def test_random_state():
df = pd.DataFrame(np.arange(1000).reshape(100, 10))
# test passes if function raises a TypeError
with pytest.raises(TypeError):
params = {"X": df, "test_size": 0.2, "random_state": "hello"}
split = Split()
split.train_test_split(params=params)
def test_without_target_col():
df = pd.DataFrame(np.arange(1000).reshape(100, 10))
params = {"train_df": df, "test_size": 0.2, "random_state": 420}
split = Split()
split.train_test_split(params=params)
assert params["X_train"].shape[0] == 80
assert params["X_train"].shape[1] == 10
assert params["X_test"].shape[0] == 20
assert params["X_test"].shape[1] == 10
def test_without_target_label():
df_x = pd.DataFrame(np.arange(1000).reshape(100, 10))
df_y = pd.Series(np.arange(100))
with pytest.raises(ValueError):
params = {
"X": df_x,
"y": df_y,
"test_size": 0.2,
"random_state": 420,
}
split = Split()
split.train_test_split(params=params)
def __init__(
self,
train_df_path=None,
test_df_path=None,
steps=None,
config_file=None,
params=None,
custom_reader=None,
):
steps = [
Parser().parse_dataset,
NullValuesHandler().execute,
Encoder().encode,
HandleOutlier().handle_outliers,
Scaler().execute,
SelectKBest().fit_transform,
Split().train_test_split,
]
super().__init__(
train_df_path=train_df_path,
test_df_path=test_df_path,
steps=steps,
config_file=config_file,
params=params,
custom_reader=custom_reader,
)
import numpy as np
import pandas as pd
import pytest
from preprocessy.resampling import Split
split = Split()
class TestSplitting:
def test_without_target_label(self):
df = pd.DataFrame(np.arange(1000).reshape(100, 10))
train, test = split.train_test_split(X=df,
test_size=0.2,
random_state=420)
assert train.shape[0] == 80
assert train.shape[1] == 10
assert test.shape[0] == 20
assert test.shape[1] == 10
def test_random_state(self):
df = pd.DataFrame(np.arange(1000).reshape(100, 10))
# test passes if function raises a TypeError
with pytest.raises(TypeError):
train, test = split.train_test_split(X=df,
test_size=0.2,
random_state="hello")