app.py

import gradio as gr
import PyMuPDF as fitz  # Importing PyMuPDF as fitz

# Function to extract text from a PDF
def extract_pdf_text(file):
    doc = fitz.open(file.name)  # Open the PDF file using PyMuPDF
    text = ""
    for page in doc:
        text += page.get_text()  # Extract text from each page
    return text

# Gradio interface
output_format_dropdown = gr.Dropdown(
    choices=["txt", "pdf", "docx"],
    label="Output Format",
    default="txt"
)

iface = gr.Interface(
    fn=extract_pdf_text, 
    inputs=gr.File(label="Upload PDF File"), 
    outputs=[gr.Textbox(label="Extracted Text"), output_format_dropdown],
    live=True
)

if __name__ == "__main__":
    iface.launch()