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 (
    INTRODUCTION_TEXT,
    LLM_BENCHMARKS_TEXT,
    TITLE,
)
from src.display.css_html_js import custom_css
from src.display.utils import (
    AutoEvalColumn,
    fields,
)
from src.envs import API, EVAL_RESULTS_PATH, REPO_ID, RESULTS_REPO, TOKEN


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

### Space initialisation
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 = pd.read_csv(f"{EVAL_RESULTS_PATH}/results.csv")
# multiply each column by 100, except first column
for col in LEADERBOARD_DF.columns[1:]:
    LEADERBOARD_DF[col] = LEADERBOARD_DF[col] * 100

# add average column
LEADERBOARD_DF["Average"] = LEADERBOARD_DF.mean(axis=1)

# # move average to the front
LEADERBOARD_DF = LEADERBOARD_DF[["model", "Average"] + [col for col in LEADERBOARD_DF.columns if col not in ["model", "Average"]]]

# round to nearest tenth
LEADERBOARD_DF = LEADERBOARD_DF.round(1)

# sort by average
LEADERBOARD_DF = LEADERBOARD_DF.sort_values(by="Average", ascending=False)


def init_leaderboard(dataframe):
    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(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],
        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)

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


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