app.py

import gradio as gr
from gradio_leaderboard import Leaderboard, ColumnFilter, SelectColumns
import pandas as pd
from apscheduler.schedulers.background import BackgroundScheduler
from huggingface_hub import snapshot_download
import plotly.graph_objects as go
import plotly.express as px
from src.about import Tasks, AssetTasks, UncertaintyTasks

from src.about import (
    CITATION_BUTTON_LABEL,
    CITATION_BUTTON_TEXT,
    EVALUATION_QUEUE_TEXT,
    INTRODUCTION_TEXT,
    LLM_BENCHMARKS_TEXT,
    TITLE,
)
from src.display.css_html_js import custom_css
from src.display.utils import (
    BENCHMARK_COLS,
    ASSET_BENCHMARK_COLS,
    UNCERTAINTY_BENCHMARK_COLS,
    COLS,
    ASSET_COLS,
    UNCERTAINTY_COLS,
    EVAL_COLS,
    EVAL_TYPES,
    AutoEvalColumn,
    AutoEvalColumnAsset,
    AutoEvalColumnUncertainty,
    ModelType,
    fields,
    WeightType,
    Precision
)
from src.envs import API, EVAL_REQUESTS_PATH, EVAL_RESULTS_PATH, QUEUE_REPO, REPO_ID, RESULTS_REPO, TOKEN
from src.populate import get_evaluation_queue_df, get_leaderboard_df
from src.submission.submit import add_new_eval
from src.about import Tasks, AssetTasks


def restart_space():
    API.restart_space(repo_id=REPO_ID)

### Space initialisation
try:
    print(EVAL_REQUESTS_PATH)
    snapshot_download(
        repo_id=QUEUE_REPO, local_dir=EVAL_REQUESTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
    )
except Exception:
    restart_space()
try:
    print(EVAL_RESULTS_PATH)
    snapshot_download(
        repo_id=RESULTS_REPO, local_dir=EVAL_RESULTS_PATH, repo_type="dataset", tqdm_class=None, etag_timeout=30, token=TOKEN
    )
except Exception:
    print('error')
    restart_space()


LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, COLS, BENCHMARK_COLS, Tasks)

print(ASSET_COLS)

ASSET_LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, ASSET_COLS, ASSET_BENCHMARK_COLS, AssetTasks)

UNCERTAINTY_LEADERBOARD_DF = get_leaderboard_df(EVAL_RESULTS_PATH, EVAL_REQUESTS_PATH, UNCERTAINTY_COLS, UNCERTAINTY_BENCHMARK_COLS, UncertaintyTasks)
missing_uncertainties = (UNCERTAINTY_LEADERBOARD_DF[UNCERTAINTY_BENCHMARK_COLS] == 0).all(axis=1)
UNCERTAINTY_LEADERBOARD_DF = UNCERTAINTY_LEADERBOARD_DF[~missing_uncertainties]
UNCERTAINTY_LEADERBOARD_DF = UNCERTAINTY_LEADERBOARD_DF.loc[:,~UNCERTAINTY_LEADERBOARD_DF.columns.duplicated()]

(
    finished_eval_queue_df,
    running_eval_queue_df,
    pending_eval_queue_df,
) = get_evaluation_queue_df(EVAL_REQUESTS_PATH, EVAL_COLS)


def init_asset_plot(df):
    acc_col = 'Average ⬆️'
    df = df.sort_values(acc_col).tail(3)
    fig = go.Figure()
    asset_tasks = [task.value.col_name for task in AssetTasks]
    for _, row in df.iterrows():
        fig.add_trace(go.Scatterpolar(
            r=[row[asset_task] for asset_task in asset_tasks],
            theta=asset_tasks,
            opacity = 0.5,
            fill='toself',
            name=row['Model'],
        )
    )
    fig.update_layout(
        autosize=False,
        width=1000,
        height=700,
        title=f"Top 3 accuracies breakdown"
    )
    return fig

def init_perf_plot(df):
    df = df.copy()
    params_col = '#Params (B)'
    df["symbol"] = 2  # Triangle
    df["color"] = ""
    df.loc[df["Model"].str.contains("granite"), "color"] = "grey"
    acc_col = 'Average ⬆️'
    fig = go.Figure()
    for i in df.index:
        fig.add_trace(
            go.Scatter(
                x=[df.loc[i, params_col]],
                y=[df.loc[i, acc_col]],
                name=df.loc[i, "Model"],
                # hovertemplate="<b>%{text}</b><br><br>",
                text=[df.loc[i, "Model"]]
            )
        )

    fig.update_layout(
        autosize=False,
        width=650,
        height=600,
        title=f"Model Size Vs Accuracy",
        xaxis_title=f"{params_col}",
        yaxis_title="Accuracy",
    )
    return fig

