app.py

import spaces
import PyPDF2 
from docx import Document
import os
from unittest.mock import patch
from TTS.api import TTS
import torch
import gradio as gr
from torchaudio import load

print("completed imported packages")
os.environ["COQUI_TOS_AGREED"] = "1"


# Function to always return 'y'
def always_yes(*args, **kwargs):
    return 'y'

# Patch the input function to always return 'y'
with patch('builtins.input', always_yes):
    device = "cuda:0" if torch.cuda.is_available() else "cpu"
    tts = TTS(model_name="tts_models/multilingual/multi-dataset/xtts_v2", progress_bar=False).to(device)



print("loded the model")
sa = 'sampleaudio.wav'

@spaces.GPU(enable_queue=True)

def text_to_speech(text):
    save_path = "outputz.wav"
    try:
        tts.tts_to_file(
            text,
            speaker_wav=sa,
            language="en",
            file_path=save_path,
            split_sentences=True,
        )
        return save_path
    except Exception as e:
        return str(e)
        
print("first function")

def extract_text_from_pdf(file_path):
    with open(file_path, 'rb') as file:
        reader = PyPDF2.PdfReader(file)
        text = "".join([page.extract_text() for page in reader.pages])
    return text

def extract_text_from_doc(file_path):
    doc = Document(file_path)
    return "\n".join([para.text for para in doc.paragraphs])

def file_to_audio(file_path):
    file_type = file_path.split('.')[-1]
    if file_type == 'pdf':
        text = extract_text_from_pdf(file_path)
    elif file_type in ['doc', 'docx']:
        text = extract_text_from_doc(file_path)
    else:
        return "Unsupported file format"

    return text_to_speech(text)
def handle_inputs(text, file):
    if text and file:
        return None, "Please provide only one input at a time: text or file."
    elif text:
        audio_path = text_to_speech(text)
        return audio_path, "Thanks for your text input"
    elif file:
        audio_path = file_to_audio(file)
        return audio_path, 'Thanks for your File'
    else:
        return None, "No input provided." 


print("last  function")



with gr.Blocks() as demo:
    gr.Markdown("## Welcome to Text to Speech 📝 ➡️ 🎧")

    with gr.Row():
        text_input = gr.Textbox(label="Text to Speak", placeholder="Enter text...")
        file_input = gr.File(type="filepath", label="Upload a PDF or Word file", file_count="single")

    submit_btn = gr.Button("Generate Voice")

    output_audio = gr.Audio(label="Audio Output", type="filepath")
    output_message = gr.Textbox(label="Message")

    submit_btn.click(fn=handle_inputs, inputs=[text_input, file_input], outputs=[output_audio, output_message])

# Launch the Gradio app
demo.launch()