update

app.py

@@ -1,15 +1,70 @@
 import gradio as gr
 import numpy as np
-
 import requests
 from PIL import Image
 from io import BytesIO
 import time
 
-# TODO: PNGinfo追加, 使用データの展開(コピーボタン)
 MAX_SEED = np.iinfo(np.int32).max
 MAX_IMAGE_SIZE = 1344
 
+title = """
+        # Stability AI - Developer Platform WebUI
+        ### UI for using the stable image api
+
+        API Key is required to use this service.
+        https://platform.stability.ai/account/keys
+        
+        Contact: D̷ELL@Stability AI - Advocate (https://x.com/xqdior) / Author: umise (https://x.com/UiE029)
+"""
+
+title_jp = """
+        # Stability AI - Developer Platform WebUI
+        ### このSpaceは、Stable Image APIを使用するためのWEB-UIです。
+
+        このサービスを利用するにはAPIキーが必要です。以下のリンクから取得してください。
+        https://platform.stability.ai/account/keys
+
+        お問い合わせ先: D̷ELL@Stability AI - Advocate (https://x.com/xqdior) / Author: umise (https://x.com/UiE029)
+"""
+
+overview = """
+
+        **Overview**
+        Stability AI’s Stable Image services offer a growing set of APIs for developers to build the best in class image applications.
+
+        - **Disrupting Content Creation:** Stability’s Image APIs are the foundation for applications disrupting publishing, media, gaming, marketing, advertising, design, and more.
+        - **For Developers:** Application developers can build advanced features for designers, photographers, content creators, and a variety of B2C customers.
+        - **Simple APIs:** Stability AI is focused on delivering simple APIs for easy integration into applications with a high bar for quality, alignment, speed, and safety.
+
+        Get Started Now: https://platform.stability.ai/docs/getting-started/stable-image
+        """
+
+overview_jp = """
+        各種モードについて
+        - テキストから生成:
+            プロンプトを基に画像を生成します。
+        - アップスケール:
+            - 通常のアップスケール:
+                できるだけ絵を変更せずにアップスケールします
+            - クリエィティブなアップスケール:
+                絵全体をいい感じになるようにアップスケールします
+        - 画像の編集:
+            - 消去:
+                マスクした要素を削除します
+            - インペイント:
+                マスクされた部分とプロンプトを基に再生成します
+            - アウトペイント:
+                指定した範囲分、画像を拡張します
+            - 検索と置き換え:
+                検索プロンプトに入力されたオブジェクトを探し、プロンプトをもとにオブジェクトを再生成します
+        - コントロール:
+            - 構図:
+                入力画像の構図を基に新しい画像を生成します
+            - スケッチ:
+                ラフなスケッチとプロンプトを基に生成します
+        """
+
 model_url = {
     "ImageUltra": "https://api.stability.ai/v2beta/stable-image/generate/ultra",
     "ImageCore": "https://api.stability.ai/v2beta/stable-image/generate/core",
@@ -29,6 +84,90 @@ service_url = {
 }
 
