app.py

# app.py
import gradio as gr
import pdfplumber
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

# Load the model
model_name = "ibm-granite/granite-docling-258m-demo"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

# Function to extract text from PDF
def extract_text_from_pdf(pdf_file):
    text = ""
    with pdfplumber.open(pdf_file.name) as pdf:
        for page in pdf.pages:
            page_text = page.extract_text()
            if page_text:
                text += page_text + "\n"
    return text

# Function to generate JSON from text
def pdf_to_json(pdf_file):
    text = extract_text_from_pdf(pdf_file)
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=2048)
    outputs = model.generate(**inputs, max_new_tokens=1024)
    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return result

# Gradio interface
interface = gr.Interface(
    fn=pdf_to_json,
    inputs=gr.File(file_types=[".pdf"]),
    outputs=gr.Textbox(label="Generated JSON"),
    title="PDF to JSON using Granite DocLing",
    description="Upload a PDF and get a JSON output using the ibm-granite/granite-docling-258m-demo model."
)

interface.launch()