app.py

import gradio as gr
import os
from docx import Document
from utils import extract_text_from_pdf, parse_pdf_to_dict

def replace_red_text_with_data(doc_path, data_dict):
    doc = Document(doc_path)

    for para in doc.paragraphs:
        for run in para.runs:
            if run.font.color and run.font.color.rgb and run.font.color.rgb.hex == "FF0000":
                text = run.text.strip()
                if text in data_dict:
                    run.text = data_dict[text]

    return doc

def process_files(pdf_file, template_docx):
    # Save uploaded files temporarily
    pdf_path = "temp_input.pdf"
    doc_path = "temp_template.docx"
    output_path = "filled_output.docx"

    pdf_file.save(pdf_path)
    template_docx.save(doc_path)

    # Extract and parse PDF
    raw_text = extract_text_from_pdf(pdf_path)
    data_dict = parse_pdf_to_dict(raw_text)

    # Replace red text with data
    final_doc = replace_red_text_with_data(doc_path, data_dict)

    # Save final document
    final_doc.save(output_path)

    return output_path

demo = gr.Interface(
    fn=process_files,
    inputs=[
        gr.File(label="Upload PDF Report", file_types=[".pdf"]),
        gr.File(label="Upload Word Template (.docx)", file_types=[".docx"])
    ],
    outputs=gr.File(label="Download Filled Report (.docx)"),
    title="Audit Report Generator",
    description="Upload a PDF and a Word template. This tool will auto-fill red-highlighted fields with data from the PDF."
)

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