 
+translations = {
+    "en": {
+        "api_key": "API Key",
+        "api_key_placeholder": "Enter your API key",
+        "model_label": "Model",
+        "mode_label": "Mode",
+        "prompt_placeholder": "Enter your prompt",
+        "negative_prompt_placeholder": "Enter a negative prompt",
+        "seed_label": "Seed",
+        "randomize_seed_label": "Randomize seed",
+        "aspect_label": "Aspect ratio",
+        "run_button": "Run",
+        "result_label": "Result",
+        "copy_field_placeholder": "Copy the field",
+        "Negative_prompt": "Negative prompt",
+        "Advanced_Settings": "Advanced Settings",
+        "Example": "Example",
+        "Generate": "Generate",
+        "Upscale": "Upscale",
+        "Edit": "Edit",
+        "Control": "Control",
+        "Submode": "Submode",
+        "Conservative": "Conservative",
+        "Creative": "Creative",
+        "Erase": "Erase",
+        "Inpaint": "Inpaint",
+        "Outpaint": "Outpaint",
+        "Structure": "Structure",
+        "Sketch": "Sketch",
+        "Search_and_Replace": "Search and Replace",
+        "Remove_Background": "Remove Background",
+        "input_image": "Input Image",
+        "style_preset": "Style preset",
+        "preset_description": "This parameter is only available for ImageCore model.",
+        "Search_prompt_placeholder": "Enter a search prompt",
+        "Control_Strength": "Control Strength",
+        "overview": overview,
+        "overview_label": "How to use",
+        "title": title,
+    },
+    "ja": {
+        "api_key": "APIキー",
+        "api_key_placeholder": "APIキーを入力してください",
+        "model_label": "モデル",
+        "mode_label": "モード",
+        "prompt_placeholder": "プロンプトを入力してください",
+        "negative_prompt_placeholder": "ネガティブプロンプトを入力してください",
+        "seed_label": "シード",
+        "randomize_seed_label": "シードをランダム化",
+        "aspect_label": "アスペクト比",
+        "run_button": "実行",
+        "result_label": "結果",
+        "copy_field_placeholder": "ここに貼り付け用の情報が出てきます",
+        "Negative_prompt": "ネガティブプロンプト",
+        "Advanced_Settings": "追加設定",
+        "Example": "例",
+        "Generate": "テキストから生成",
+        "Upscale": "アップスケール",
+        "Edit": "画像の編集",
+        "Control": "コントロールモード",
+        "Submode": "サブモード",
+        "Conservative": "通常のアップスケール",
+        "Creative": "クリエィティブなアップスケール",
+        "Erase": "消去",
+        "Inpaint": "インペイント",
+        "Outpaint": "アウトペイント(拡張)",
+        "Structure": "構図",
+        "Sketch": "スケッチ",
+        "Search_and_Replace": "検索と置き換え",
+        "Remove_Background": "背景削除",
+        "input_image": "入力画像",
+        "style_preset": "スタイルのプリセット",
+        "preset_description": "このパラメータはimage coreのときにだけ有効になります",
+        "Search_prompt_placeholder": "探したい要素を入力してください",
+        "Control_Strength": "コントロールネットの適用強度",
+        "overview": overview_jp,
+        "overview_label": "使い方",
+        "title": title_jp,
+    },
+}
+
+lang = "ja"
+
+
 def bytes_to_image(image):
     image = BytesIO(image)
     image = Image.open(image).convert("RGB")
@@ -83,12 +222,15 @@ def generate(
         "seed": seed,
         "aspect_ratio": aspect,
     }
+    data_rmbg = {
+        "output_format": "png",
+    }
     if input_image is not None:
         file["image"] = image_to_bytes(input_image)
         if mask is not None:
             file["mask"] = image_to_bytes(mask)
 
-    if mode == "Generate":
+    if mode == translations[lang]["Generate"]:
         file["none"] = ""
         if model == "Stable Image Ultra (8B + workflow)":
             url = model_url["ImageUltra"]
@@ -104,44 +246,46 @@ def generate(
         elif model == "Stable Diffusion 3 Large Turbo (8B Turbo)":
             url = model_url["StableDiffusion3"]
             data["model"] = "sd3-large-turbo"
-
         else:
             raise ValueError("Invalid model type")
 
-    elif mode == "Upscale":
-        if submode == "Conservative":
+    elif mode == translations[lang]["Upscale"]:
+        if submode == translations[lang]["Conservative"]:
             url = service_url["Conservative_Upscale"]
-        elif submode == "Creative":
+        elif submode == translations[lang]["Creative"]:
             url = service_url["Creative_Upscale"]
 
-    elif mode == "Edit":
-        if submode == "Erase":
+    elif mode == translations[lang]["Edit"]:
+        if submode == translations[lang]["Erase"]:
             url = service_url["Erase"]
-        elif submode == "Inpaint":
+        elif submode == translations[lang]["Inpaint"]:
             url = service_url["Inpaint"]
-        elif submode == "Outpaint":
+        elif submode == translations[lang]["Outpaint"]:
             url = service_url["Outpaint"]
             data["left"] = op_left
             data["right"] = op_right
             data["up"] = op_up
             data["down"] = op_down
-        elif submode == "Search and Replace":
+        elif submode == translations[lang]["Search_and_Replace"]:
             url = service_url["SR"]
             data["search_prompt"] = search_prompt
-        elif submode == "Remove Background":
+        elif submode == translations[lang]["Remove_Background"]:
+            data = data_rmbg
             url = service_url["RMBG"]
 
-    elif mode == "Control":
+    elif mode == translations[lang]["Control"]:
         data["control_strength"] = CNstrength
-        if submode == "Sketch":
+        if submode == translations[lang]["Sketch"]:
             url = service_url["Sketch"]
-        elif submode == "Structure":
+        elif submode == translations[lang]["Structure"]:
             url = service_url["Structure"]
-
     response = send_request(url, api_key, file, data)
-
+    
     if response.status_code == 200:
-        if mode == "Upscale" and submode == "Creative":
+        if (
+            mode == translations[lang]["Upscale"]
+            and submode == translations[lang]["Creative"]
+        ):
             generation_id = response.json().get("id")
             if not generation_id:
                 raise Exception("No generation ID returned for creative upscale")
@@ -200,31 +344,41 @@ def update_mode(mode):
     img_input_update = gr.update(visible=False)
     mask_update = gr.update(visible=False)
 
