3daedc17ddab948752f3d6b90e56b177ea77ec42,nilearn/mass_univariate/permuted_least_squares.py,,permuted_ols,#Any#Any#Any#Any#Any#Any#Any#,252

Before Change

    h0_fmax = np.sort(np.ravel(np.concatenate(ret)))
    // convert scores into negative log10 p-values
    // TODO: to speed this up, we could threshold scores_original_data
    n_scores = n_descriptors * n_regressors
    ravelized_scores = np.ravel(scores_original_data)
    ret = joblib.Parallel(n_jobs=n_jobs)(joblib.delayed(_convert_to_pvalues)
          (h0_fmax, ravelized_scores[chunk])
          for chunk in gen_even_slices(
            n_scores + 1, min(n_scores, n_jobs)))
    pvals = np.concatenate(ret).reshape((n_regressors, n_descriptors))

    return pvals, scores_original_data, h0_fmax

After Change

           intercept_test=intercept_test, random_state=0)
          for n_perm_chunk in n_perm_chunks)
    // reduce results
    scores_as_ranks_parts, h0_fmax_parts = zip(*ret)
    h0_fmax = np.hstack((h0_fmax_parts))
    scores_as_ranks = np.zeros((n_regressors, n_descriptors))
    for scores_as_ranks_part in scores_as_ranks_parts:

Instances