import io

import pandas as pd

from src.leaderboard.columns import RequiredInputColumns
from src.utils import download_github_file_content

# GitHub API endpoint for the file (handles Git LFS files)
LEADERBOARD_API_URL = "https://api.github.com/repos/upgini/mle-bench/contents/rankings/low/tabular/overall_ranks.csv"


def load_csv_from_github() -> pd.DataFrame:
    """Load the leaderboard CSV from GitHub."""
    csv_content = download_github_file_content(LEADERBOARD_API_URL, timeout=30)

    df = pd.read_csv(io.StringIO(csv_content))

    if df.empty:
        return pd.DataFrame(columns=RequiredInputColumns.values())

    missing_cols = [col for col in RequiredInputColumns.values() if col not in df.columns]
    if missing_cols:
        raise ValueError(f"Leaderboard is missing expected columns: {', '.join(missing_cols)}")

    return df