app.py

import gradio as gr

from PyPDF2 import PdfReader
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
from gtts import gTTS
from io import BytesIO

# Define model and tokenizer
model_name = "ArtifactAI/led_large_16384_arxiv_summarization"
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
tokenizer = AutoTokenizer.from_pretrained(model_name)

def summarize_pdf_abstract(pdf_data):
    """
    Reads a PDF file, extracts the abstract, and summarizes it in one sentence.

    Args:
        pdf_data: A byte string containing the PDF data.

    Returns:
        A dictionary containing the one-sentence summary of the abstract and the generated audio.
    """

    reader = PdfReader(BytesIO(pdf_data))
    abstract_text = ""

    for page in reader.pages:
        if "Abstract" in page.extract_text() or "Introduction" in page.extract_text():
            abstract_text = page.extract_text()
            break

    inputs = tokenizer(abstract_text, return_tensors="pt")
    outputs = model.generate(**inputs)
    summary = tokenizer.decode(outputs[0], skip_special_tokens=True)

    speech = gTTS(summary, lang="en")
    speech_bytes = speech.get_wav_data()

    return {"summary": summary, "audio": speech_bytes}

# Create Gradio interface
interface = gr.Interface(
    fn=summarize_pdf_abstract,
    inputs=[gr.File(label="Upload PDF", mimetypes=["application/pdf"])],
    outputs=[gr.Text(label="One-sentence summary"), gr.Audio(label="Summary audio")],
)

# Launch the Hugging Face Space
interface.launch(title="PDF Abstract Summarizer")