def test_dataframe(self, data):
"""helper.split_dataframe: DataFrame output"""
s = helper.split_dataframe(data, "split", type="DataFrame")
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, pd.DataFrame)
pd.testing.assert_frame_equal(d1, d2)
def test_array(self, data):
"""helper.split_dataframe: array output"""
s = helper.split_dataframe(data, "split", type="array")
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, np.ndarray)
np.testing.assert_array_equal(d1, d2.values)
def plot(file=None):
ax.cla()
d = state["trc"].loc[f_sel.value]
for p, trc in helper.split_dataframe(d, ("fret", "particle"),
[("donor", "x"),
("donor", "y")]):
ax.plot(trc[:, 0], trc[:, 1])
ax.text(*trc[0], str(p))
fig.canvas.draw()
def test_dataframe(self):
"""helpers.pandas.split_dataframe: DataFrame output"""
s = split_dataframe(self.data, "split", type="DataFrame")
expected = list(self.data.groupby("split"))
self.assertEqual(len(s), len(expected))
for (i1, d1), (i2, d2) in zip(s, expected):
self.assertEqual(i1, i2)
self.assertIsInstance(d1, pd.DataFrame)
pd.testing.assert_frame_equal(d1, d2)
def test_array_select_columns(self):
"""helpers.pandas.split_dataframe: selected columns, array output"""
s = split_dataframe(self.data, "split", ["B", "A"], type="array")
expected = list(self.data.groupby("split"))
self.assertEqual(len(s), len(expected))
for (i1, d1), (i2, d2) in zip(s, expected):
self.assertEqual(i1, i2)
self.assertIsInstance(d1, np.ndarray)
np.testing.assert_equal(d1, d2[["B", "A"]].values)
def test_array_ambiguous_index(self):
"""helpers.pandas.split_dataframe: ambiguous index, array output"""
self.data.index = np.full(len(self.data), 1)
s = split_dataframe(self.data, "split", type="array")
expected = list(self.data.groupby("split"))
self.assertEqual(len(s), len(expected))
for (i1, d1), (i2, d2) in zip(s, expected):
self.assertEqual(i1, i2)
self.assertIsInstance(d1, np.ndarray)
np.testing.assert_equal(d1, d2.values)
def test_array_multiindex(self, data):
"""helper.split_dataframe: multiindex, array output"""
d = pd.concat([data, data], axis=1, keys=["x", "y"])
s = helper.split_dataframe(d, ("x", "split"), [("y", "B"), ("x", "A")],
type="array")
expected = list(d.groupby([("x", "split")]))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, np.ndarray)
np.testing.assert_equal(d1, d2[[("y", "B"), ("x", "A")]].values)
def test_array_multiindex(self):
"""helpers.pandas.split_dataframe: multiindex, array output"""
d = pd.concat([self.data, self.data], axis=1, keys=["x", "y"])
s = split_dataframe(d, ("x", "split"), [("y", "B"), ("x", "A")],
type="array")
expected = list(d.groupby([("x", "split")]))
self.assertEqual(len(s), len(expected))
for (i1, d1), (i2, d2) in zip(s, expected):
self.assertEqual(i1, i2)
self.assertIsInstance(d1, np.ndarray)
np.testing.assert_equal(d1, d2[[("y", "B"), ("x", "A")]].values)
def test_array_list(self, data):
"""helper.split_dataframe: list of arrays output"""
data["boolean"] = True
s = helper.split_dataframe(data, "split", type="array_list")
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, list)
assert len(d1) == d2.shape[1]
for c1, (_, c2) in zip(d1, d2.items()):
np.testing.assert_array_equal(c1, c2.values)
def test_array_list_select_columns(self, data):
"""helper.split_dataframe: selected columns, list of arrays output"""
cols = ["B", "A"]
s = helper.split_dataframe(data, "split", cols, type="array_list")
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, list)
assert len(d1) == len(cols)
for c1, c2_name in zip(d1, cols):
np.testing.assert_array_equal(c1, d2[c2_name].values)
def test_array_keep_index(self):
"""helpers.pandas.split_dataframe: keep_index=True, array output"""
s = split_dataframe(self.data, "split", type="array", keep_index=True)
expected = list(self.data.groupby("split"))
self.assertEqual(len(s), len(expected))
for i, e in expected:
e.insert(0, "index", e.index)
for (i1, d1), (i2, d2) in zip(s, expected):
self.assertEqual(i1, i2)
self.assertIsInstance(d1, np.ndarray)
np.testing.assert_equal(d1, d2.values)
def test_array_list_multiindex(self, data):
"""helper.split_dataframe: multiindex, list of arrays output"""
cols = [("y", "B"), ("x", "A")]
d = pd.concat([data, data], axis=1, keys=["x", "y"])
s = helper.split_dataframe(d, ("x", "split"), cols, type="array_list")
expected = list(d.groupby([("x", "split")]))
assert len(s) == len(expected)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, list)
assert len(d1) == len(cols)
for c1, c2_name in zip(d1, cols):
np.testing.assert_array_equal(c1, d2[c2_name].values)
def test_array_keep_index(self, data):
"""helper.split_dataframe: keep_index=True, array output"""
s = helper.split_dataframe(data,
"split",
type="array",
keep_index=True)
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for i, e in expected:
e.insert(0, "index", e.index)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, np.ndarray)
np.testing.assert_array_equal(d1, d2.values)
def test_array_list_keep_index(self, data):
"""helper.split_dataframe: keep_index=True, list of arrays output"""
s = helper.split_dataframe(data,
"split",
type="array_list",
keep_index=True)
expected = list(data.groupby("split"))
assert len(s) == len(expected)
for i, e in expected:
e.insert(0, "index", e.index)
for (i1, d1), (i2, d2) in zip(s, expected):
assert i1 == i2
assert isinstance(d1, list)
assert len(d1) == d2.shape[1]
for c1, (_, c2) in zip(d1, d2.items()):
np.testing.assert_array_equal(c1, c2.values)