def init_leaderboard(dataframe, auto_eval_col_class):
    if dataframe is None or dataframe.empty:
        raise ValueError("Leaderboard DataFrame is empty or None.")
    return Leaderboard(
        value=dataframe,
        datatype=[c.type for c in fields(auto_eval_col_class)],
        select_columns=SelectColumns(
            default_selection=[c.name for c in fields(auto_eval_col_class) if c.displayed_by_default],
            cant_deselect=[c.name for c in fields(auto_eval_col_class) if c.never_hidden],
            label="Select Columns to Display:",
        ),
        search_columns=[auto_eval_col_class.model.name, auto_eval_col_class.license.name],
        hide_columns=[c.name for c in fields(auto_eval_col_class) if c.hidden],
        filter_columns=[
            ColumnFilter(auto_eval_col_class.model_type.name, type="checkboxgroup", label="Model types"),
            ColumnFilter(auto_eval_col_class.precision.name, type="checkboxgroup", label="Precision"),
            ColumnFilter(
                auto_eval_col_class.params.name,
                type="slider",
                min=0.01,
                max=800,
                label="Select the number of parameters (B)",
            ),
            ColumnFilter(
                auto_eval_col_class.still_on_hub.name, type="boolean", label="Deleted/incomplete", default=False
            ),
        ],
        bool_checkboxgroup_label="Hide models",
        interactive=False,
    )


demo = gr.Blocks(css=custom_css)
with demo:
    gr.HTML(TITLE)
    gr.Markdown(INTRODUCTION_TEXT, elem_classes="markdown-text")

    with gr.Tabs(elem_classes="tab-buttons") as tabs:
        with gr.TabItem("🏅 LLM Benchmark", elem_id="llm-benchmark-tab-table", id=0):
            leaderboard = init_leaderboard(LEADERBOARD_DF, AutoEvalColumn)
        
        with gr.TabItem("🛠️ Asset Benchmark", elem_id="llm-benchmark-asset-tab-table", id=1):
            leaderboard = init_leaderboard(ASSET_LEADERBOARD_DF, AutoEvalColumnAsset)
        
        with gr.TabItem("😵‍💫 Uncertainty Benchmark", elem_id="llm-benchmark-asset-tab-table", id=2):
            leaderboard = init_leaderboard(UNCERTAINTY_LEADERBOARD_DF, AutoEvalColumnUncertainty)

        with gr.TabItem("📊 Performance Plot", elem_id="llm-benchmark-tab-table", id=3):
            print(LEADERBOARD_DF.columns)
            # gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
            perf_plot = gr.components.Plot(
                # value=init_perf_plot(LEADERBOARD_DF),
                value=init_asset_plot(ASSET_LEADERBOARD_DF),
                elem_id="bs1-plot",
                show_label=False,
            )

        with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=4):
            gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")

        with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=5):
            with gr.Column():
                with gr.Row():
                    gr.Markdown(EVALUATION_QUEUE_TEXT, elem_classes="markdown-text")

                with gr.Column():
                    with gr.Accordion(
                        f"✅ Finished Evaluations ({len(finished_eval_queue_df)})",
                        open=False,
                    ):
                        with gr.Row():
                            finished_eval_table = gr.components.Dataframe(
                                value=finished_eval_queue_df,
                                headers=EVAL_COLS,
                                datatype=EVAL_TYPES,
                                row_count=5,
                            )
                    with gr.Accordion(
                        f"🔄 Running Evaluation Queue ({len(running_eval_queue_df)})",
                        open=False,
                    ):
                        with gr.Row():
                            running_eval_table = gr.components.Dataframe(
                                value=running_eval_queue_df,
                                headers=EVAL_COLS,
                                datatype=EVAL_TYPES,
                                row_count=5,
                            )

                    with gr.Accordion(
                        f"⏳ Pending Evaluation Queue ({len(pending_eval_queue_df)})",
                        open=False,
                    ):
                        with gr.Row():
                            pending_eval_table = gr.components.Dataframe(
                                value=pending_eval_queue_df,
                                headers=EVAL_COLS,
                                datatype=EVAL_TYPES,
                                row_count=5,
                            )
            with gr.Row():
                gr.Markdown("# ✉️✨ Submit your model here!", elem_classes="markdown-text")

            with gr.Row():
                with gr.Column():
                    model_name_textbox = gr.Textbox(label="Model name")
                    revision_name_textbox = gr.Textbox(label="Revision commit", placeholder="main")
                    model_type = gr.Dropdown(
                        choices=[t.to_str(" : ") for t in ModelType if t != ModelType.Unknown],
                        label="Model type",
                        multiselect=False,
                        value=None,
                        interactive=True,
                    )

                with gr.Column():
                    precision = gr.Dropdown(
                        choices=[i.value.name for i in Precision if i != Precision.Unknown],
                        label="Precision",
                        multiselect=False,
                        value="float16",
                        interactive=True,
                    )
                    weight_type = gr.Dropdown(
                        choices=[i.value.name for i in WeightType],
                        label="Weights type",
                        multiselect=False,
                        value="Original",
                        interactive=True,
                    )
                    base_model_name_textbox = gr.Textbox(label="Base model (for delta or adapter weights)")

            submit_button = gr.Button("Submit Eval")
            submission_result = gr.Markdown()
            submit_button.click(
                add_new_eval,
                [
                    model_name_textbox,
                    base_model_name_textbox,
                    revision_name_textbox,
                    precision,
                    weight_type,
                    model_type,
                ],
                submission_result,
            )

    with gr.Row():
        with gr.Accordion("📙 Citation", open=False):
            citation_button = gr.Textbox(
                value=CITATION_BUTTON_TEXT,
                label=CITATION_BUTTON_LABEL,
                lines=20,
                elem_id="citation-button",
                show_copy_button=True,
            )

scheduler = BackgroundScheduler()
scheduler.add_job(restart_space, "interval", seconds=1800)
scheduler.start()
demo.queue(default_concurrency_limit=40).launch()