Skip to content

Commit

Permalink
Remove redundancies in test
Browse files Browse the repository at this point in the history
  • Loading branch information
@KarelZe
KarelZe committed Dec 22, 2023
1 parent 9a85119 commit 07c00bf
Showing 1 changed file with 79 additions and 130 deletions.
209 changes: 79 additions & 130 deletions tests/test_classical_classifier.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -259,6 +259,33 @@ def test_mid(self, x_train: pd.DataFrame, subset: str) -> None:
)
assert (y_pred == y_test).all()

def _apply_rule(
self,
x_train: pd.DataFrame,
x_test: pd.DataFrame,
y_test: pd.DataFrame,
layers: list[tuple[str, str]],
random_state: int = 7,
) -> None:
"""Apply rule-based classification.
Args:
x_train (pd.DataFrame): training features
x_test (pd.DataFrame): test features
y_test (pd.DataFrame): true labels
layers (list[tuple[str, str]]): layers
random_state (int, optional): random state. Defaults to 7.
"""
y_pred = (
ClassicalClassifier(
layers=layers,
random_state=random_state,
)
.fit(x_train[x_test.columns])
.predict(x_test)
)
assert (y_pred == y_test).all()

@pytest.mark.parametrize("subset", ["all", "ex"])
def test_tick_rule(self, x_train: pd.DataFrame, subset: str) -> None:
"""Test, if tick rule is correctly applied.
Expand All @@ -269,24 +296,14 @@ def test_tick_rule(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = ["trade_price", f"price_{subset}_lag"]

x_test = pd.DataFrame(
[[1, 2], [2, 1], [1, 1], [1, np.nan]],
columns=columns,
columns=["trade_price", f"price_{subset}_lag"],
)

# first two by rule (see p. 28 Grauer et al.), remaining two by random chance.
y_test = pd.Series([-1, 1, 1, -1])
y_pred = (
ClassicalClassifier(
layers=[("tick", subset)],
random_state=7,
)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("tick", subset)], 7)

@pytest.mark.parametrize("subset", ["all", "ex"])
def test_rev_tick_rule(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -298,21 +315,13 @@ def test_rev_tick_rule(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = ["trade_price", f"price_{subset}_lead"]

x_test = pd.DataFrame(
[[1, 2], [2, 1], [1, 1], [1, np.nan]],
columns=columns,
columns=["trade_price", f"price_{subset}_lead"],
)

# first two by rule (see p. 28 Grauer et al.), remaining two by random chance.
y_test = pd.Series([-1, 1, 1, -1])
y_pred = (
ClassicalClassifier(layers=[("rev_tick", subset)], random_state=7)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("rev_tick", subset)], 7)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_quote_rule(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -324,8 +333,6 @@ def test_quote_rule(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = ["trade_price", f"bid_{subset}", f"ask_{subset}"]

