|
import streamlit as st |
|
from newspaper import Article |
|
from transformers import pipeline |
|
from transformers import T5Tokenizer, T5Model, T5ForConditionalGeneration |
|
|
|
|
|
@st.cache_resource |
|
def load_summarizer(): |
|
model_name = "cahya/t5-base-indonesian-summarization-cased" |
|
tokenizer = T5Tokenizer.from_pretrained(model_name) |
|
model = T5ForConditionalGeneration.from_pretrained(model_name) |
|
|
|
return pipeline("summarization", model=model, tokenizer=tokenizer) |
|
|
|
summarizer = load_summarizer() |
|
|
|
st.title("📰 Indonesian News Summarizer") |
|
st.write("Enter a URL from an Indonesian news website (e.g. Detik.com)") |
|
|
|
url = st.text_input("Paste the article URL here:") |
|
|
|
if st.button("Show Article Text"): |
|
if url: |
|
try: |
|
article = Article(url, language='id') |
|
article.download() |
|
article.parse() |
|
st.subheader("Full Article:") |
|
st.write(article.text) |
|
st.session_state.article_text = article.text |
|
except Exception as e: |
|
st.error(f"Failed to fetch article: {str(e)}") |
|
else: |
|
st.warning("Please input a valid URL.") |
|
|
|
if st.button("Summarize"): |
|
if "article_text" in st.session_state: |
|
with st.spinner("Summarizing..."): |
|
input_text = "ringkasan: " + st.session_state.article_text |
|
summary = summarizer(input_text, max_length=150, min_length=40, do_sample=False) |
|
st.subheader("Summary:") |
|
st.success(summary[0]['summary_text']) |
|
else: |
|
st.warning("No article text found. Please load the article first.") |
|
|
