Update app.py

app.py

@@ -1,10 +1,10 @@
 import gradio as gr
 import edge_tts
-import asyncio
 import tempfile
 import os
+import asyncio
 
-# Map readable labels to Edge TTS internal IDs
+# خريطة الأصوات
 VOICE_MAP = {
     "رجل (مصري)": "ar-EG-ShakirNeural",
     "سيدة (مصرية)": "ar-EG-SalmaNeural",
@@ -14,76 +14,68 @@ VOICE_MAP = {
     "English (US) F": "en-US-AriaNeural"
 }
 
-async def text_to_speech_edge(text, voice, emotion, is_symbol, rate, pitch):
-    """
-    Generates speech using Edge TTS.
-    Args:
-        text (str): The text to speak.
-        voice (str): Voice ID or Label.
-        emotion (str): Ignored (kept for compatibility).
-        is_symbol (bool): Ignored (kept for compatibility).
-        rate (str): Speed change (e.g., "+0%", "+20%").
-        pitch (str): Pitch change (e.g., "+0Hz", "+2st").
-    """
-    
-    # 1. Validation
+async def generate_speech(text, voice, emotion, is_symbol, rate, pitch):
+    # 1. التحقق من المدخلات الأساسية
     if not text or not text.strip():
         return None
-
-    # 2. Voice Resolution
-    # Check if the input is a known Key (Label) or Value (ID)
-    selected_voice = "ar-SA-HamedNeural" # Default
+    
+    # 2. معالجة القيم الافتراضية لتجنب الأخطاء
+    # إذا وصلت القيم فارغة من الواجهة الأمامية، نستخدم القيم الافتراضية
+    final_rate = rate if rate and isinstance(rate, str) and len(rate.strip()) > 0 else "+0%"
+    final_pitch = pitch if pitch and isinstance(pitch, str) and len(pitch.strip()) > 0 else "+0Hz"
+    
+    # 3. تحديد هوية الصوت
+    # نحاول البحث في القاموس، إذا لم نجد الاسم، نستخدم القيمة كما هي (على افتراض أنها ID)
+    selected_voice = "ar-SA-HamedNeural" # صوت افتراضي
     if voice in VOICE_MAP:
         selected_voice = VOICE_MAP[voice]
     elif voice in VOICE_MAP.values():
         selected_voice = voice
     
-    # 3. Parameter Sanitization
-    # EdgeTTS crashes if rate/pitch are None or empty strings.
-    # We ensure they always have a valid default value.
-    final_rate = rate if rate and isinstance(rate, str) and len(rate.strip()) > 0 else "+0%"
-    final_pitch = pitch if pitch and isinstance(pitch, str) and len(pitch.strip()) > 0 else "+0Hz"
-    
-    print(f"Generating: TextLen={len(text)}, Voice={selected_voice}, Rate={final_rate}, Pitch={final_pitch}")
-
-    # 4. Temp File Creation
-    # delete=False is required for Gradio to serve the file after the function returns
-    output_file = tempfile.NamedTemporaryFile(suffix=".mp3", delete=False)
-    output_path = output_file.name
-    output_file.close()
+    print(f"Processing: TextLen={len(text)}, Voice={selected_voice}, Rate={final_rate}, Pitch={final_pitch}")
 
     try:
-        # 5. Generation
+        # 4. إنشاء ملف مؤقت
+        output_file = tempfile.NamedTemporaryFile(suffix=".mp3", delete=False)
+        output_path = output_file.name
+        output_file.close()
+        
+        # 5. التوليد باستخدام Edge TTS
         communicate = edge_tts.Communicate(text, selected_voice, rate=final_rate, pitch=final_pitch)
         await communicate.save(output_path)
-        return output_path
         
-    except Exception as e:
-        print(f"CRITICAL ERROR: {str(e)}")
-        # Return None or raise a Gradio Error to notify the frontend
-        raise gr.Error(f"Generation Failed: {str(e)}")
-
-# Define the Interface
-# The order of inputs MUST match the array sent from React:
-# [text, voice, emotion, is_symbol, rate, pitch]
-inputs = [
-    gr.Textbox(label="Text"),
-    gr.Dropdown(label="Voice", choices=list(VOICE_MAP.keys()) + list(VOICE_MAP.values())),
-    gr.Textbox(label="Emotion", value="neutral"), 
-    gr.Checkbox(label="Is Symbol", value=True),   
-    gr.Textbox(label="Rate", value="+0%"),
-    gr.Textbox(label="Pitch", value="+0Hz")
-]
+        return output_path
 
-outputs = gr.Audio(label="Generated Audio", type="filepath")
+    except Exception as e:
+        print(f"Error generating speech: {str(e)}")
+        # في حال حدوث خطأ، نرجح None ليتم التعامل معه في الواجهة
+        raise gr.Error(f"TTS Error: {str(e)}")
 
-demo = gr.Interface(
-    fn=text_to_speech_edge,
-    inputs=inputs,
-    outputs=outputs,
-    title="Natiq Pro API",
-    allow_flagging="never"
-)
+# استخدام Blocks بدلاً من Interface لتفادي مشاكل النسخ الجديدة
+with gr.Blocks(title="Natiq Pro API") as demo:
+    gr.Markdown("# Natiq Pro API Endpoints")
+    
+    # تعريف المدخلات (يجب أن تطابق الترتيب المرسل من React)
+    with gr.Row(visible=False): # نجعلها مخفية لأننا نستخدمها كـ API فقط
+        input_text = gr.Textbox(label="Text")
+        input_voice = gr.Textbox(label="Voice")
+        input_emotion = gr.Textbox(label="Emotion", value="neutral")
+        input_symbol = gr.Checkbox(label="Is Symbol", value=True)
+        input_rate = gr.Textbox(label="Rate", value="+0%")
+        input_pitch = gr.Textbox(label="Pitch", value="+0Hz")
+        
+    output_audio = gr.Audio(label="Generated Audio", type="filepath")
+    
+    btn = gr.Button("Generate", visible=False)
+    
+    # ربط الدالة بالزر وتعريف اسم الـ API بشكل صريح
+    # api_name="text_to_speech_edge" هو المفتاح الذي يبحث عنه تطبيق React
+    btn.click(
+        fn=generate_speech,
+        inputs=[input_text, input_voice, input_emotion, input_symbol, input_rate, input_pitch],
+        outputs=[output_audio],
+        api_name="text_to_speech_edge"
+    )
 
 if __name__ == "__main__":
     demo.queue().launch()