import streamlit as st
import pandas as pd
from simpletransformers.classification import ClassificationModel
import torch
import random
import os
import numpy as np
import subprocess

import requests

random.seed(4)
np.random.seed(4)
torch.manual_seed(4)
np.random.seed(4)


#def start_fastapi():
    

# Start the FastAPI server
#start_fastapi()






# Set the Streamlit app title
st.title("Molecule Toxicity Predictions")

# Set the model path
path = 'ToxicityPrediction/Models/transformers/checkpoint-149-epoch-1'

# Load the model from the stage
#loaded_model = ClassificationModel('roberta', path, use_cuda = False)
#rob_chem_model = ClassificationModel('roberta', 'seyonec/SMILES_tokenized_PubChem_shard00_160k',use_cuda=False ,args={'evaluate_each_epoch':True , 'evaluate_during_training_verbose':True, 'seed':4})
# Predict based on the input
rob_chem_model = ClassificationModel('roberta', 'BasselAhmed/RobertaChemClinToxTuned',use_cuda=False ,args={'evaluate_each_epoch':True , 'evaluate_during_training_verbose':True, 'seed':4})

rob_chem_model.model.eval() 
#target_name= st.text_input('Enter a SMILES string:')
target_name = st.text_area("Enter smiles (one per line):", "")

target_name_list = target_name.splitlines()
target_name_list = [x.strip() for x in target_name_list]
predict_toxicity = st.button('Predict Toxicity')
if predict_toxicity:
    predictions, raw_outputs = rob_chem_model.predict(target_name_list)
    df_pred = pd.DataFrame({'Smiles':target_name_list,'Predictions': predictions})
    st.dataframe(df_pred)