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

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,
    COLS,
    EVAL_COLS,
    EVAL_TYPES,
    AutoEvalColumn,
    fields,
)
from src.envs import (
    API,
    EVAL_DETAILED_RESULTS_PATH,
    EVAL_RESULTS_PATH,
    EVAL_DETAILED_RESULTS_REPO,
    REPO_ID,
    RESULTS_REPO,
    TOKEN,
)
from src.populate import get_leaderboard_df


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


### Space initialisation
try:
    print(EVAL_DETAILED_RESULTS_REPO)
    snapshot_download(
        repo_id=EVAL_DETAILED_RESULTS_REPO,
        local_dir=EVAL_DETAILED_RESULTS_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:
    restart_space()


LEADERBOARD_DF = get_leaderboard_df(RESULTS_REPO)


def init_leaderboard(dataframes):
    if dataframes is None or not dataframes:
        raise ValueError("Leaderboard data is empty or None.")
    
    def create_leaderboard(df):
        return Leaderboard(
            value=df,
            datatype=[c.type for c in fields(AutoEvalColumn)],
            select_columns=SelectColumns(
                default_selection=[c.name for c in fields(AutoEvalColumn) if c.displayed_by_default],
                cant_deselect=[c.name for c in fields(AutoEvalColumn) if c.never_hidden],
                label="Select Columns to Display:",
            ),
            search_columns=[AutoEvalColumn.model.name],
            hide_columns=[c.name for c in fields(AutoEvalColumn) if c.hidden],
            filter_columns=[],
            interactive=False,
        )
    
    subset_names = list(dataframes.keys())
    selected_subset = gr.Dropdown(choices=subset_names, label="Select Dataset Subset", value=subset_names[0])
    
    leaderboard = gr.Dynamic(create_leaderboard, inputs=[selected_subset], outputs="output")
    selected_subset.change(
        fn=lambda x: create_leaderboard(dataframes[x]),
        inputs=[selected_subset],
        outputs=leaderboard
    )
    
    return leaderboard

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("🏅 LiveBench Results", elem_id="llm-benchmark-tab-table", id=0):
            leaderboard = init_leaderboard(LEADERBOARD_DF)

        with gr.TabItem("📝 About", elem_id="llm-benchmark-tab-table", id=2):
            gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")

    # 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()