Rename LabelLeakageDataCheck to TargetLeakageDataCheck. (#1319)

* Rename LabelLeakageDataCheck to TargetLeakageDataCheck.

* Adding PR 1319 to relase notes.

* Sorting imports in default_data_checks.py

* Renaming tests from _label_ to _target_

docs/source/api_reference.rst

@@ -435,7 +435,7 @@ Data Check Classes
     InvalidTargetDataCheck
     HighlyNullDataCheck
     IDColumnsDataCheck
-    LabelLeakageDataCheck
+    TargetLeakageDataCheck
     OutliersDataCheck
     NoVarianceDataCheck
     ClassImbalanceDataCheck

docs/source/release_notes.rst

@@ -29,6 +29,11 @@ Release Notes
         * Cleaned up ``make_pipeline`` tests to test for all estimators :pr:`1257`
         * Added a test to check conda build after merge to main :pr:`1247`
 
+.. warning::
+
+    **Breaking Changes**
+        * Renamed ``LabelLeakageDataCheck`` to ``TargetLeakageDataCheck`` :pr:`1319`
+
 
 **v0.14.1 Sep. 29, 2020**
     * Enhancements

evalml/data_checks/__init__.py

@@ -7,7 +7,7 @@
 from .invalid_targets_data_check import InvalidTargetDataCheck
 from .highly_null_data_check import HighlyNullDataCheck
 from .id_columns_data_check import IDColumnsDataCheck
-from .label_leakage_data_check import LabelLeakageDataCheck
+from .target_leakage_data_check import TargetLeakageDataCheck
 from .outliers_data_check import OutliersDataCheck
 from .no_variance_data_check import NoVarianceDataCheck
 from .class_imbalance_data_check import ClassImbalanceDataCheck

evalml/data_checks/default_data_checks.py

@@ -2,17 +2,17 @@
 from .highly_null_data_check import HighlyNullDataCheck
 from .id_columns_data_check import IDColumnsDataCheck
 from .invalid_targets_data_check import InvalidTargetDataCheck
-from .label_leakage_data_check import LabelLeakageDataCheck
 from .no_variance_data_check import NoVarianceDataCheck
+from .target_leakage_data_check import TargetLeakageDataCheck
 
 
 class DefaultDataChecks(DataChecks):
     """A collection of basic data checks that is used by AutoML by default.
-    Includes HighlyNullDataCheck, IDColumnsDataCheck, LabelLeakageDataCheck, InvalidTargetDataCheck,
+    Includes HighlyNullDataCheck, IDColumnsDataCheck, TargetLeakageDataCheck, InvalidTargetDataCheck,
     and NoVarianceDataCheck."""
 
     _DEFAULT_DATA_CHECK_CLASSES = [HighlyNullDataCheck, IDColumnsDataCheck,
-                                   LabelLeakageDataCheck, InvalidTargetDataCheck, NoVarianceDataCheck]
+                                   TargetLeakageDataCheck, InvalidTargetDataCheck, NoVarianceDataCheck]
 
     def __init__(self, problem_type):
         """

evalml/data_checks/label_leakage_data_check.py → evalml/data_checks/target_leakage_data_check.py

@@ -6,7 +6,7 @@
 from evalml.utils.gen_utils import numeric_and_boolean_dtypes
 
 
-class LabelLeakageDataCheck(DataCheck):
+class TargetLeakageDataCheck(DataCheck):
     """Check if any of the features are highly correlated with the target."""
 
     def __init__(self, pct_corr_threshold=0.95):
@@ -32,7 +32,7 @@ def validate(self, X, y):
             y (pd.Series): The target data
 
         Returns:
-            list (DataCheckWarning): List with a DataCheckWarning if there is label leakage detected.
+            list (DataCheckWarning): List with a DataCheckWarning if target leakage is detected.
 
