app.py

import gradio as gr
from Bio import PDB
import requests
import os
import pandas as pd

# Function to download and parse the PDB file
def fetch_and_parse_pdb(url):
    # Download the PDB file
    response = requests.get(url)
    if response.status_code != 200:
        return "Failed to download the PDB file. Please check the URL."

    # Save the file locally
    pdb_filename = url.split("/")[-1]
    with open(pdb_filename, 'wb') as f:
        f.write(response.content)

    # Parse the PDB file
    parser = PDB.PDBParser()
    structure = parser.get_structure(pdb_filename, pdb_filename)
    
    # Prepare the header information for output
    headers = structure.header
    header_items = []

    for key, value in headers.items():
        if isinstance(value, list):
            for item in value:
                header_items.append((key, item))
        else:
            header_items.append((key, value))

    # Convert to DataFrame for better formatting
    df = pd.DataFrame(header_items, columns=['Header', 'Content'])

    # Clean up the downloaded file
    os.remove(pdb_filename)

    return df

# Create Gradio interface
iface = gr.Interface(
    fn=fetch_and_parse_pdb,
    inputs="text",
    outputs=gr.Dataframe(interactive=True),
    title="PDB Metadata Explorer",
    description="Input the URL of a Protein Data Bank (PDB) file and retrieve its metadata"
)

# Launch the interface
iface.launch()