Update app.py

app.py

@@ -6,6 +6,8 @@ import tempfile
 import datetime
 import shutil
 import re
+from tqdm import tqdm
+import time
 
 # --- Coqui TTS 授權同意 ---
 os.environ["COQUI_TOS_AGREED"] = "1"
@@ -25,11 +27,9 @@ except Exception as e:
     print(f"警告：無法將安全全局變數加入 PyTorch 白名單: {e}")
     print("如果遇到模型載入錯誤，請檢查 PyTorch 和 TTS 庫版本。")
 
-# --- 設備選擇 ---
 device = "cuda" if torch.cuda.is_available() else "cpu"
 print(f"使用設備: {device}")
 
-# --- 全局變數 ---
 tts = None
 model_load_error = None
 
@@ -60,62 +60,78 @@ try:
 except Exception as e:
     model_load_error = f"載入 Coqui TTS XTTS-v2 模型時發生錯誤: {e}"
 
-def generate_speech(text, language, uploaded_speaker_audio_path, progress=gr.Progress(track_tqdm=True)):
+def generate_speech(text, language, uploaded_speaker_audio_path):
     if model_load_error:
         return None, f"應用程式啟動錯誤：{model_load_error}"
 
-    progress(0.05, desc="檢查模型狀態")
-    if tts is None:
-        return None, "TTS 模型未成功載入，無法生成語音。"
-
-    progress(0.1, desc="檢查輸入")
-    if not text:
-        return None, "請輸入一些文字！"
-    if not language:
-        return None, "請選擇一個語言！"
-
-    speaker_wav_to_use = None
-    status_message = ""
-
-    progress(0.2, desc="處理語音參考檔案")
-    if uploaded_speaker_audio_path:
-        speaker_wav_to_use = uploaded_speaker_audio_path
-        try:
-            timestamp_ref = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-            original_ext = os.path.splitext(uploaded_speaker_audio_path)[1]
-            saved_ref_file_name = f"{timestamp_ref}_uploaded_ref{original_ext}"
-            saved_ref_file_path = os.path.join(SAVE_UPLOADED_REFERENCES_DIR, saved_ref_file_name)
-            shutil.copy(uploaded_speaker_audio_path, saved_ref_file_path)
-            status_message += f"參考語音已儲存到：{saved_ref_file_path}\n"
-        except Exception as e:
-            status_message += f"警告：儲存參考語音失敗: {e}\n"
-    else:
-        speaker_wav_to_use = DEFAULT_SPEAKER_WAV
-        if not os.path.exists(speaker_wav_to_use):
-            return None, f"錯誤：預設語音參考檔案 ({DEFAULT_SPEAKER_WAV}) 未找到。請上傳一個檔案或確保預設檔案存在。"
-
-    output_file = None
-    try:
-        with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
-            output_file = fp.name
-
-        progress(0.5, desc="生成語音")
-        tts.tts_to_file(text=text, language=language, speaker_wav=speaker_wav_to_use, file_path=output_file)
-
-        progress(0.8, desc="儲存語音檔案")
-        timestamp_gen = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
-        sanitized_text = sanitize_filename(text)
-        saved_file_name = f"{timestamp_gen}_{language}_{sanitized_text}.wav"
-        saved_file_path = os.path.join(SAVE_GENERATED_AUDIO_DIR, saved_file_name)
-        shutil.copy(output_file, saved_file_path)
-        status_message += f"語音生成成功！已儲存為：{saved_file_path}"
-
-        progress(1.0, desc="完成")
-        return output_file, status_message
-    except Exception as e:
-        if output_file and os.path.exists(output_file):
-            os.remove(output_file)
-        return None, f"生成語音失敗: {e}"
+    steps = [
+        "檢查模型狀態",
+        "檢查輸入",
+        "處理語音參考檔案",
+        "生成語音",
+        "儲存語音檔案",
+        "完成"
+    ]
+
+    # tqdm 進度條會出現在 logs/終端機
+    for i, step in enumerate(tqdm(steps, desc="語音生成流程", ncols=80)):
+        if step == "檢查模型狀態":
+            if tts is None:
+                return None, "TTS 模型未成功載入，無法生成語音。"
+            time.sleep(0.1)
+        elif step == "檢查輸入":
+            if not text:
+                return None, "請輸入一些文字！"
+            if not language:
+                return None, "請選擇一個語言！"
+            time.sleep(0.1)
+        elif step == "處理語音參考檔案":
+            global speaker_wav_to_use
+            speaker_wav_to_use = None
+            global status_message
+            status_message = ""
+            if uploaded_speaker_audio_path:
+                speaker_wav_to_use = uploaded_speaker_audio_path
+                try:
+                    timestamp_ref = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
+                    original_ext = os.path.splitext(uploaded_speaker_audio_path)[1]
+                    saved_ref_file_name = f"{timestamp_ref}_uploaded_ref{original_ext}"
+                    saved_ref_file_path = os.path.join(SAVE_UPLOADED_REFERENCES_DIR, saved_ref_file_name)
+                    shutil.copy(uploaded_speaker_audio_path, saved_ref_file_path)
+                    status_message += f"參考語音已儲存到：{saved_ref_file_path}\n"
+                except Exception as e:
+                    status_message += f"警告：儲存參考語音失敗: {e}\n"
+            else:
+                speaker_wav_to_use = DEFAULT_SPEAKER_WAV
+                if not os.path.exists(speaker_wav_to_use):
+                    return None, f"錯誤：預設語音參考檔案 ({DEFAULT_SPEAKER_WAV}) 未找到。請上傳一個檔案或確保預設檔案存在。"
+            time.sleep(0.1)
+        elif step == "生成語音":
+            global output_file
+            output_file = None
+            try:
+                with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as fp:
+                    output_file = fp.name
+                tts.tts_to_file(text=text, language=language, speaker_wav=speaker_wav_to_use, file_path=output_file)
+            except Exception as e:
+                if output_file and os.path.exists(output_file):
+                    os.remove(output_file)
+                return None, f"生成語音失敗: {e}"
+        elif step == "儲存語音檔案":
+            try:
+                timestamp_gen = datetime.datetime.now().strftime("%Y%m%d_%H%M%S")
+                sanitized_text = sanitize_filename(text)
+                saved_file_name = f"{timestamp_gen}_{language}_{sanitized_text}.wav"
+                saved_file_path = os.path.join(SAVE_GENERATED_AUDIO_DIR, saved_file_name)
+                shutil.copy(output_file, saved_file_path)
+                status_message += f"語音生成成功！已儲存為：{saved_file_path}"
+            except Exception as e:
+                return None, f"儲存語音檔案失敗: {e}"
+        elif step == "完成":
+            pass
+        # 每步都可加 sleep 模擬耗時（真實環境下可移除）
+        time.sleep(0.1)
+    return output_file, status_message
 
 def list_saved_audio_files():
     audio_files = []
@@ -183,6 +199,4 @@ with gr.Blocks(title="Coqui TTS XTTS-v2 語音生成") as demo:
         demo.load(list_uploaded_reference_files, outputs=[saved_uploaded_ref_files_output])
         refresh_uploaded_ref_button.click(list_uploaded_reference_files, outputs=[saved_uploaded_ref_files_output])
 
-demo.queue()  # 啟用進度條功能
-if __name__ == "__main__":
-    demo.launch()
+demo.launch()