import os
import warnings
import numpy as np
from scipy import stats


significant_level = 0.05 # five percent significant level for one-sided t-test

def hypothesis_test(*results):
    signif = [False] * len(results)

    # choose the index with the maximum mean
    i_best = np.argmax([result.mean() for result in results])

    for i in range(len((results))):
        if i == i_best:
            signif[i] = True
        else:
            signif[i] = [0]*len(results[i])
            if not np.array_equal(results[i], results[i_best]):
                t_stat, p_ = stats.ttest_ind(results[i], results[i_best])
                p = 0.5 * p_  # two-sided p-value -> one-sided p-value
                signif[i] = p > significant_level

    return signif

# score of each method with 5 trials
method_1_scores = np.array([10, 11, 12, 14, 16])
method_2_scores = np.array([14, 21, 20, 17, 12])
method_3_scores = np.array([14, 15, 11, 14, 12])

results = np.array([method_1_scores, method_2_scores, method_3_scores])
print(results)
print(hypothesis_test(*results))