         Example:
             >>> X = pd.DataFrame({
@@ -41,8 +41,8 @@ def validate(self, X, y):
             ...    'y': [12, 5, 13, 74, 24],
             ... })
             >>> y = pd.Series([10, 42, 31, 51, 40])
-            >>> label_leakage_check = LabelLeakageDataCheck(pct_corr_threshold=0.8)
-            >>> assert label_leakage_check.validate(X, y) == [DataCheckWarning("Column 'leak' is 80.0% or more correlated with the target", "LabelLeakageDataCheck")]
+            >>> target_leakage_check = TargetLeakageDataCheck(pct_corr_threshold=0.8)
+            >>> assert target_leakage_check.validate(X, y) == [DataCheckWarning("Column 'leak' is 80.0% or more correlated with the target", "TargetLeakageDataCheck")]
         """
         if not isinstance(X, pd.DataFrame):
             X = pd.DataFrame(X)

evalml/tests/data_checks_tests/test_data_checks.py

@@ -64,7 +64,7 @@ def test_default_data_checks_classification():
     y = pd.Series([0, 1, np.nan, 1, 0])
     data_checks = DefaultDataChecks("binary")
 
-    leakage = [DataCheckWarning("Column 'has_label_leakage' is 95.0% or more correlated with the target", "LabelLeakageDataCheck")]
+    leakage = [DataCheckWarning("Column 'has_label_leakage' is 95.0% or more correlated with the target", "TargetLeakageDataCheck")]
 
     assert data_checks.validate(X, y) == messages[:3] + leakage + messages[3:]
 

evalml/tests/data_checks_tests/test_target_leakage_data_check.py

@@ -0,0 +1,71 @@
+import pandas as pd
+import pytest
+
+from evalml.data_checks.data_check_message import DataCheckWarning
+from evalml.data_checks.target_leakage_data_check import TargetLeakageDataCheck
+
+
+def test_target_leakage_data_check_init():
+    target_leakage_check = TargetLeakageDataCheck()
+    assert target_leakage_check.pct_corr_threshold == 0.95
+
+    target_leakage_check = TargetLeakageDataCheck(pct_corr_threshold=0.0)
+    assert target_leakage_check.pct_corr_threshold == 0
+
+    target_leakage_check = TargetLeakageDataCheck(pct_corr_threshold=0.5)
+    assert target_leakage_check.pct_corr_threshold == 0.5
+
+    target_leakage_check = TargetLeakageDataCheck(pct_corr_threshold=1.0)
+    assert target_leakage_check.pct_corr_threshold == 1.0
+
+    with pytest.raises(ValueError, match="pct_corr_threshold must be a float between 0 and 1, inclusive."):
+        TargetLeakageDataCheck(pct_corr_threshold=-0.1)
+    with pytest.raises(ValueError, match="pct_corr_threshold must be a float between 0 and 1, inclusive."):
+        TargetLeakageDataCheck(pct_corr_threshold=1.1)
+
+
+def test_target_leakage_data_check_warnings():
+    y = pd.Series([1, 0, 1, 1])
+    X = pd.DataFrame()
+    X["a"] = y * 3
+    X["b"] = y - 1
+    X["c"] = y / 10
+    X["d"] = ~y
+    X["e"] = [0, 0, 0, 0]
+    y = y.astype(bool)
+
+    leakage_check = TargetLeakageDataCheck(pct_corr_threshold=0.5)
+    assert leakage_check.validate(X, y) == [DataCheckWarning("Column 'a' is 50.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                            DataCheckWarning("Column 'b' is 50.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                            DataCheckWarning("Column 'c' is 50.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                            DataCheckWarning("Column 'd' is 50.0% or more correlated with the target", "TargetLeakageDataCheck")]
+
+
+def test_target_leakage_data_check_input_formats():
+    leakage_check = TargetLeakageDataCheck(pct_corr_threshold=0.8)
+
+    # test empty pd.DataFrame, empty pd.Series
+    assert leakage_check.validate(pd.DataFrame(), pd.Series()) == []
+
+    y = pd.Series([1, 0, 1, 1])
+    X = pd.DataFrame()
+    X["a"] = y * 3
+    X["b"] = y - 1
+    X["c"] = y / 10
+    X["d"] = ~y
+    X["e"] = [0, 0, 0, 0]
+    y = y.astype(bool)
+
+    expected_messages = [DataCheckWarning("Column 'a' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                         DataCheckWarning("Column 'b' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                         DataCheckWarning("Column 'c' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                         DataCheckWarning("Column 'd' is 80.0% or more correlated with the target", "TargetLeakageDataCheck")]
+
+    #  test y as list
+    assert leakage_check.validate(X, y.values) == expected_messages
+
+    # test X as np.array
+    assert leakage_check.validate(X.to_numpy(), y) == [DataCheckWarning("Column '0' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                                       DataCheckWarning("Column '1' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                                       DataCheckWarning("Column '2' is 80.0% or more correlated with the target", "TargetLeakageDataCheck"),
+                                                       DataCheckWarning("Column '3' is 80.0% or more correlated with the target", "TargetLeakageDataCheck")]