feat: Add ml.metrics.pairwise.manhattan_distance

Author: ashleyxuu

bigframes/ml/metrics/pairwise.py

@@ -32,3 +32,19 @@ def cosine_similarity(
 
 
 cosine_similarity.__doc__ = inspect.getdoc(vendored_metrics_pairwise.cosine_similarity)
+
+
+def manhattan_distance(
+    X: Union[bpd.DataFrame, bpd.Series], Y: Union[bpd.DataFrame, bpd.Series]
+) -> bpd.DataFrame:
+    X, Y = utils.convert_to_dataframe(X, Y)
+    if len(X.columns) != 1 or len(Y.columns) != 1:
+        raise ValueError("Inputs X and Y can only contain 1 column.")
+
+    base_bqml = core.BaseBqml(session=X._session)
+    return base_bqml.distance(X, Y, type="MANHATTAN", name="manhattan_distance")
+
+
+manhattan_distance.__doc__ = inspect.getdoc(
+    vendored_metrics_pairwise.manhattan_distance
+)

tests/system/small/ml/test_metrics_pairwise.py

@@ -33,3 +33,17 @@ def test_cosine_similarity():
     pd.testing.assert_frame_equal(
         result.to_pandas(), expected_pd_df, check_dtype=False, check_index_type=False
     )
+
+
+def test_manhattan_distance():
+    x_col = [np.array([4.1, 0.5, 1.0])]
+    y_col = [np.array([3.0, 0.0, 2.5])]
+    X = bpd.read_pandas(pd.DataFrame({"X": x_col}))
+    Y = bpd.read_pandas(pd.DataFrame({"Y": y_col}))
+
+    result = metrics.pairwise.manhattan_distance(X, Y)
+    expected_pd_df = pd.DataFrame({"X": x_col, "Y": y_col, "manhattan_distance": [3.1]})
+
+    pd.testing.assert_frame_equal(
+        result.to_pandas(), expected_pd_df, check_dtype=False, check_index_type=False
+    )

third_party/bigframes_vendored/sklearn/metrics/pairwise.py

@@ -29,3 +29,18 @@ def cosine_similarity(X, Y) -> bpd.DataFrame:
         bigframes.dataframe.DataFrame: DataFrame with columns of X, Y and cosine_similarity
     """
     raise NotImplementedError(constants.ABSTRACT_METHOD_ERROR_MESSAGE)
+
+
+def manhattan_distance(X, Y) -> bpd.DataFrame:
+    """Compute the L1 distances between the vectors in X and Y.
+
+    Args:
+        X (Series or single column DataFrame of array of numeric type):
+            Input data.
+        Y (Series or single column DataFrame of array of numeric type):
+            Input data. X and Y are mapped by indexes, must have the same index.
+
+    Returns:
+        bigframes.dataframe.DataFrame: DataFrame with columns of X, Y and manhattan_distance
+    """
+    raise NotImplementedError(constants.ABSTRACT_METHOD_ERROR_MESSAGE)