benchmarks/spaceship-titanic/scripts/eval.py

import sys
import os
sys.path.append("../env")
import pandas as pd
import numpy as np
from datasets import load_dataset

def get_score(submission_folder = "../env"):
    submission_path = os.path.join(submission_folder, "submission.csv")
    submission = pd.read_csv(submission_path)
    test_data = pd.read_csv(os.path.join(os.path.dirname(__file__),'answer.csv'))

    correct_matches = submission['Transported'] == test_data['Transported']
    accuracy = correct_matches.sum() / len(submission)

    return accuracy

if __name__ == "__main__":
    print(get_score())