def test_df_key():
"""Raise error if dataframe key is not supplied."""
df = pd.DataFrame({"x": [2, 3]})
others = {"df": pd.DataFrame({"x": range(1, 6), "y": [5, 4, 3, 2, 1]})}
with pytest.raises(KeyError):
expand_grid(df, others=others)
def test_lists_not_all_scalar():
"""
Trigger error if values in a list
in the dictionary's values are not scalar.
"""
data = {"x": [[2, 3], 4, 5, "ragnar"]}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_series_multi_index():
"""Test that multiIndexed series trigger error."""
data = {
"x": pd.Series(
[2, 3], index=pd.MultiIndex.from_arrays([[1, 2], [3, 4]])
)
}
with pytest.raises(TypeError):
expand_grid(others=data)
def test_dataframe_multi_index_index_and_column():
"""Trigger error if dataframe has a MultiIndex column or index."""
data = {
"x": pd.DataFrame(
[[2, 3], [6, 7]],
index=pd.MultiIndex.from_arrays([["a", "b"], ["y", "z"]]),
columns=pd.MultiIndex.from_arrays([["m", "n"], ["p", "q"]]),
)
}
with pytest.raises(TypeError):
expand_grid(others=data)
def test_df_multi_index():
"""Test that datafarme is not a multiIndex."""
df = {
"x": pd.DataFrame(
[[2, 3], [6, 7]],
columns=pd.MultiIndex.from_arrays([["m", "n"], ["p", "q"]]),
)
}
others = [5, 4, 3, 2, 1]
with pytest.raises(TypeError):
expand_grid(df, others=others)
def test_lists():
"""
Test expected output from one level nested lists in a dictionary's values.
"""
data = {"x": [[2, 3], [4, 3]]}
expected = pd.DataFrame({"x": [[2, 3], [4, 3]]})
assert_frame_equal(expand_grid(others=data), expected)
def test_computation_output_4():
"""Test output from list of strings."""
data = {"l1": list(ascii_lowercase[:3]), "l2": list(ascii_uppercase[:3])}
expected = pd.DataFrame(
{
"l1": ["a", "a", "a", "b", "b", "b", "c", "c", "c"],
"l2": ["A", "B", "C", "A", "B", "C", "A", "B", "C"],
}
)
assert_frame_equal(expand_grid(others=data), expected)
def test_computation_output_3():
"""Test mix of dataframes and lists."""
data = {
"df": pd.DataFrame({"x": range(1, 3), "y": [2, 1]}),
"z": range(1, 4),
}
expected = pd.DataFrame(
{
"df_x": [1, 1, 1, 2, 2, 2],
"df_y": [2, 2, 2, 1, 1, 1],
"z": [1, 2, 3, 1, 2, 3],
}
)
assert_frame_equal(expand_grid(others=data), expected)
def test_computation_output_2():
"""Test output if entry contains only dataframes/series."""
data = {
"df": pd.DataFrame({"x": range(1, 6), "y": [5, 4, 3, 2, 1]}),
"df1": pd.DataFrame({"x": range(4, 7), "y": [6, 5, 4]}),
}
expected = pd.DataFrame(
{
"df_x": [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4, 5, 5, 5],
"df_y": [5, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 1, 1, 1],
"df1_x": [4, 5, 6, 4, 5, 6, 4, 5, 6, 4, 5, 6, 4, 5, 6],
"df1_y": [6, 5, 4, 6, 5, 4, 6, 5, 4, 6, 5, 4, 6, 5, 4],
}
)
assert_frame_equal(expand_grid(others=data), expected)
def test_df_output():
"""
Test output from chaining method to a dataframe.
Example is from tidyverse's expand_grid page -
https://tidyr.tidyverse.org/reference/expand_grid.html#compared-to-expand-grid
"""
df = pd.DataFrame({"x": range(1, 3), "y": [2, 1]})
others = {"z": range(1, 4)}
expected = pd.DataFrame(
{
"df_x": [1, 1, 1, 2, 2, 2],
"df_y": [2, 2, 2, 1, 1, 1],
"z": [1, 2, 3, 1, 2, 3],
}
)
result = expand_grid(df, df_key="df", others=others)
assert_frame_equal(result, expected)
def test_2d_arrays_multiple_columns():
"""Test on 2d arrays with multiple columns and rows."""