-    if mode == "Generate":
+    if mode == translations[lang]["Generate"]:
         submode_update = gr.update(visible=False)
-    elif mode == "Upscale":
+    elif mode == translations[lang]["Upscale"]:
         submode_update = gr.update(
-            choices=["Conservative", "Creative"], value="Conservative", visible=True
+            choices=[
+                translations[lang]["Conservative"],
+                translations[lang]["Creative"],
+            ],
+            value=translations[lang]["Conservative"],
+            visible=True,
         )
         img_input_update = gr.update(visible=True)
         image_label_update = gr.update(visible=True)
-    elif mode == "Edit":
+    elif mode == translations[lang]["Edit"]:
         submode_update = gr.update(
             choices=[
-                "Erase",
-                "Inpaint",
-                "Outpaint",
-                "Search and Replace",
-                "Remove Background",
+                translations[lang]["Erase"],
+                translations[lang]["Inpaint"],
+                translations[lang]["Outpaint"],
+                translations[lang]["Search_and_Replace"],
+                translations[lang]["Remove_Background"],
             ],
-            value="Erase",
+            value=translations[lang]["Erase"],
             visible=True,
         )
         img_input_update = gr.update(visible=True)
         image_label_update = gr.update(visible=True)
-    elif mode == "Control":
+    elif mode == translations[lang]["Control"]:
         submode_update = gr.update(
-            choices=["Sketch", "Structure"], value="Sketch", visible=True
+            choices=[
+                translations[lang]["Structure"],
+                translations[lang]["Sketch"],
+            ],
+            value=translations[lang]["Structure"],
+            visible=True,
         )
         img_input_update = gr.update(visible=True)
         image_label_update = gr.update(visible=True)
@@ -238,17 +392,17 @@ def update_submode(submode):
     cn = gr.update(visible=False)
     search_prompt = gr.update(visible=False)
 
-    if submode in ["Erase", "Inpaint"]:
+    if submode in [translations[lang]["Erase"], translations[lang]["Inpaint"]]:
         mask = gr.update(visible=True)
 
     else:
-        if submode == "Outpaint":
+        if submode == translations[lang]["Outpaint"]:
             outpaint = gr.update(visible=True)
 
-        elif submode == "Control":
+        elif submode in [translations[lang]["Structure"], translations[lang]["Sketch"]]:
             cn = gr.update(visible=True)
 
-        elif submode == "Search and Replace":
+        elif submode == translations[lang]["Search_and_Replace"]:
             search_prompt = gr.update(visible=True)
 
     return mask, outpaint, cn, search_prompt
@@ -257,44 +411,50 @@ def update_submode(submode):
 with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
     with gr.Column(elem_id="col-container"):
         gr.Markdown(
-        f"""
-        # Stability AI - Developer Platform WebUI
-        **Overview**  
-        Stability AI’s Stable Image services offer a growing set of APIs for developers to build the best in class image applications.
-
-        - **Disrupting Content Creation:** Stability’s Image APIs are the foundation for applications disrupting publishing, media, gaming, marketing, advertising, design, and more.
-        - **For Developers:** Application developers can build advanced features for designers, photographers, content creators, and a variety of B2C customers.
-        - **Simple APIs:** Stability AI is focused on delivering simple APIs for easy integration into applications with a high bar for quality, alignment, speed, and safety.
-
-        Get Started Now: https://platform.stability.ai/docs/getting-started/stable-image    
-
-        Contact: D̷ELL@Stability AI - Advocate (https://x.com/xqdior) / Author: umise (https://x.com/UiE029)
-        """
+            translations[lang]["title"],
         )
+        with gr.Accordion(translations[lang]["overview_label"], open=False):
+            gr.Markdown(
+                translations[lang]["overview"],
+            )
 
         with gr.Row():
             api_key = gr.Text(
-                label="API Key",
+                label=translations[lang]["api_key"],
                 type="password",
-                placeholder="Enter your API key",
+                placeholder=translations[lang]["api_key_placeholder"],
                 max_lines=1,
                 container=False,
             )
 
