app.py

import gradio as gr
import argparse
import os
import random
import numpy as np
import torch
import torch.backends.cudnn as cudnn
from minigpt4.common.config import Config
from minigpt4.common.dist_utils import get_rank
from minigpt4.common.registry import registry
from minigpt4.conversation.conversation_esm import Chat, CONV_VISION
import esm

# ProteinGPT Initialization Function
def initialize_chat(args):
    cfg = Config(args)
    model_config = cfg.model_cfg
    model_config.device_8bit = 0
    model_cls = registry.get_model_class(model_config.arch)
    model = model_cls.from_config(model_config).to('cpu')
    vis_processor_cfg = cfg.datasets_cfg.cc_sbu_align.vis_processor.train
    vis_processor = registry.get_processor_class(vis_processor_cfg.name).from_config(vis_processor_cfg)
    chat = Chat(model, vis_processor, device='cpu')
    return chat

# Gradio Reset Function
def gradio_reset(chat_state, img_list):
    if chat_state is not None:
        chat_state.messages = []
    if img_list is not None:
        img_list = []
    return None, gr.update(value=None, interactive=True), gr.update(placeholder='Please upload your protein structure and sequence first', interactive=False), gr.update(value="Upload & Start Chat", interactive=True), chat_state, img_list

# Upload Function
def upload_protein(structure, sequence, text_input, chat_state):
    # Check if structure and sequence files are valid
    if structure is None or not structure.endswith(".pt"):
        return (None, None, None, gr.update(placeholder="Invalid structure file, must be a .pt file.", interactive=True), chat_state, None)
    if sequence is None or not sequence.endswith(".pt"):
        return (None, None, None, gr.update(placeholder="Invalid sequence file, must be a .pt file.", interactive=True), chat_state, None)
    
    # Load protein structure and sequence
    pdb_embedding = torch.load(structure, map_location=torch.device('cpu'))
    sample_pdb = pdb_embedding.to('cpu')
    
    seq_embedding = torch.load(sequence, map_location=torch.device('cpu'))
    sample_seq = seq_embedding.to('cpu')

    # Initialize the conversation state
    chat_state = CONV_VISION.copy()
    img_list = []
    
    # Upload protein data
    llm_message = chat.upload_protein(sample_pdb, sample_seq, chat_state, img_list)
    
    # Return the required outputs
    return (gr.update(interactive=False),  # Disable structure file input
            gr.update(interactive=False),  # Disable sequence file input
            gr.update(interactive=True, placeholder='Type and press Enter'),  # Enable the text input box
            gr.update(value="Start Chatting", interactive=False),  # Update upload button state
            chat_state,  # Return the conversation state
            img_list)  # Return the list of images (if any)
# Ask Function
def gradio_ask(user_message, chatbot, chat_state):
    if len(user_message) == 0:
        return gr.update(interactive=True, placeholder='Input should not be empty!'), chatbot, chat_state
    chat.ask(user_message, chat_state)
    chatbot = chatbot + [[user_message, None]]
    return '', chatbot, chat_state

# Answer Function
def gradio_answer(chatbot, chat_state, img_list, num_beams, temperature):
    img_list = [mat.half() for mat in img_list]
    llm_message = chat.answer(conv=chat_state, img_list=img_list, max_new_tokens=300, num_beams=num_beams, temperature=temperature, max_length=2000)[0]
    chatbot[-1][1] = llm_message
    return chatbot, chat_state, img_list

# Command-line Argument Parsing
def parse_args():
    parser = argparse.ArgumentParser(description="Demo")
    parser.add_argument("--cfg-path", help="path to configuration file.", default='configs/evaluation.yaml')
    parser.add_argument(
        "--options",
        nargs="+",
        help="override some settings in the used config, the key-value pair "
        "in xxx=yyy format will be merged into config file (deprecate), "
        "change to --cfg-options instead.",
    )
    args = parser.parse_args()
    return args

# Demo Gradio Interface
title = """<h1 align="center">Demo of ProteinGPT</h1>"""
description = """<h3>Upload your protein sequence and structure and start chatting with your protein!</h3>"""
article = """<div style='display:flex; gap: 0.25rem; '><a href='https://huggingface.co/AI-BIO/ProteinGPT-Llama3'><img src='https://img.shields.io/badge/Project-Page-Green'></a><a href='https://github.com'><img src='https://img.shields.io/badge/Github-Code-blue'></a></div>"""

args = parse_args()  # Parse arguments to get config and model info
chat = initialize_chat(args)  # Initialize ProteinGPT model

with gr.Blocks() as demo:
    gr.Markdown(title)
    gr.Markdown(description)
    gr.Markdown(article)

    with gr.Row():
        with gr.Column(scale=0.5):
            structure = gr.File(type="filepath", label="Upload Protein Structure", show_label=True)
            sequence = gr.File(type="filepath", label="Upload Protein Sequence", show_label=True)
            upload_button = gr.Button(value="Upload & Start Chat", interactive=True, variant="primary")
            clear = gr.Button("Restart")
            num_beams = gr.Slider(minimum=1, maximum=5, value=1, step=1, interactive=True, label="Beam search numbers")
            temperature = gr.Slider(minimum=0.1, maximum=2.0, value=1.0, step=0.1, interactive=True, label="Temperature")

        with gr.Column():
            chat_state = gr.State()
            img_list = gr.State()
            chatbot = gr.Chatbot(label='ProteinGPT')
            text_input = gr.Textbox(label='User', placeholder='Please upload your image first', interactive=False)
    
    upload_button.click(upload_protein, 
                    [structure, sequence, text_input, chat_state], 
                    [structure, sequence, text_input, upload_button, chat_state, img_list])
    text_input.submit(gradio_ask, [text_input, chatbot, chat_state], [text_input, chatbot, chat_state]).then(gradio_answer, [chatbot, chat_state, img_list, num_beams, temperature], [chatbot, chat_state, img_list])
    clear.click(gradio_reset, [chat_state, img_list], [chatbot, structure, sequence, text_input, upload_button, chat_state, img_list], queue=False)

demo.launch(share=True)