import gradio as gr
from huggingface_hub import InferenceClient
from transformers import pipeline
 
advice_model = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
def calculate_advice(food, clothes, utilities):
    total_expenses = food + clothes + utilities
    if total_expenses == 0:
        return "Please enter valid expense values."
    food_percentage = (food / total_expenses) * 100
    clothes_percentage = (clothes / total_expenses) * 100
    utilities_percentage = (utilities / total_expenses) * 100
 
    advice_input = [{"role": "user", "content": f"Food: {food_percentage:.2f}%, Clothes: {clothes_percentage:.2f}%, Utilities: {utilities_percentage:.2f}%."}]
    advice_output = advice_model.chat_completion(advice_input, max_tokens=100)
 
    return advice_output.choices[0].message.content
with gr.Blocks() as demo:
    gr.Markdown("## Budget Advisor")
    food_input = gr.Number(label="Food", value=0)
    clothes_input = gr.Number(label="Clothes", value=0)
    utilities_input = gr.Number(label="Utilities", value=0)
 
    def update_fields():
        additional_expenses.value = add_input_field(additional_expenses.value)
        return additional_expenses.value
 
    advice_button = gr.Button("Advise Me")
    output_text = gr.Textbox(label="Advice Output", interactive=False)
 
    advice_button.click(
        fn=calculate_advice, 
        inputs=[food_input, clothes_input, utilities_input], 
        outputs=output_text
    )
 
if __name__ == "__main__":
    demo.launch()