Update src/text_to_speech.py

src/text_to_speech.py

@@ -14,74 +14,37 @@ def generate_audio(text, filename="output.mp3", gender="female", speed="normal")
         gender (str): "male" (use MMS-TTS) or "female" (use gTTS).
         speed (str): "slow", "normal", or "fast" (only for gTTS).
     """
-    print("\n[DEBUG] Function: generate_audio")
-    print(f"Received parameters -> text: {text[:50]}..., filename: {filename}, gender: {gender}, speed: {speed}")
-    
     lang = "vi"
     
     if gender.lower() == "female":
-        print("[DEBUG] Using gTTS for female voice...")
+        # gTTS chỉ có giọng nữ
         speed_mapping = {"slow": True, "normal": False, "fast": False}
         slow = speed_mapping.get(speed.lower(), False)
         
-        try:
-            tts = gTTS(text=text, lang=lang, slow=slow)
-            tts.save(filename)
-            print(f"✅ Audio saved as {filename}")
-        except Exception as e:
-            print(f"[ERROR] Failed to generate audio: {e}")
-    else:
-        print("[ERROR] Male voice generation not implemented yet!")
-
-    # Debug check for file existence
-    if os.path.exists(filename):
-        print(f"✅ Verified: {filename} exists.")
+        tts = gTTS(text=text, lang=lang, slow=slow)
+        tts.save(filename)
+        print(f"✅ Audio saved as {filename}")
+    
+    elif gender.lower() == "male":
+        # MMS-TTS cho giọng nam
+        model = VitsModel.from_pretrained("facebook/mms-tts-vie")
+        tokenizer = AutoTokenizer.from_pretrained("facebook/mms-tts-vie")
+        
+        inputs = tokenizer(text, return_tensors="pt")
+        with torch.no_grad():
+            output = model(**inputs).waveform
+        
+        # Lưu file âm thanh
+        torchaudio.save(filename, output, 24000, backend="sox_io")
+        print(f"✅ Audio saved as {filename}")
+    
     else:
-        print(f"[ERROR] {filename} was not created.")
-import os
-
+        print("⚠️ Giọng không hợp lệ! Chỉ hỗ trợ 'male' hoặc 'female'.")
 def text_to_speech(gender, speed):
-    """
-    Convert text files in the folder to speech and save as audio files.
-    
-    Parameters:
-        gender (str): "male" (use MMS-TTS) or "female" (use gTTS).
-        speed (str): "slow", "normal", or "fast" (only for gTTS).
-    """
-    print("\n[DEBUG] Function: text_to_speech")
-    print(f"Received parameters -> gender: {gender}, speed: {speed}")
-
     text_folder = "./"
-    print(f"[DEBUG] Looking for text files in: {os.path.abspath(text_folder)}")
-
-    # Get all valid text files (excluding specific ones)
-    text_files = sorted([
-        f for f in os.listdir(text_folder) 
-        if f.endswith('.txt') and f not in ["text.txt", "requirements.txt"]
-    ])
-
-    print(f"[DEBUG] Found text files: {text_files}")
-
-    if not text_files:
-        print("[WARNING] No text files found for conversion.")
-
+    text_files = sorted([f for f in os.listdir(text_folder) if f.endswith('.txt') and f != "text.txt" and f != "requirements.txt"])
     for text_file in text_files:
-        file_path = os.path.join(text_folder, text_file)
-        try:
-            print(f"[DEBUG] Reading file: {file_path}")
-            with open(file_path, "r", encoding="utf-8") as file:
-                content = file.read()
-            
-            audio_file = text_file.replace("txt", "mp3")
-            print(f"[DEBUG] Generating audio for: {text_file} -> {audio_file}")
-
-            generate_audio(content, audio_file, gender=gender, speed=speed)
-
-            # Verify if audio was created
-            if os.path.exists(audio_file):
-                print(f"✅ Audio file created: {audio_file}")
-            else:
-                print(f"[ERROR] Audio file {audio_file} was not created!")
-
-        except Exception as e:
-            print(f"[ERROR] Failed to process {text_file}: {e}")
+        with open(f"{text_file}", "r", encoding="utf-8") as file:
+            content = file.read()
+        audio_file = text_file.replace("txt","mp3")
+        generate_audio(content, f"{audio_file}", gender=gender, speed=speed)