import streamlit as st
from transformers import pipeline

# ---------------------- Page Config ----------------------
st.set_page_config(page_title="🧠 Smart NLP Assistant", layout="centered")

# ---------------------- Sidebar --------------------------
st.sidebar.title("🧠 Smart NLP Assistant")
task = st.sidebar.selectbox("📌 Select an NLP Task:", 
                            ["📝 Text Classification", "❓ Question Answering", "📰 Summarization"])

st.title("🌟 NLP Companion App")
st.markdown("Interact with powerful **Transformer models** for different Natural Language Processing tasks right in your browser!")

# -------------------- Load Models ------------------------
@st.cache_resource
def load_pipelines():
    sentiment_model = pipeline("text-classification", model="cardiffnlp/twitter-roberta-base-sentiment-latest")
    qa_model = pipeline("question-answering", model="deepset/roberta-base-squad2")
    summarizer_model = pipeline("summarization", model="facebook/bart-large-cnn")
    return sentiment_model, qa_model, summarizer_model

text_classifier, qa_pipeline, summarizer = load_pipelines()

# -------------------- Task: Sentiment Analysis ---------------------
if task == "📝 Text Classification":
    st.subheader("📝 Sentiment Analysis")
    user_input = st.text_input("✍️ Enter a sentence to analyze sentiment:")
    
    if user_input.strip():
        with st.spinner("🔍 Analyzing sentiment..."):
            result = text_classifier(user_input)[0]
        st.success(f"📣 **Sentiment:** {result['label'].capitalize()}")
        st.info(f"🎯 **Confidence Score:** {result['score']*100:.2f}%")
    else:
        st.warning("⚠️ Please enter some text above to proceed.")

# -------------------- Task: Question Answering ---------------------
elif task == "❓ Question Answering":
    st.subheader("❓ Ask a Question")
    context = st.text_area("📄 Paste the context passage:")
    question = st.text_input("🔍 Ask your question here:")
    
    if context.strip() and question.strip():
        with st.spinner("💡 Finding the answer..."):
            result = qa_pipeline(question=question, context=context)
        st.success(f"✅ **Answer:** {result['answer']}")
        st.info(f"🎯 **Confidence Score:** {result['score']*100:.2f}%")
    elif question or context:
        st.warning("⚠️ Please provide both context and question.")

# -------------------- Task: Summarization ---------------------
elif task == "📰 Summarization":
    st.subheader("📰 Summarize Text")
    long_text = st.text_area("📃 Paste or type the long text to summarize:")
    
    if long_text.strip():
        with st.spinner("🧠 Generating summary..."):
            summary = summarizer(long_text, max_length=150, min_length=30, do_sample=False)[0]
        st.success("📝 **Summary:**")
        st.write(summary["summary_text"])
    else:
        st.warning("⚠️ Please enter content to summarize.")