app.py

import numpy as np
import pandas as pd
import torch
from torch import nn
from torch.utils.data import Dataset, DataLoader
from transformers import AutoTokenizer, BertModel
from sklearn import metrics
import streamlit as st

class ToxicityDataset(Dataset):
    def __init__(self, dataframe, tokenizer, max_len):
        self.tokenizer = tokenizer
        self.data = dataframe
        self.text = self.data.comment_text
        self.targets = self.data.labels
        self.max_len = max_len

    def __len__(self):
        return len(self.text)

    def __getitem__(self, index):
        text = str(self.text[index])
        text = " ".join(text.split())

        inputs = self.tokenizer.encode_plus(
            text,
            None,
            add_special_tokens=True,
            max_length=self.max_len,
            padding="max_length",
            truncation=True,
            return_token_type_ids=True,
        )
        ids = inputs["input_ids"]
        mask = inputs["attention_mask"]
        token_type_ids = inputs["token_type_ids"]
        return {
            "ids": torch.tensor(ids, dtype=torch.long),
            "mask": torch.tensor(mask, dtype=torch.long),
            "token_type_ids": torch.tensor(token_type_ids, dtype=torch.long),
            "targets": torch.tensor(self.targets[index], dtype=torch.float),
        }


def inference():
    model.eval()
    final_targets = []
    final_outputs = []
    with torch.no_grad():
        for _, data in enumerate(testing_loader, 0):
            ids = data["ids"].to(device, dtype=torch.long)
            mask = data["mask"].to(device, dtype=torch.long)
            token_type_ids = data["token_type_ids"].to(device, dtype=torch.long)
            targets = data["targets"].to(device, dtype=torch.float)
            outputs = model(ids, mask, token_type_ids)
            final_targets.extend(targets.cpu().detach().numpy().tolist())
            final_outputs.extend(torch.sigmoid(outputs).cpu().detach().numpy().tolist())
    return final_outputs, final_targets

bert_path = "bert-base-uncased"
bert_tokenizer = AutoTokenizer.from_pretrained(bert_path)
bert_model = BertForSequenceClassification.from_pretrained(bert_path, num_labels=6)
tuned_model = model = torch.load("pytorch_bert_toxic.bin")

tweets_raw = pd.read_csv("test.csv", nrows=20)
labels_raw = pd.read_csv("test_labels.csv", nrows=20)
label_set = ["toxic", "severe_toxic", "obscene", "threat", "insult", "identity_hate"]


MAX_LENGTH = 100
TEST_BATCH_SIZE = 128
test_dataset = ToxicityDataset(test_df, tokenizer, MAX_LENGTH)
test_params = {"batch_size": TEST_BATCH_SIZE, "shuffle": True, "num_workers": 0}
testing_loader = DataLoader(test_dataset, **test_params)

option = st.selectbox("Select a text analysis model:", ("BERT", "Fine-tuned BERT"))
if option == "BERT":
    tokenizer = bert_tokenizer
    model = bert_model
else:
    tokenizer = bert_tokenizer
    model = tuned_model

prediction, targets = inference()
prediction = np.array(prediction) >= 0.5

targets = np.argmax(targets, axis=1)
prediction = np.argmax(prediction, axis=1)
accuracy = metrics.accuracy_score(targets, prediction)
f1_score_micro = metrics.f1_score(targets, prediction, average="micro")
f1_score_macro = metrics.f1_score(targets, prediction, average="macro")

print(f"Accuracy Score = {accuracy}")
print(f"F1 Score (Micro) = {f1_score_micro}")
print(f"F1 Score (Macro) = {f1_score_macro}")

# Write results
st.write("Classification Probabilities")
st.write(f"{neutralProb:.4f} - NEUTRAL")
st.write(f"{toxicProb:.4f} - TOXIC")