app.py

import gradio as gr
from transformers import pipeline
from PyPDF2 import PdfReader
from ebooklib import epub
from bs4 import BeautifulSoup

# Load the VITS model from Hugging Face
text_to_speech = pipeline("text-to-speech", model="efficient-speech/lite-whisper-large-v3-turbo")

# Function to extract text from PDF
def extract_pdf_text(file):
    reader = PdfReader(file)
    text = ""
    for page in reader.pages:
        if page and page.extract_text():
            text += page.extract_text()
    return text

# Function to extract text from EPUB
def extract_epub_text(file):
    book = epub.read_epub(file)
    text = ""
    for item in book.get_items():
        if item.get_type() == epub.ITEM_DOCUMENT:
            soup = BeautifulSoup(item.content, 'html.parser')
            text += soup.get_text()
    return text

# Unified function to convert text to speech
def convert_to_audio(file, file_type):
    if file_type == 'PDF':
        text = extract_pdf_text(file)
    elif file_type == 'EPUB':
        text = extract_epub_text(file)
    else:
        text = file.read().decode('utf-8')

    if not text.strip():
        return "No text found in the file."

    # Convert text to speech
    audio = text_to_speech(text[:5000])  # Limiting input to avoid model constraints
    return (audio["audio"],)

# Gradio interface
demo = gr.Interface(
    fn=convert_to_audio,
    inputs=[
        gr.File(label="Upload PDF, EPUB, or Text File"),
        gr.Radio(["PDF", "EPUB", "TXT"], label="File Type")
    ],
    outputs="audio",
    title="Unlimited Text-to-Speech Converter",
    description="Upload PDF, EPUB, or text files — convert them into audio with no limits!"
)

demo.launch()