app.py

import torch
import torch.nn as nn
from transformers import AutoTokenizer,AutoModelForSequenceClassification,AutoConfig
import numpy as np
import pandas as pd
import re
from Bio.Seq import Seq
from collections import OrderedDict	
from transformers import set_seed
import random
import gradio as gr

def setup_seed(seed):
    set_seed(seed)
    torch.manual_seed(seed)
    torch.cuda.manual_seed_all(seed)
    np.random.seed(seed)
    random.seed(seed)
    torch.backends.cudnn.deterministic = True
setup_seed(4)

device = "cpu"
model_checkpoint = "facebook/esm2_t6_8M_UR50D"

config = AutoConfig.from_pretrained(model_checkpoint)
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)


def conotoxinfinder(files):
    fr=open(files, 'r')
    seqs = []
    for line in fr:
        if not line.startswith('>'):
            line = line.replace('\n','')
            line = line.replace(' ','')
            if line.islower():
                seqs.append(str((Seq(line).translate())))
            else:
                seqs.append(line)

    model = AutoModelForSequenceClassification.from_pretrained(model_checkpoint, num_labels=1)
    model.load_state_dict(torch.load("best_model.pth"))
    model = model.to(device)

    value_all = []
    for i in seqs:
        tokenizer_test = tokenizer(i, return_tensors='pt').to(device)
        with torch.no_grad():
            value = model(**tokenizer_test)
            value_all.append(np.exp(value["logits"][0].item()))

    summary = OrderedDict()
    summary['Seq'] = seqs
    summary['Value'] = value_all
    summary_df = pd.DataFrame(summary)
    summary_df.to_csv('output.csv', index=False)
    return 'output.csv'

with open("conotoxinfinder.md", "r") as f:
    description = f.read()
iface = gr.Interface(fn=conotoxinfinder,
                     title="ConotoxinFinder α7 regression",
                     inputs=["file"
                             ],
                     outputs= "file",
                     description=description
                     )
iface.launch()