app.py

import streamlit as st
from transformers import pipeline
import tensorflow as tf
import numpy as np
import pandas as pd
from tensorflow.keras.layers import TextVectorization
from tensorflow import keras

model = tf.keras.models.load_model('toxicity_model.h5')

dataset = pd.read_csv('train.csv')
comments = dataset['comment_text']

vectorizer = TextVectorization(max_tokens = 2500000,
                               output_sequence_length=1800,
                               output_mode='int')

vectorizer.adapt(comments.values)

st.title('Toxicity Classifier')

st.header('Write a message here:')
text = st.text_area('The toxicity of the message will be evaluated.',
                   value = "You're fucking ugly.")

input_str = vectorizer(text)
res = model.predict(np.expand_dims(input_str,0))
classification = res[0].tolist()

toxicity = classification[0]
toxicity_severe = classification[1]
obscene = classification[2]
threat = classification[3]
insult = classification[4]
identity_hate = classification[5]

highest_class = "Severe toxicity"
highest_class_rating = toxicity_severe
if(obscene > highest_class_rating):
    highest_class = "Obscenity"
    highest_class_rating = obscene
if(threat > highest_class_rating):
    highest_class = "Threat"
    highest_class_rating = threat
if(insult > highest_class_rating):
    highest_class = "Insult"
    highest_class_rating = insult
if(identity_hate > highest_class_rating):
    highest_class = "Identity hate"
    highest_class_rating = identity_hate

st.write("---")
st.write("Overall toxicity rating: " +str(toxicity))
st.write("---")
st.write("Classifications:")
if(toxicity_severe > 0.5):
    st.write("Severely toxic - " +str(toxicity_severe))
if(obscene > 0.5):
    st.write("Obscene - " +str(obscene))
if(threat > 0.5):
    st.write("Threat - " +str(threat))
if(insult > 0.5):
    st.write("Insult - " +str(insult))
if(identity_hate > 0.5):
    st.write("Identity hate - " +str(identity_hate))
st.write("---")
st.write("Invalid classifications:")
if(toxicity_severe <= 0.5):
    st.write("Severely toxic - " +str(toxicity_severe))
if(obscene <= 0.5):
    st.write("Obscene - " +str(obscene))
if(threat <= 0.5):
    st.write("Threat - " +str(threat))
if(insult <= 0.5):
    st.write("Insult - " +str(insult))
if(identity_hate <= 0.5):
    st.write("Identity hate - " +str(identity_hate))