejaVuSans-Bold.ttf", font_size)
+        except:
+            font = ImageFont.load_default()
+
+        # Calculate text positioning
+        y_position = int(img.height * 0.005)
+        x_margin = int(img.width * 0.005)
+        available_width = img.width - (2 * x_margin)
+        
+        # Wrap text to fit image width
+        wrapped_text = textwrap.fill(text, width=int(available_width / (font_size * 0.6)))
+        
+        # Add white outline to text for better readability
+        outline_color = (255, 255, 255)
+        text_color = (0, 0, 0)
+        offsets = [-2, -1, 1, 2]
+
+        # Draw text outline
+        for dx in offsets:
+            for dy in offsets:
+                draw.multiline_text(
+                    (x_margin + dx, y_position + dy),
+                    wrapped_text,
+                    font=font,
+                    fill=outline_color
+                )
+
+        # Draw main text
+        draw.multiline_text(
+            (x_margin, y_position),
+            wrapped_text,
+            font=font,
+            fill=text_color
+        )
+
+        return img
+        
+    except Exception as e:
+        print(f"Error in overlay_text_on_image: {e}")
+        return None
+        
 def generate_combined_audio_from_story(story_text, voice='af_heart', speed=1):
     clear_memory()
     
@@ -555,19 +601,19 @@ def create_interface():
         
         with gr.Row():
             prompts_btn = gr.Button("4. Generate Scene Prompts")
-        with gr.Row():
+        with gr.Row():   
             prompts_output = gr.Textbox(label="Generated Scene Prompts", lines=20)
         
         with gr.Row():
             generate_scenes_btn = gr.Button("5. Generate Story Scenes", variant="primary")
             
         with gr.Row():
-            scene_prompts_display = gr.Textbox(
-                label="Scenes Being Generated",
-                lines=8,
+            scene_progress = gr.Textbox(
+                label="Generation Progress",
+                lines=6,
                 interactive=False
             )
-        
+            
         with gr.Row():
             gallery = gr.Gallery(
                 label="Story Scenes",
@@ -576,6 +622,13 @@ def create_interface():
                 height="auto",
                 interactive=False
             )
+            
+        with gr.Row():
+            scene_prompts_display = gr.Textbox(
+                label="Scene Descriptions",
+                lines=8,
+                interactive=False
+            )
         
         with gr.Row():
             add_text_btn = gr.Button("6. Add Text to Scenes", variant="primary")
@@ -627,7 +680,7 @@ def create_interface():
         generate_scenes_btn.click(
             fn=generate_all_scenes,
             inputs=[prompts_output],
-            outputs=[gallery, scene_prompts_display]
+            outputs=[gallery, scene_prompts_display, scene_progress]
         )
         
         add_text_btn.click(
@@ -643,7 +696,6 @@ def create_interface():
         )
 
     return demo
-
 if __name__ == "__main__":
     demo = create_interface()
     demo.launch()