BudgetX/GradioPDFComparer.original from qwen.py

import gradio as gr
import camelot
import pandas as pd

# Global variables to store PDF data
pdf1_data = None
pdf2_data = None
dropdown_options = []
table_data = []

def process_pdf(file_path):
    """Extract tables from PDF using Camelot"""
    try:
        tables = camelot.read_pdf(file_path)
        return [table.df for table in tables]
    except Exception as e:
        return f"Error: {str(e)}"

def update_dropdown(pdf1, pdf2):
    """Process both PDFs and update dropdown options"""
    global pdf1_data, pdf2_data, dropdown_options
    
    # Process first PDF
    if pdf1:
        result = process_pdf(pdf1.name)
        if isinstance(result, list):
            pdf1_data = result
        else:
            return [], f"Error processing first PDF: {result}", []
    
    # Process second PDF
    if pdf2:
        result = process_pdf(pdf2.name)
        if isinstance(result, list):
            pdf2_data = result
        else:
            return [], f"Error processing second PDF: {result}", []
    
    # Generate dropdown options when both PDFs are processed
    if pdf1_data and pdf2_data:
        # Create options combining tables from both PDFs
        dropdown_options = []
        for i, table in enumerate(pdf1_data):
            dropdown_options.append(f"PDF1-Table{i+1}")
        for i, table in enumerate(pdf2_data):
            dropdown_options.append(f"PDF2-Table{i+1}")
        return dropdown_options, "Both PDFs processed successfully", []
    
    return [], "Upload both PDFs to see options", []

def update_table(selected_option):
    """Update table based on selected option"""
    global pdf1_data, pdf2_data
    
    if not selected_option:
        return []
    
    # Determine which PDF and table to use
    if selected_option.startswith("PDF1"):
        idx = int(selected_option.split("Table")[1]) - 1
        table_df = pdf1_data[idx]
    else:
        idx = int(selected_option.split("Table")[1]) - 1
        table_df = pdf2_data[idx]
    
    # Convert to list of lists for Gradio
    headers = [f"Col {i+1}" for i in range(len(table_df.columns))]
    rows = [headers] + table_df.values.tolist()
    return rows

with gr.Blocks() as app:
    gr.Markdown("## PDF Table Comparison Tool")
    
    with gr.Row():
        # Left column for first PDF
        with gr.Column():
            pdf1_input = gr.File(label="Upload First PDF", file_types=[".pdf"])
            pdf1_preview = gr.File(label="First PDF Preview")
        
        # Right column for second PDF
        with gr.Column():
            pdf2_input = gr.File(label="Upload Second PDF", file_types=[".pdf"])
            pdf2_preview = gr.File(label="Second PDF Preview")
    
    # Error/warning textbox
    error_output = gr.Textbox(label="Status", interactive=False)
    
    # Dropdown for selecting tables
    dropdown = gr.Dropdown(label="Select Table", choices=[])
    
    # Output table
    output_table = gr.Dataframe(label="Table Content", headers=["Column 1", "Column 2", "Column 3"])

    # Event handling
    pdf1_input.change(
        fn=lambda x: x, 
        inputs=pdf1_input, 
        outputs=pdf1_preview
    )
    
    pdf2_input.change(
        fn=lambda x: x, 
        inputs=pdf2_input, 
        outputs=pdf2_preview
    )
    
    # Update dropdown when both files are uploaded
    inputs = [pdf1_input, pdf2_input]
    pdf1_input.change(update_dropdown, inputs=inputs, outputs=[dropdown, error_output, output_table])
    pdf2_input.change(update_dropdown, inputs=inputs, outputs=[dropdown, error_output, output_table])
    
    # Update table when dropdown selection changes
    dropdown.change(update_table, inputs=dropdown, outputs=output_table)

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