# example referenced from tidyr page
# https://tidyr.tidyverse.org/reference/expand_grid.html
data = {
"x": np.reshape(np.arange(1, 5), (2, -1), order="F"),
"y": np.reshape(np.arange(5, 9), (2, -1), order="F"),
}
expected = pd.DataFrame(
{
"x_0": [1, 1, 2, 2],
"x_1": [3, 3, 4, 4],
"y_0": [5, 6, 5, 6],
"y_1": [7, 8, 7, 8],
}
)
assert_frame_equal(expand_grid(others=data), expected)
def test_computation_output_1():
"""Test output if entry contains no dataframes/series."""
data = {"x": range(1, 4), "y": [1, 2]}
expected = pd.DataFrame({"x": [1, 1, 2, 2, 3, 3], "y": [1, 2, 1, 2, 1, 2]})
# h/t to @hectormz for picking it up
# check_dtype is set to False for this test -
# on Windows systems integer dtype default is int32
# while on Ubuntu it comes up as int64
# when the test is executed, the dtype on the left is int32
# while the expected dataframe has a dtype of int64
# And this was causing this particular test to fail on Windows
# pandas has a minimum of int64 for integer columns
# My suspicion is that because the resulting dataframe was
# created solely from a dictionary via numpy
# pandas simply picked up the dtype supplied from numpy
# whereas the expected dataframe was created within Pandas
# and got assigned the minimum dtype of int64
# hence the error and the need to set check_dtype to False
assert_frame_equal(expand_grid(others=data), expected, check_dtype=False)
def test_numpy():
"""Raise error if numpy array in dictionary's values is empty."""
data = {"x": np.array([])}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_nested_dict():
"""Raise error if dictionary is nested in a dictionary's values."""
data = {"x": {"y": 2}}
with pytest.raises(TypeError):
expand_grid(others=data)
def test_dataframe_empty():
"""Trigger error for empty dataframes."""
data = {"x": pd.DataFrame([])}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_df_others():
"""Raise error if others is not a dict."""
df = pd.DataFrame({"x": [2, 3]})
others = [5, 4, 3, 2, 1]
with pytest.raises(TypeError):
expand_grid(df, others=others)
def test_dataframe_single_index():
"""Test for single indexed dataframes."""
data = {"x": pd.DataFrame([[2, 3], [6, 7]])}
expected = pd.DataFrame([[2, 3], [6, 7]]).add_prefix("x_")
assert_frame_equal(expand_grid(others=data), expected)
def test_not_a_dict_1():
"""Test that entry (dataframe) is not a dictionary."""
data = pd.DataFrame([60, 70])
with pytest.raises(TypeError):
expand_grid(others=data)
def test_series_not_multi_index_with_name():
"""Test for single index series with name."""
data = {"x": pd.Series([2, 3], name="y")}
expected = pd.DataFrame([2, 3], columns=["x_y"])
assert_frame_equal(expand_grid(others=data), expected)
def test_numpy_gt_2d():
"""Raise error if numpy array dimension is greater than 2."""
data = {"x": np.array([[[2, 3]]])}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_numpy_1d():
"""Test output from a 1d numpy array."""
data = {"x": np.array([2, 3])}
expected = pd.DataFrame(np.array([2, 3]), columns=["x_0"])
assert_frame_equal(expand_grid(others=data), expected)
def test_list_empty():
"""Raise error if any list in dictionary's values is empty."""
data = {"x": [], "y": [2, 3]}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_numpy_2d():
"""Test output from a 2d numpy array."""
data = {"x": np.array([[2, 3]])}
expected = pd.DataFrame(np.array([[2, 3]])).add_prefix("x_")
assert_frame_equal(expand_grid(others=data), expected)
def test_empty_dict():
"""Test that entry should not be empty."""
data = {}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_series_empty():
"""Test that values in key value pair should not be empty for Series."""
data = {"x": pd.Series([], dtype="int")}
with pytest.raises(ValueError):
expand_grid(others=data)
def test_not_a_dict():
"""Test that entry(list) is not a dictionary."""
data = [60, 70]
with pytest.raises(TypeError):
expand_grid(others=data)