         with gr.Row():
             model = gr.Dropdown(
-                label="Model",
-                choices=["Stable Image Ultra (8B + workflow)", "Stable Image Core (2B + workflow)", "Stable Diffusion 3 Large Turbo (8B Turbo)", "Stable Diffusion 3 Large (8B)", "Stable Diffusion 3 Medium (2B)"],
-                value="Ultra",
+                label=translations[lang]["model_label"],
+                choices=[
+                    "Stable Image Ultra (8B + workflow)",
+                    "Stable Image Core (2B + workflow)",
+                    "Stable Diffusion 3 Large Turbo (8B Turbo)",
+                    "Stable Diffusion 3 Large (8B)",
+                    "Stable Diffusion 3 Medium (2B)",
+                ],
+                value="Stable Image Ultra (8B + workflow)",
             )
             mode = gr.Dropdown(
-                label="Mode",
-                choices=["Generate", "Upscale", "Edit", "Control"],
-                value="Generate",
+                label=translations[lang]["mode_label"],
+                choices=[
+                    translations[lang]["Generate"],
+                    translations[lang]["Upscale"],
+                    translations[lang]["Edit"],
+                    translations[lang]["Control"],
+                ],
+                value=translations[lang]["Generate"],
             )
 
         submode = gr.Dropdown(
-            label="Submode", choices=["None"], visible=False, value="None"
+            label=translations[lang]["Submode"],
+            choices=["None"],
+            visible=False,
+            value="None",
         )
 
         with gr.Row():
@@ -303,7 +463,7 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                     label="Prompt",
                     show_label=False,
                     max_lines=1,
-                    placeholder="Enter your prompt",
+                    placeholder=translations[lang]["prompt_placeholder"],
                     container=False,
                 )
                 search_prompt = gr.Text(
@@ -311,14 +471,18 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                     visible=False,
                     show_label=False,
                     max_lines=1,
-                    placeholder="Enter a search prompt",
+                    placeholder=translations[lang]["Search_prompt_placeholder"],
                 )
-
-            run_button = gr.Button("Run", scale=0)
-
+            run_button = gr.Button(translations[lang]["run_button"], scale=0)
+        with gr.Row():
+            gr.Examples(
+                label=translations[lang]["Example"], examples=examples, inputs=[prompt]
+            )
         with gr.Row():
             with gr.Column():
-                image_label = gr.Markdown(value="input image", visible=False)
+                image_label = gr.Markdown(
+                    value=translations[lang]["input_image"], visible=False
+                )
                 image = gr.Image(
                     type="pil",
                     label="img input",
@@ -342,30 +506,34 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                 )
 
         with gr.Row():
-            result = gr.Image(label="Result", width="20vw", height="20%")
+            result = gr.Image(
+                label=translations[lang]["result_label"], width="20vw", height="20%"
+            )
 
-        with gr.Accordion("Advanced Settings", open=False):
+        with gr.Accordion(translations[lang]["Advanced_Settings"], open=False):
             negative_prompt = gr.Text(
-                label="Negative prompt",
+                label=translations[lang]["Negative_prompt"],
                 max_lines=1,
-                placeholder="Enter a negative prompt",
+                placeholder=translations[lang]["negative_prompt_placeholder"],
             )
             seed = gr.Slider(
-                label="Seed",
+                label=translations[lang]["seed_label"],
                 minimum=0,
                 maximum=MAX_SEED,
                 step=1,
                 value=0,
             )
             CN_strength = gr.Slider(
-                label="Control Strength",
+                label=translations[lang]["Control_Strength"],
                 minimum=0,
                 maximum=1,
                 step=0.01,
                 value=0.5,
                 visible=False,
             )
-            randomize_seed = gr.Checkbox(label="Randomize seed", value=True)
+            randomize_seed = gr.Checkbox(
+                label=translations[lang]["randomize_seed_label"], value=True
+            )
             aspect = gr.Radio(
                 choices=[
                     "1:1",
@@ -378,7 +546,7 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                     "9:16",
                     "9:21",
                 ],
-                label="Aspect raito",
+                label=translations[lang]["aspect_label"],
                 value="1:1",
             )
             with gr.Row(visible=False) as style:
@@ -402,9 +570,9 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                         "pixel-art",
                         "tile-texture",
                     ],
-                    label="Style_preset",
+                    label=translations[lang]["style_preset"],
                     value="anime",
-                    info="This parameter is only available for ImageCore model.",
+                    info=translations[lang]["preset_description"],
                 )
             with gr.Row(visible=False) as outpaint_scale:
                 paint = gr.Markdown(value="Outpain Scale")
@@ -421,18 +589,16 @@ with gr.Blocks(css=css, theme="NoCrypt/miku") as demo:
                     label="down", minimum=0, maximum=2000, step=4, value=200
                 )
 
-        gr.Examples(examples=examples, inputs=[prompt])
-
         copy_filed = gr.TextArea(
             value="",
             label="Copy Field",
             max_lines=1,
-            placeholder="Copy the field",
+            placeholder=translations[lang]["copy_field_placeholder"],
             show_copy_button=True,
             container=False,
         )
         gr.Markdown(
-        f"""
+            f"""
         ## License
         This work is licensed under a
         [Creative Commons Attribution-NonCommercial 4.0 International License][cc-by-nc].