app.py

import streamlit as st
from transformers import pipeline
from textblob import TextBlob
from transformers import BertForSequenceClassification, AdamW, BertConfig
st.set_page_config(layout='wide', initial_sidebar_state='expanded')
col1, col2= st.columns(2)
with col2:
    text = st.text_input("Enter the text you'd like to analyze for spam.")
    aButton = st.button('Analyze') 
with col1:
    st.title("Spamd: Turkish Spam Detector")
    st.markdown("Message spam detection tool for Turkish language. Due the small size of the dataset, I decided to go with transformers technology Google BERT. Using the Turkish pre-trained model BERTurk, I imporved the accuracy of the tool by 18 percent compared to the previous model which used fastText.")
    st.markdown("Original file is located at")
    st.markdown("https://colab.research.google.com/drive/1QuorqAuLsmomesZHsaQHEZgzbPEM8YTH")
 
import torch
import numpy as np
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("dbmdz/bert-base-turkish-uncased")
from transformers import AutoModel
model = BertForSequenceClassification.from_pretrained("NimaKL/spamd_model")
token_id = []
attention_masks = []
def preprocessing(input_text, tokenizer):
    '''
                  Returns <class transformers.tokenization_utils_base.BatchEncoding> with the following fields:
                    - input_ids: list of token ids
                    - token_type_ids: list of token type ids
                    - attention_mask: list of indices (0,1) specifying which tokens should considered by the model (return_attention_mask = True).
    '''
    return tokenizer.encode_plus(
        input_text,
        add_special_tokens = True,
        max_length = 32,
        pad_to_max_length = True,
        return_attention_mask = True,
        return_tensors = 'pt'
            )
device = 'cpu'
    
def predict(new_sentence):
    # We need Token IDs and Attention Mask for inference on the new sentence
    test_ids = []
    test_attention_mask = []
    # Apply the tokenizer
    encoding = preprocessing(new_sentence, tokenizer)
    # Extract IDs and Attention Mask
    test_ids.append(encoding['input_ids'])
    test_attention_mask.append(encoding['attention_mask'])
    test_ids = torch.cat(test_ids, dim = 0)
    test_attention_mask = torch.cat(test_attention_mask, dim = 0)
    # Forward pass, calculate logit predictions
    with torch.no_grad():
        output = model(test_ids.to(device), token_type_ids = None, attention_mask = test_attention_mask.to(device))
        prediction = 'Spam' if np.argmax(output.logits.cpu().numpy()).flatten().item() == 1 else 'Normal'
        pred = 'Predicted Class: '+ prediction
        return pred      

if text or aButton:
    with col2:
        with st.spinner('Wait for it...'):
            st.success(predict(text))