#https://huggingface.co/spaces/gianb/PDF_Summarized_TTS

# Here are the imports

import gradio as gr
import PyPDF2
import pdfplumber
from transformers import pipeline, AutoProcessor, AutoModel, AutoTokenizer
from PyPDF2 import PdfReader
import torch
import soundfile as sf
from IPython.display import Audio
from datasets import load_dataset
from pdfminer.high_level import extract_pages, extract_text
import io

#Here is the code

summarization = pipeline ('summarization', model = "pszemraj/long-t5-tglobal-base-16384-book-summary")

def summarize_and_speech(pdf_file):
        pdf_bytes_io = io.BytesIO(pdf_file)
        pdf_reader = PyPDF2.PdfReader(pdf_bytes_io)
        abstract_text = pdf_reader.pages[0].extract_text()
        summary = summarization(abstract_text, max_length=13, min_length=10)[0]['summary_text']
        
        # Use a text-to-speech model to generate audio
        synthesiser = pipeline("text-to-speech", "facebook/mms-tts-eng")
        tts_output = synthesiser(summary)
        audio_data = tts_output[0]["audio"]


        return summary, audio_data

iface = gr.Interface(
    fn= summarize_and_speech,
    inputs=gr.File(label="Upload PDF", type="binary"),
    outputs=[gr.Textbox(label="Abstract Summary:"), gr.Audio(type="filepath", label="Summary_Speech")],
    live=True,
    title="Abstract_Research_Paper_Summarizer",
    description="Upload a Research Paper PDF File. The model will generate a one line summary of the Abstract section and a speech audio."
)

iface.launch()