app.py

import streamlit as st
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, pipeline
import torch

# Load grammar correction model
@st.cache_resource
def load_grammar_model():
    model_name = "prithivida/grammar_error_correcter_v1"
    tokenizer = AutoTokenizer.from_pretrained(model_name)
    model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
    return tokenizer, model

# Load explanation model
@st.cache_resource
def load_explainer():
    explainer = pipeline("text2text-generation", model="google/flan-t5-base", max_length=256)
    return explainer

grammar_tokenizer, grammar_model = load_grammar_model()
explanation_model = load_explainer()

# Grammar correction function
def correct_grammar(text):
    input_text = "gec: " + text
    inputs = grammar_tokenizer.encode(input_text, return_tensors="pt", truncation=True)
    outputs = grammar_model.generate(inputs, max_length=512, num_beams=4, early_stopping=True)
    corrected_text = grammar_tokenizer.decode(outputs[0], skip_special_tokens=True)
    return corrected_text

# Explanation function using second model
def explain_correction(original, corrected):
    prompt = f"Explain the grammar improvements made from: \"{original}\" to: \"{corrected}\""
    result = explanation_model(prompt)[0]['generated_text']
    return result

# Streamlit App UI
st.title("📝 Smart Grammar Correction with Explanations")
st.write("Enter your text below. The AI will correct grammar **and explain why** the changes were made using grammar principles.")

user_input = st.text_area("Your Text", height=200, placeholder="Type or paste your text here...")

if st.button("Correct and Explain"):
    if user_input.strip():
        with st.spinner("Correcting grammar..."):
            corrected = correct_grammar(user_input)

        with st.spinner("Explaining corrections..."):
            explanation = explain_correction(user_input, corrected)

        st.subheader("✅ Corrected Text")
        st.success(corrected)

        st.subheader("📘 Explanation (Why it was changed)")
        st.markdown(explanation)
    else:
        st.warning("Please enter some text to correct.")