Update app.py

app.py

@@ -20,6 +20,7 @@ from transformers import GroundingDinoForObjectDetection, GroundingDinoProcessor
 from diffusers import FluxPipeline
 from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM
 import gc
+from PIL import Image, ImageDraw, ImageFont
 
 def clear_memory():
     """메모리 정리 함수"""
@@ -339,16 +340,14 @@ def on_change_bbox(prompts: dict[str, Any] | None):
 def on_change_prompt(img: Image.Image | None, prompt: str | None, bg_prompt: str | None = None):
     return gr.update(interactive=bool(img and prompt))
 
-# process_prompt 함수 수정
 def process_prompt(img: Image.Image, prompt: str, bg_prompt: str | None = None, 
                   aspect_ratio: str = "1:1", position: str = "bottom-center", 
                   scale_percent: float = 100, text_params: dict | None = None) -> tuple[Image.Image, Image.Image]:
     try:
-        
-
         if img is None or prompt.strip() == "":
             raise gr.Error("Please provide both image and prompt")
         
+        
         print(f"Processing with position: {position}, scale: {scale_percent}")
         
         try:
@@ -358,6 +357,7 @@ def process_prompt(img: Image.Image, prompt: str, bg_prompt: str | None = None,
         except Exception as e:
             print(f"Translation error (continuing with original text): {str(e)}")
         
+        # 기존 처리 로직...
         results, _ = _process(img, prompt, bg_prompt, aspect_ratio)
         
         if bg_prompt:
@@ -368,14 +368,13 @@ def process_prompt(img: Image.Image, prompt: str, bg_prompt: str | None = None,
                     position=position,
                     scale_percent=scale_percent
                 )
-                print(f"Combined image created with position: {position}")
-                return combined, results[2]
             except Exception as e:
                 print(f"Combination error: {str(e)}")
-                return results[1], results[2]
-        
-        return results[1], results[2]
+                combined = results[1]
+        else:
+            combined = results[1]
 
+        # 텍스트 추가 로직을 여기로 이동
         if text_params and text_params.get('text'):
             combined = add_text_to_image(combined, text_params)
         
@@ -386,7 +385,6 @@ def process_prompt(img: Image.Image, prompt: str, bg_prompt: str | None = None,
     finally:
         clear_memory()
 
-
 def process_bbox(img: Image.Image, box_input: str) -> tuple[Image.Image, Image.Image]:
     try:
         if img is None or box_input.strip() == "":
@@ -556,6 +554,7 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
     """)
 
     with gr.Row():
+        # 입력 컬럼
         with gr.Column(scale=1):
             input_image = gr.Image(
                 type="pil",
@@ -567,6 +566,8 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
                 placeholder="Enter what you want to extract...",
                 interactive=True
             )
+            
+            # 배경 및 비율 설정
             with gr.Row():
                 bg_prompt = gr.Textbox(
                     label="Background Prompt (optional)",
@@ -583,10 +584,12 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
                     scale=1
                 )
 
+            # 오브젝트 컨트롤
             with gr.Row(visible=False) as object_controls:
+                # 위치 컨트롤
                 with gr.Column(scale=1):
+                    position = gr.State(value="bottom-center")
                     with gr.Row():
-                        position = gr.State(value="bottom-center")
                         btn_top_left = gr.Button("↖")
                         btn_top_center = gr.Button("↑")
                         btn_top_right = gr.Button("↗")
@@ -598,6 +601,8 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
                         btn_bottom_left = gr.Button("↙")
                         btn_bottom_center = gr.Button("↓")
                         btn_bottom_right = gr.Button("↘")
+                
+                # 크기 컨트롤
                 with gr.Column(scale=1):
                     scale_slider = gr.Slider(
                         minimum=10,
@@ -607,65 +612,99 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
                         label="Object Size (%)"
                     )
 
+            # 텍스트 입력 섹션
+            with gr.Group() as text_group:
+                text_input = gr.Textbox(
+                    label="Text to Add",
+                    placeholder="Enter text..."
+                )
+                with gr.Row():
+                    with gr.Column(scale=1):
+                        font_size = gr.Slider(
+                            minimum=10,
+                            maximum=800,
+                            value=400,
+                            step=10,
+                            label="Font Size"
+                        )
+                        thickness = gr.Slider(
+                            minimum=0,
+                            maximum=20,
+                            value=0,
+                            step=1,
+                            label="Text Thickness"
+                        )
+                        color_dropdown = gr.Dropdown(
+                            choices=["White", "Black", "Red", "Green", "Blue", "Yellow", "Purple"],
+                            value="White",
+                            label="Text Color"
+                        )
+                    with gr.Column(scale=1):
+                        opacity_slider = gr.Slider(
+                            minimum=0,
+                            maximum=255,
+                            value=255,
+                            step=1,
+                            label="Opacity"
+                        )
+                        text_x_position = gr.Slider(
+                            minimum=0,
+                            maximum=100,
+                            value=50,
+                            step=1,
+                            label="Text X Position (%)"
+                        )
+                        text_y_position = gr.Slider(
+                            minimum=0,
+                            maximum=100,
+                            value=50,
+                            step=1,
+                            label="Text Y Position (%)"
+                        )
+
+            # 처리 버튼
             process_btn = gr.Button(
                 "Process",
                 variant="primary",
                 interactive=False
             )
 
