app.py

import streamlit as st
from transformers import pipeline
from translate import Translator
import re

# Load models
def load_models():
    """Loads the sentiment analysis and summarization models."""
    sentiment_analyzer = pipeline("text-classification", model="miltonc/distilbert-base-uncased_ft_5")
    summarizer = pipeline("summarization", model="FelixChao/T5-Chinese-Summarization")
    return sentiment_analyzer, summarizer

# Max length for models
SENTIMENT_MAX_LENGTH = 512  # Maximum input length for DistilBERT sentiment analysis
SUMMARIZER_MAX_LENGTH = 512  # Maximum input length for T5 summarization (approximate)

def sentiment_analysis(text, sentiment_analyzer):
    """Analyzes the sentiment of the input text."""
    try:
        result = sentiment_analyzer(text)[0]
        return result
    except Exception as e:
        print(f"sentiment_analysis error for '{text}': {e}. Returning 'sentiment_analysis Failed'")
        return {"label": "sentiment_analysis Failed", "score": 0.0}

def summarize_news(text, summarizer):
    """Summarizes the input news text."""
    try:
        summary = summarizer(text, max_length=30, min_length=10)[0]['summary_text']
        return summary
    except Exception as e:
        print(f"Summarization error for '{text}': {e}. Returning 'Summarization Failed'")
        return "Summarization Failed"

def translate_text(text, target_language='en', source_language='zh-TW'):
    """Translates Chinese text to English."""
    translator = Translator(to_lang=target_language, from_lang=source_language)
    try:
        translation = translator.translate(text)
        return translation
    except Exception as e:
        print(f"Translation error for '{text}': {e}. Returning 'Translation Failed'")
        return "Translation Failed"

# Check if the input text is Chinese
def is_chinese(text):
    """Checks if the input text contains Chinese characters."""
    return bool(re.search(r'[\u4e00-\u9fff]', text))

# Main Streamlit app
def main():
    """Main Streamlit application."""
    st.title("AI-Powered Sentiment Analysis and Summarization")

    sentiment_analyzer, summarizer = load_models()

    text = st.text_area(f"Enter the Chinese text here (Max {max(SENTIMENT_MAX_LENGTH, SUMMARIZER_MAX_LENGTH)} characters).....", height=200)

    if st.button("Analyze"):
        if text:
            if not is_chinese(text):
                st.error("Error: Please enter Chinese text.")
                return

            if len(text) > SENTIMENT_MAX_LENGTH:
                st.error(f"Error: Input text for sentiment analysis exceeds the maximum length of {SENTIMENT_MAX_LENGTH} characters. Please shorten your text.")
                return

            if len(text) > SUMMARIZER_MAX_LENGTH:
                st.error(f"Error: Input text for summarization exceeds the maximum length of {SUMMARIZER_MAX_LENGTH} characters. Please shorten your text.")
                return

            with st.spinner("Analyzing sentiment..."):
                text_en = translate_text(text, target_language='en', source_language='zh-TW')
                sentiment_result = sentiment_analysis(text_en, sentiment_analyzer)
                label_map = {"LABEL_0": 'negative', "LABEL_1": 'neutral', "LABEL_2": 'positive'}
                try:
                    sentiment_label = label_map.get(sentiment_result['label'], 'Unknown')
                    sentiment_score = sentiment_result['score']
                    st.write(f"### Sentiment: {sentiment_label}")
                    st.write(f"Score: {sentiment_score:.4f}")
                except (KeyError, ValueError):
                    st.write("Error processing sentiment result.")
                    st.write(sentiment_result)

            with st.spinner("Summarizing News..."):
                story = summarize_news(text, summarizer)
                st.write("### Summarized News:")
                st.write(story)

if __name__ == "__main__":
    main()