import gradio as gr
from transformers import pipeline

model_name = "cornelliusyudhawijaya/AG_News_Classification_DistillBert"
classifier = pipeline("text-classification", model=model_name, tokenizer=model_name)

label_names = {0: 'World', 1: 'Sports', 2: 'Business', 3: 'Sci/Tech'}

def classify_text(text):
    result = classifier(text)[0]
    label_id = int(result['label'].split('_')[-1]) 
    label_name = label_names[label_id] 
    return f"Label: {label_name}, Score: {result['score']:.4f}"

# Define the Gradio interface using the new API
iface = gr.Interface(
    fn=classify_text,
    inputs=gr.Textbox(lines=2, placeholder="Enter text here..."),
    outputs="text",
    title="News Classification",
    description="Enter text to classify the news."
)

if __name__ == "__main__":
    iface.launch()