import gradio as gr

from ml_tabular.service import TabularMLService


service = TabularMLService()


def train_model(file_obj, target_column, task_type, test_size, random_state):
    file_path = getattr(file_obj, "name", file_obj)
    return service.run(file_path, target_column, task_type, float(test_size), int(random_state))


with gr.Blocks(
    title="Machine Learning Tabular Lab",
    theme=gr.themes.Soft(primary_hue="green", secondary_hue="gray"),
) as demo:
    gr.Markdown(
        """
        # Machine Learning Tabular Lab
        Upload a CSV dataset, choose the target column, and train a classic machine learning model on free CPU.
        """
    )

    dataset_input = gr.File(label="CSV Dataset", file_types=[".csv"])
    target_input = gr.Textbox(label="Target Column", placeholder="target")
    task_input = gr.Dropdown(
        choices=["Auto", "Classification", "Regression"],
        value="Auto",
        label="Task Type",
    )

    with gr.Row():
        test_size_input = gr.Slider(0.1, 0.4, value=0.2, step=0.05, label="Test Split")
        random_state_input = gr.Slider(1, 999, value=42, step=1, label="Random State")

    run_button = gr.Button("Train Model", variant="primary")

    model_output = gr.Textbox(label="Selected Model", lines=1)
    metric_output = gr.Textbox(label="Metrics", lines=6)
    preview_output = gr.Textbox(label="Feature Preview", lines=8)
    status_output = gr.Textbox(label="Status", lines=3)

    run_button.click(
        fn=train_model,
        inputs=[dataset_input, target_input, task_input, test_size_input, random_state_input],
        outputs=[model_output, metric_output, preview_output, status_output],
    )


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