app.py

import streamlit as st
from transformers import T5Tokenizer, T5ForConditionalGeneration
from happytransformer import HappyTextToText, TTSettings

# Initialize the grammar correction model
happy_tt = HappyTextToText("T5", "vennify/t5-base-grammar-correction")
args = TTSettings(num_beams=5, min_length=1)

# Initialize T5 spelling correction tokenizer and model
tokenizer = T5Tokenizer.from_pretrained("thaboe01/t5-spelling-corrector")
model = T5ForConditionalGeneration.from_pretrained("thaboe01/t5-spelling-corrector")

# Ensure the model is on the CPU
device = "cpu"
model = model.to(device)

# Function to split text into chunks
def split_text(text, chunk_size=500):
    chunks = []
    for i in range(0, len(text), chunk_size):
        chunks.append(text[i:i+chunk_size])
    return chunks

# Function to correct spelling using T5 model
def correct_spelling(text):
    input_ids = tokenizer(text, return_tensors="pt").input_ids.to(device)
    outputs = model.generate(input_ids)
    corrected_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return corrected_text

# Streamlit app
def main():
    st.title("WordWarden: Spelling and Grammar Checker")
    st.markdown("Welcome to WordWarden! Enter your text below and click the 'Check' button to see the spelling and grammar corrections.")
    
    # Input text area
    text_input = st.text_area("Enter your text here:")
    
    # Initialize counters
    spelling_counter = 0
    grammar_counter = 0
    
    # Check button
    if st.button("Check"):
        # Split the text into manageable chunks
        text_chunks = split_text(text_input)
        
        corrected_spelling_chunks = []
        corrected_grammar_chunks = []
        
        # Process each chunk
        for chunk in text_chunks:
            try:
                # Spelling correction
                corrected_spelling = correct_spelling(chunk)
                corrected_spelling_chunks.append(corrected_spelling)
                
                # Grammar correction
                result = happy_tt.generate_text(f"grammar: {chunk}", args=args)
                corrected_grammar = result.text
                corrected_grammar_chunks.append(corrected_grammar)
                
                # Increment counters if corrections were made
                if corrected_spelling != chunk:
                    spelling_counter += 1
                if corrected_grammar != chunk:
                    grammar_counter += 1
            except Exception as e:
                st.error(f"Error processing chunk: {chunk}\n{e}")
        
        # Combine chunks back into full text
        corrected_spelling_text = ' '.join(corrected_spelling_chunks)
        corrected_grammar_text = ' '.join(corrected_grammar_chunks)
        
        # Display corrected text
        st.subheader("Corrected Text:")
        st.write(corrected_grammar_text)  # Display grammar-corrected text
    
    # Display counters
    st.subheader("Corrections Summary")
    st.write(f"Spelling Corrections: {spelling_counter}")
    st.write(f"Grammar Corrections: {grammar_counter}")

if __name__ == "__main__":
    main()