def test_categorical_encoder_unfitted_fail():
unfitted_categorical_encoder = column_encoders.CategoricalEncoder(
["col_1"])
assert not unfitted_categorical_encoder.is_fitted()
with pytest.raises(column_encoders.NotFittedError):
unfitted_categorical_encoder.transform(
pd.DataFrame({"col_1": ['a', 'b']}))
def test_categorical_encoder_numeric_nan():
df = pd.DataFrame({'brand': [1, 2, 3, None]})
try:
column_encoders.CategoricalEncoder("brand").fit(df)
except TypeError:
pytest.fail(
"fitting categorical encoder on integers with nulls should not fail"
)
def test_categorical_encoder_max_token():
categorical_encoder = column_encoders.CategoricalEncoder(
['labels'], max_tokens=1e4).fit(df)
assert categorical_encoder.max_tokens == 2
def test_categorical_encoder_numeric_transform():
df = pd.DataFrame({'brand': [1, 2, 3, 1, 2, 1, np.nan, None]})
col_enc = column_encoders.CategoricalEncoder("brand").fit(df)
assert np.array_equal(col_enc.transform(df),
np.array([[1], [2], [3], [1], [2], [1], [0], [0]]))
import numpy as np
import pandas as pd
import pytest
from datawig import column_encoders
df = pd.DataFrame({
'features': [
'xwcxG pQldP Cel0n 5LaWO 2cjTu', '2cjTu YizDY u1aEa Cel0n SntTK',
'2cjTu YizDY u1aEa Cel0n SntTK'
],
'labels': ['xwcxG', 'SntTK', 'SntTK']
})
categorical_encoder = column_encoders.CategoricalEncoder(['labels'],
max_tokens=3).fit(df)
sequential_encoder = column_encoders.SequentialEncoder(['features'],
max_tokens=50,
seq_len=3).fit(df)
# CategoricalEncoder Tests
def test_categorical_encoder_unfitted_fail():
unfitted_categorical_encoder = column_encoders.CategoricalEncoder(
["col_1"])
assert not unfitted_categorical_encoder.is_fitted()
with pytest.raises(column_encoders.NotFittedError):
unfitted_categorical_encoder.transform(
pd.DataFrame({"col_1": ['a', 'b']}))