Update voice_of_the_doctor.py

voice_of_the_doctor.py

@@ -1,35 +1,13 @@
-import os
-from gtts import gTTS
-from elevenlabs.client import ElevenLabs
-
-# Load ElevenLabs API key from environment
-ELEVENLABS_API_KEY = os.environ.get("ELEVENLABS_API_KEY")
-
-
-def text_to_speech_with_gtts(input_text, output_filepath="doctor_voice.mp3"):
-    """Generate voice using gTTS (free) and save as MP3"""
-    tts = gTTS(text=input_text, lang="en", slow=False)
-    tts.save(output_filepath)
-    return output_filepath  # Return file for frontend to play
-
-
-def text_to_speech_with_elevenlabs(input_text, output_filepath="doctor_voice.mp3"):
-    """Generate voice using ElevenLabs (paid API)"""
-    if not ELEVENLABS_API_KEY:
-        raise ValueError("ELEVENLABS_API_KEY not set in environment variables")
-
-    client = ElevenLabs(api_key=ELEVENLABS_API_KEY)
-
-    # This returns a generator of audio chunks
-    audio_stream = client.text_to_speech.convert(
-        voice_id="EXAVITQu4vr4xnSDxMaL",  # Example voice_id (replace with yours)
-        model_id="eleven_turbo_v2",
-        text=input_text
-    )
-
-    # Collect chunks into a file
-    with open(output_filepath, "wb") as f:
-        for chunk in audio_stream:
-            f.write(chunk)
-
+import soundfile as sf
+from transformers import pipeline
+
+# Load Hugging Face TTS pipeline
+pipe = pipeline("text-to-speech", model="maya-research/Veena")
+
+def text_to_speech_with_veena(input_text, output_filepath="doctor_voice.wav"):
+    """Generate speech using Hugging Face Veena TTS model"""
+    output = pipe(input_text)
+    audio_array = output["audio"]
+    sampling_rate = output["sampling_rate"]
+    sf.write(output_filepath, audio_array, sampling_rate)
     return output_filepath