app.py

import gradio as gr
from gtts import gTTS
import io
import os
import tempfile

def text_to_speech(text, language, pitch):
    tts = gTTS(text=text, lang=language, slow=False)
    
    # Save to a temporary file
    with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as fp:
        tts.save(fp.name)
        
    # Apply pitch shift using ffmpeg
    output_file = tempfile.NamedTemporaryFile(delete=False, suffix=".mp3").name
    pitch_shift = pitch - 1  # Adjust pitch shift value
    os.system(f"ffmpeg -i {fp.name} -af asetrate=44100*{2**pitch_shift},aresample=44100 {output_file}")
    
    # Clean up the original file
    os.unlink(fp.name)
    
    return output_file

def gradio_tts_interface(text, language, pitch):
    audio_file = text_to_speech(text, language, pitch)
    return audio_file

iface = gr.Blocks(theme="Hev832/Applio")

with iface:
    gr.Markdown("# Text-to-Speech Demo")
    
    with gr.Row():
        with gr.Column():
            text_input = gr.Textbox(label="Enter text to convert to speech", lines=3)
            language_input = gr.Dropdown(["en", "fr", "es", "de", "it", "id", "ja"], label="Select Language")
            pitch_input = gr.Slider(minimum=0.5, maximum=2.0, value=1.0, step=0.1, label="Pitch (0.5 for lower/male, 1.0 for normal, 2.0 for higher/female)")
            submit_button = gr.Button("Convert to Speech")
        
        with gr.Column():
            audio_output = gr.Audio(label="Generated Speech")
    
    submit_button.click(
        fn=gradio_tts_interface,
        inputs=[text_input, language_input, pitch_input],
        outputs=audio_output
    )

iface.launch()