# first two by rule (see p. 28 Grauer et al.), remaining four by random chance.
x_test = pd.DataFrame(
[
Expand All @@ -336,15 +343,10 @@ def test_quote_rule(self, x_train: pd.DataFrame, subset: str) -> None:
[1, np.nan, 1],
[3, np.nan, np.nan],
],
columns=columns,
columns=["trade_price", f"bid_{subset}", f"ask_{subset}"],
)
y_test = pd.Series([-1, 1, 1, -1, -1, 1])
y_pred = (
ClassicalClassifier(layers=[("quote", subset)], random_state=45)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("quote", subset)], 45)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_lr(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -356,24 +358,18 @@ def test_lr(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = [
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lag",
]
# first two by quote rule, remaining two by tick rule.
x_test = pd.DataFrame(
[[1, 1, 3, 0], [3, 1, 3, 0], [1, 1, 1, 0], [3, 2, 4, 4]],
columns=columns,
columns=[
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lag",
],
)
y_test = pd.Series([-1, 1, 1, -1])
y_pred = (
ClassicalClassifier(layers=[("lr", subset)], random_state=7)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("lr", subset)], 7)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_rev_lr(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -385,12 +381,6 @@ def test_rev_lr(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = [
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lead",
]
# first two by quote rule, two by tick rule, and two by random chance.
x_test = pd.DataFrame(
[
Expand All @@ -401,15 +391,15 @@ def test_rev_lr(self, x_train: pd.DataFrame, subset: str) -> None:
[1, 1, np.nan, np.nan],
[1, 1, np.nan, np.nan],
],
columns=columns,
columns=[
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lead",
],
)
y_test = pd.Series([-1, 1, 1, -1, -1, 1])
y_pred = (
ClassicalClassifier(layers=[("rev_lr", subset)], random_state=42)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("rev_lr", subset)], 42)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_emo(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -421,12 +411,6 @@ def test_emo(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e.g., best
"""
columns = [
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lag",
]
# first two by quote rule, two by tick rule, two by random chance.
x_test = pd.DataFrame(
[
Expand All @@ -437,15 +421,15 @@ def test_emo(self, x_train: pd.DataFrame, subset: str) -> None:
[1, 1, np.inf, np.nan],
[1, 1, np.nan, np.nan],
],
columns=columns,
columns=[
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lag",
],
)
y_test = pd.Series([-1, 1, 1, -1, -1, 1])
y_pred = (
ClassicalClassifier(layers=[("emo", subset)], random_state=42)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("emo", subset)], 42)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_rev_emo(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -457,12 +441,6 @@ def test_rev_emo(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = [
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lead",
]
# first two by quote rule, two by tick rule, two by random chance.
x_test = pd.DataFrame(
[
Expand All @@ -473,15 +451,15 @@ def test_rev_emo(self, x_train: pd.DataFrame, subset: str) -> None:
[1, 1, np.inf, np.nan],
[1, 1, np.nan, np.nan],
],
columns=columns,
columns=[
"trade_price",
f"bid_{subset}",
f"ask_{subset}",
f"price_{subset}_lead",
],
)
y_test = pd.Series([-1, 1, 1, -1, -1, 1])
y_pred = (
ClassicalClassifier(layers=[("rev_emo", subset)], random_state=42)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("rev_emo", subset)], 42)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -493,12 +471,6 @@ def test_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = [
"trade_price",
f"ask_{subset}",
f"bid_{subset}",
f"price_{subset}_lag",
]
# first two by quote rule, two by tick rule, two by random chance.
x_test = pd.DataFrame(
[
Expand All @@ -509,15 +481,15 @@ def test_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
[1.7, 3, 1, 0], # tick rule
[1.3, 3, 1, 1], # quote rule
],
columns=columns,
columns=[
"trade_price",
f"ask_{subset}",
f"bid_{subset}",
f"price_{subset}_lag",
],
)
y_test = pd.Series([1, -1, 1, -1, 1, -1])
y_pred = (
ClassicalClassifier(layers=[("clnv", subset)], random_state=42)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("clnv", subset)], 42)

@pytest.mark.parametrize("subset", ["best", "ex"])
def test_rev_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
Expand All @@ -529,12 +501,6 @@ def test_rev_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
x_train (pd.DataFrame): training set
subset (str): subset e. g., 'ex'
"""
columns = [
"trade_price",
f"ask_{subset}",
f"bid_{subset}",
f"price_{subset}_lead",
]
x_test = pd.DataFrame(
[
[5, 3, 1, 0], # rev tick rule
Expand All @@ -552,19 +518,13 @@ def test_rev_clnv(self, x_train: pd.DataFrame, subset: str) -> None:
],
)
y_test = pd.Series([1, -1, 1, -1, 1, -1])
y_pred = (
ClassicalClassifier(layers=[("rev_clnv", subset)], random_state=5)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("rev_clnv", subset)], 5)

def test_trade_size(self, x_train: pd.DataFrame) -> None:
"""Test, if the trade size algorithm is correctly applied.
Tests cases where relevant data is present or missing.
"""
columns = ["trade_size", "ask_size_ex", "bid_size_ex"]
# first two by trade size, random, at bid size, random, random.
x_test = pd.DataFrame(
[
Expand All @@ -575,28 +535,16 @@ def test_trade_size(self, x_train: pd.DataFrame) -> None:
[1, np.inf, 2],
[1, np.inf, 2],
],
columns=columns,
columns=["trade_size", "ask_size_ex", "bid_size_ex"],
)
y_test = pd.Series([-1, 1, -1, 1, -1, 1])
y_pred = (
ClassicalClassifier(layers=[("trade_size", "ex")], random_state=42)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("trade_size", "ex")], 42)

def test_depth(self, x_train: pd.DataFrame) -> None:
"""Test, if the depth rule is correctly applied.
Tests cases where relevant data is present or missing.
"""
columns = [
"ask_size_ex",
"bid_size_ex",
"ask_ex",
"bid_ex",
"trade_price",
]
# first three by depth, all other random as mid is different from trade price.
x_test = pd.DataFrame(
[
Expand All @@ -606,12 +554,13 @@ def test_depth(self, x_train: pd.DataFrame) -> None:
[2, 1, 2, 4, 2],
[2, 1, 2, 4, 2],
],
columns=columns,
columns=[
"ask_size_ex",
"bid_size_ex",
"ask_ex",
"bid_ex",
"trade_price",
],
)
y_test = pd.Series([1, -1, 1, 1, -1])
y_pred = (
ClassicalClassifier(layers=[("depth", "ex")], random_state=5)
.fit(x_train[columns])
.predict(x_test)
)
assert (y_pred == y_test).all()
self._apply_rule(x_train, x_test, y_test, [("depth", "ex")], 5)

0 comments on commit 07c00bf

Please sign in to comment.