How to use the rsmtool.preprocess.transform_feature function in rsmtool

def test_transform_feature():
    name = 'dpsec'
    data = np.array([1, 2, 3, 4])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'add_one_inverse')
        assert_array_equal(transform_feature(name, data, 'inv'), 1/data)
        assert_array_equal(transform_feature(name, data, 'raw'), data)
        assert_array_equal(transform_feature(name, data, 'org'), data)
        assert_array_equal(transform_feature(name, data, 'log'), np.log(data))
        assert_array_equal(transform_feature(name, data, 'addOneInv'), 1/(data+1))
        assert_array_equal(transform_feature(name, data, 'addOneLn'), np.log(data+1))

def test_transform_feature():
    name = 'dpsec'
    data = np.array([1, 2, 3, 4])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'add_one_inverse')
        assert_array_equal(transform_feature(name, data, 'inv'), 1/data)
        assert_array_equal(transform_feature(name, data, 'raw'), data)
        assert_array_equal(transform_feature(name, data, 'org'), data)
        assert_array_equal(transform_feature(name, data, 'log'), np.log(data))
        assert_array_equal(transform_feature(name, data, 'addOneInv'), 1/(data+1))
        assert_array_equal(transform_feature(name, data, 'addOneLn'), np.log(data+1))

def test_transform_feature_with_warning():
    name = 'dpsec'
    data = np.array([-1, 0, 2, 3])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_array_equal(transform_feature(name, data, 'sqrt', raise_error=False),
                           np.sqrt(data))
        assert_array_equal(transform_feature(name, data, 'inv', raise_error=False),
                           1/data)
        assert_array_equal(transform_feature(name, data, 'addOneInv', raise_error=False),
                           1/(data+1))        
        assert_array_equal(transform_feature(name, data, 'log', raise_error=False),
                           np.log(data))        
        assert_array_equal(transform_feature(name, data, 'addOneLn', raise_error=False),
                           np.log(data+1))

def test_transform_feature_with_warning():
    name = 'dpsec'
    data = np.array([-1, 0, 2, 3])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_array_equal(transform_feature(name, data, 'sqrt', raise_error=False),
                           np.sqrt(data))
        assert_array_equal(transform_feature(name, data, 'inv', raise_error=False),
                           1/data)
        assert_array_equal(transform_feature(name, data, 'addOneInv', raise_error=False),
                           1/(data+1))        
        assert_array_equal(transform_feature(name, data, 'log', raise_error=False),
                           np.log(data))        
        assert_array_equal(transform_feature(name, data, 'addOneLn', raise_error=False),
                           np.log(data+1))

def test_transform_feature():
    name = 'dpsec'
    data = np.array([1, 2, 3, 4])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'add_one_inverse')
        assert_array_equal(transform_feature(name, data, 'inv'), 1/data)
        assert_array_equal(transform_feature(name, data, 'raw'), data)
        assert_array_equal(transform_feature(name, data, 'org'), data)
        assert_array_equal(transform_feature(name, data, 'log'), np.log(data))
        assert_array_equal(transform_feature(name, data, 'addOneInv'), 1/(data+1))
        assert_array_equal(transform_feature(name, data, 'addOneLn'), np.log(data+1))

def test_transform_feature():
    name = 'dpsec'
    data = np.array([1, 2, 3, 4])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'add_one_inverse')
        assert_array_equal(transform_feature(name, data, 'inv'), 1/data)
        assert_array_equal(transform_feature(name, data, 'raw'), data)
        assert_array_equal(transform_feature(name, data, 'org'), data)
        assert_array_equal(transform_feature(name, data, 'log'), np.log(data))
        assert_array_equal(transform_feature(name, data, 'addOneInv'), 1/(data+1))
        assert_array_equal(transform_feature(name, data, 'addOneLn'), np.log(data+1))

def test_transform_feature_with_error():
    name = 'dpsec'
    data = np.array([-1, 0, 2, 3])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'sqrt')
        assert_raises(ValueError, transform_feature, name, data, 'inv')
        assert_raises(ValueError, transform_feature, name, data, 'addOneInv')
        assert_raises(ValueError, transform_feature, name, data, 'log')
        assert_raises(ValueError, transform_feature, name, data, 'addOneLn')

def test_transform_feature_with_warning():
    name = 'dpsec'
    data = np.array([-1, 0, 2, 3])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_array_equal(transform_feature(name, data, 'sqrt', raise_error=False),
                           np.sqrt(data))
        assert_array_equal(transform_feature(name, data, 'inv', raise_error=False),
                           1/data)
        assert_array_equal(transform_feature(name, data, 'addOneInv', raise_error=False),
                           1/(data+1))        
        assert_array_equal(transform_feature(name, data, 'log', raise_error=False),
                           np.log(data))        
        assert_array_equal(transform_feature(name, data, 'addOneLn', raise_error=False),
                           np.log(data+1))

def test_transform_feature():
    name = 'dpsec'
    data = np.array([1, 2, 3, 4])
    # run the test but suppress the expected runtime warnings
    with warnings.catch_warnings():
        warnings.filterwarnings('ignore', category=RuntimeWarning)
        assert_raises(ValueError, transform_feature, name, data, 'add_one_inverse')
        assert_array_equal(transform_feature(name, data, 'inv'), 1/data)
        assert_array_equal(transform_feature(name, data, 'raw'), data)
        assert_array_equal(transform_feature(name, data, 'org'), data)
        assert_array_equal(transform_feature(name, data, 'log'), np.log(data))
        assert_array_equal(transform_feature(name, data, 'addOneInv'), 1/(data+1))
        assert_array_equal(transform_feature(name, data, 'addOneLn'), np.log(data+1))

between the feature values and the human scores. See
        :ref:`documentation ` for the
        full list of transformations.
    """

    # Do not use sqrt and ln for potential negative features.
    # Do not use inv for positive features.
    if any(feature_value < 0):
        applicable_transformations = ['org', 'inv']
    else:
        applicable_transformations = ['org', 'sqrt', 'addOneInv', 'addOneLn']

    correlations = []
    for trans in applicable_transformations:
        try:
            transformed_value = transform_feature(feature_name, feature_value, trans)
            correlations.append(abs(pearsonr(transformed_value, scores)[0]))
        except ValueError:
            # If the transformation returns an error, append 0.
            correlations.append(0)
    best = np.argmax(correlations)
    best_transformation = applicable_transformations[best]
    return best_transformation