from pathlib import Path
import torch
from torch.utils.data import Dataset
from transformers import DistilBertTokenizerFast, DistilBertForSequenceClassification
from transformers import Trainer, TrainingArguments

import streamlit as st
import requests

model_one = "distilbert-base-uncased-finetuned-sst-2-english"
model_two = "Newtral/xlm-r-finetuned-toxic-political-tweets-es"
 
def toxicRating(text, model):
    model = AutoModelForSequenceClassification.from_pretrained(model)
    tokenizer = AutoTokenizer.from_pretrained(model)
    
    classifier = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
    results = classifier(text)
    return results
    
def main():
    st.title("TOXIC TWEETS, \n TOXIC OR NOT?")
    prompt = st.header("Select Model")
    selection = st.radio("Models",('Model 1', 'Model 2'))
                         
    
    input = st.text_area("Enter Tweet: ")
    if st.button('Rate') and input:
        if selection == 'Model 1':
            rating = toxicRating(input, model_one)
            st.write(f"Label: {rating[1]} \n Score  : {rating[3]}")
        elif selection == 'Model 2':
            rating = toxicRating(input, model_two)
            st.write(f"Label: {rating[1]} \n Score  : {rating[4]}")
        else:
            st.warning("Enter Tweet")

if __name__ == "__main__":
    main();