app.py

import gradio as gr
from gtts import gTTS
import io
import os
import librosa
import soundfile as sf

def text_to_speech(text, language_accent, pitch):
    lang, tld = language_tld_map[language_accent].split(',')

    # create the text-to-speech audio
    tts = gTTS(text, lang=lang, tld=tld)
    tts.save('gtts.wav')

    # Load the audio file
    y, sr = librosa.load('gtts.wav')

    # Specify the number of semitones to shift
    new_y = librosa.effects.pitch_shift(y=y, sr=sr, n_steps=pitch)

    sf.write("generated_gtts.wav", new_y, sr)

    # return the generated audio
    return 'generated_gtts.wav'


language_tld_map = {
    "English_Australia": "en,com.au",
    "English_United_Kingdom": "en,co.uk",
    "English_United_States": "en,com",
    "English_Canada": "en,ca",
    "English_Nigerian": "en,com.ng",
    "English_Ireland": "en,ie",
    "English_South Africa": "en,co.za",
    "French_Canada": "fr,ca",
    "French_France": "fr,fr",
    "Mandarin_China_Mainland": "zh-CN,com",
    "Mandarin_Taiwan": "zh-TW,com",
    "Portuguese_Brazil": "pt,com.br",
    "Portuguese_Portugal": "pt,pt",
    "Spanish_Mexico": "es,com.mx",
    "Spanish_Spain": "es,es",
    "Spanish_United_States": "es,com"
}

# create the Gradio interface
iface = gr.Interface(fn=text_to_speech, 
                     inputs=[gr.Textbox(lines=10, label="Enter your text here:"),
                             gr.Dropdown(choices=list(language_tld_map.keys()), label="Select Language & Accent:", value="English_United_Kingdom", type="value"), 
                             gr.Number(label="Pitch (0 = no variations, negative pitch makes it more masculine, + pitch makes it more feminine):", value=0)],
                     outputs=[gr.Audio(label="Audio")])

iface.launch()