-            # 각 버튼에 대한 클릭 이벤트 처리
-            def update_position(new_position):
-                return new_position
-
-            btn_top_left.click(fn=lambda: update_position("top-left"), outputs=position)
-            btn_top_center.click(fn=lambda: update_position("top-center"), outputs=position)
-            btn_top_right.click(fn=lambda: update_position("top-right"), outputs=position)
-            btn_middle_left.click(fn=lambda: update_position("middle-left"), outputs=position)
-            btn_middle_center.click(fn=lambda: update_position("middle-center"), outputs=position)
-            btn_middle_right.click(fn=lambda: update_position("middle-right"), outputs=position)
-            btn_bottom_left.click(fn=lambda: update_position("bottom-left"), outputs=position)
-            btn_bottom_center.click(fn=lambda: update_position("bottom-center"), outputs=position)
-            btn_bottom_right.click(fn=lambda: update_position("bottom-right"), outputs=position)
-
+        # 출력 컬럼
         with gr.Column(scale=1):
-            with gr.Row():
-                combined_image = gr.Image(
-                    label="Combined Result",
-                    show_download_button=True,
-                    type="pil",
-                    height=512
-                )
-            with gr.Row():
-                extracted_image = gr.Image(
-                    label="Extracted Object",
-                    show_download_button=True,
-                    type="pil",
-                    height=256
-                )
+            combined_image = gr.Image(
+                label="Combined Result",
+                show_download_button=True,
+                type="pil",
+                height=512
+            )
+            extracted_image = gr.Image(
+                label="Extracted Object",
+                show_download_button=True,
+                type="pil",
+                height=256
+            )
 
-    
-    with gr.Row():
-        with gr.Column(scale=1):
-            # 텍스트 입력 섹션 추가
-            with gr.Group(visible=True) as text_group:
-                text_input = gr.Textbox(label="Text to Add", placeholder="Enter text...")
-                font_size = gr.Slider(minimum=10, maximum=800, value=400, step=10, 
-                                    label="Font Size")
-                thickness = gr.Slider(minimum=0, maximum=20, value=0, step=1,
-                                    label="Text Thickness")
-                color_dropdown = gr.Dropdown(
-                    choices=["White", "Black", "Red", "Green", "Blue", "Yellow", "Purple"],
-                    value="White",
-                    label="Text Color"
-                )
-                opacity_slider = gr.Slider(minimum=0, maximum=255, value=255, step=1, 
-                                         label="Opacity")
-                text_x_position = gr.Slider(minimum=0, maximum=100, value=50, step=1, 
-                                          label="Text X Position (%)")
-                text_y_position = gr.Slider(minimum=0, maximum=100, value=50, step=1, 
-                                          label="Text Y Position (%)")
-    
-    # 이벤트 바인딩 수정
+    # 이벤트 핸들러
+    def get_text_params():
+        return {
+            'text': text_input.value,
+            'font_size': font_size.value,
+            'thickness': thickness.value,
+            'color': color_dropdown.value,
+            'opacity': opacity_slider.value,
+            'x_position': text_x_position.value,
+            'y_position': text_y_position.value
+        }
+
+    # 위치 버튼 이벤트
+    for btn, pos in [
+        (btn_top_left, "top-left"), (btn_top_center, "top-center"), (btn_top_right, "top-right"),
+        (btn_middle_left, "middle-left"), (btn_middle_center, "middle-center"), (btn_middle_right, "middle-right"),
+        (btn_bottom_left, "bottom-left"), (btn_bottom_center, "bottom-center"), (btn_bottom_right, "bottom-right")
+    ]:
+        btn.click(fn=lambda p=pos: p, outputs=position)
+
+    # 메인 프로세스 이벤트
     process_btn.click(
         fn=process_prompt,
         inputs=[
@@ -675,70 +714,35 @@ with gr.Blocks(theme=gr.themes.Soft(), css=css) as demo:
             aspect_ratio,
             position,
             scale_slider,
-            # 텍스트 파라미터 추가
-            gr.State(lambda: {
-                'text': text_input.value,
-                'font_size': font_size.value,
-                'thickness': thickness.value,
-                'color': color_dropdown.value,
-                'opacity': opacity_slider.value,
-                'x_position': text_x_position.value,
-                'y_position': text_y_position.value
-            })
+            gr.State(get_text_params)
         ],
         outputs=[combined_image, extracted_image]
     )
 
-    
-    # Event bindings
+    # UI 업데이트 이벤트
     input_image.change(
         fn=update_process_button,
         inputs=[input_image, text_prompt],
-        outputs=process_btn,
-        queue=False
+        outputs=process_btn
     )
     
     text_prompt.change(
         fn=update_process_button,
         inputs=[input_image, text_prompt],
-        outputs=process_btn,
-        queue=False
+        outputs=process_btn
     )
 
-    def update_controls(bg_prompt):
-        """배경 프롬프트 입력 여부에 따라 컨트롤 표시 업데이트"""
-        is_visible = bool(bg_prompt)
-        return [
-            gr.update(visible=is_visible),  # aspect_ratio
-            gr.update(visible=is_visible),  # object_controls
-        ]
-
     bg_prompt.change(
         fn=update_controls,
-        inputs=bg_prompt,
-        outputs=[aspect_ratio, object_controls],
-        queue=False
+        inputs=[bg_prompt],
+        outputs=[aspect_ratio, object_controls]
     )
 
-    process_btn.click(
-        fn=process_prompt,
-        inputs=[
-            input_image,
-            text_prompt,
-            bg_prompt,
-            aspect_ratio,
-            position,
-            scale_slider
-        ],
-        outputs=[combined_image, extracted_image],
-        queue=True
-    )
-
-
-demo.queue(max_size=5)  # 큐 크기 제한
+# 런처 설정
+demo.queue(max_size=5)
 demo.launch(
     server_name="0.0.0.0",
     server_port=7860,
     share=False,
-    max_threads=2  # 스레드 수 제한
-)
+    max_threads=2
+)