import streamlit as st
from transformers import pipeline

# loarding pipeline
sentiment_analyzer = pipeline("sentiment-analysis", model="distilbert-base-uncased-finetuned-sst-2-english")
ner_tagger = pipeline("ner", model="dslim/bert-base-NER", grouped_entities=True)

st.set_page_config(page_title="Customer Support Analyzer", layout="centered")
st.title("📞 AI Customer service Dialogue Analysis")

# Customer type
user_input = st.text_area("Please enter the question or conversation：", height=150)

if st.button("Analyse"):
    if user_input.strip() == "":
        st.warning("Please enter content")
    else:
        with st.spinner("Analysing..."):
            # Emotion
            sentiment_result = sentiment_analyzer(user_input)[0]
            st.subheader("📌 Sentiment analysis results")
            st.write(f"**Emotional type**: {sentiment_result['label']}")
            st.write(f"**Confidence degree**: {sentiment_result['score']:.2f}")

            # Command
            ner_results = ner_tagger(user_input)
            extracted_entities = [ent['word'] for ent in ner_results if ent['score'] > 0.5]

            st.subheader("🔍Problem keyword recognition")
            if extracted_entities:
                st.write(", ".join(set(extracted_entities)))
            else:
                st.write("The specific problem keywords were not identified")