import gradio as gr
from visualization import create_radar_plot


def compare_models(df, model_names=None):
    if model_names is None or len(model_names) == 0:
        model_names = [df.sort_values("Model Avg", ascending=False).iloc[0]["Model"]]

    filtered_df = df[df["Model"].isin(model_names)]
    radar_chart = create_radar_plot(df, model_names)

    # Create styled table for model info
    info_html = f"""
    <div class="dark-table-container">
        <table class="dark-styled-table">
            <thead>
                <tr>
                    <th>Model</th>
                    <th>Type</th>
                    <th>Average</th>
                    <th>I/O Cost</th>
                    <th>Single Turn</th>
                    <th>Multi Turn</th>
                </tr>
            </thead>
            <tbody>
    """

    for _, row in filtered_df.iterrows():
        info_html += f"""
            <tr>
                <td>{row['Model']}</td>
                <td>{row['Model Type']}</td>
                <td>{row['Model Avg']:.3f}</td>
                <td>${row['IO Cost']:.2f}</td>
                <td>{row['single turn perf']:.3f}</td>
                <td>{row['multi turn perf']:.3f}</td>
            </tr>
        """

    info_html += """
            </tbody>
        </table>
    </div>
    """

    return info_html, radar_chart


def create_model_comparison_tab(df, HEADER_CONTENT, CARDS):
    with gr.Tab("Model Comparison"):
        gr.HTML(HEADER_CONTENT)
        with gr.Column():
            # Filters row
            with gr.Row(equal_height=True):
                model_selector = gr.Dropdown(
                    choices=df["Model"].unique().tolist(),
                    value=df.sort_values("Model Avg", ascending=False).iloc[0]["Model"],
                    multiselect=True,
                    label="Select Models to Compare",
                )

            # Content
            model_info = gr.HTML()
            radar_plot = gr.Plot()

        model_selector.change(
            fn=lambda m: compare_models(df, m),
            inputs=[model_selector],
            outputs=[model_info, radar_plot],
        )

        return model_info, radar_plot