Adding a periodic encoder to the DatetimeEncoder

CHANGES.rst

@@ -12,12 +12,15 @@ Ongoing development
 New features
 ------------
 
-- The :class:`TableReport` now switch it's visual theme between light and dark according to the user preferences.
+- The :class:`TableReport` now switches its visual theme between light and dark according to the user preferences.
   :pr:`1201` by :user:`rouk1 <rouk1>`.
 
 - Adding a new way to control the location of the data directory, using envar `SKRUB_DATA_DIRECTORY`.
   :pr:`1215` by :user:`Thomas S. <thomass-dev>`
 
+- The :class:`DatetimeEncoder` now supports periodic encoding of the features using circular (sine/cosine) and spline
+  transformers. :pr:`1235` by :user:`Riccardo Cappuzzo<rcap107>`.
+
 Changes
 -------
 

skrub/__init__.py

@@ -6,7 +6,7 @@
 from ._agg_joiner import AggJoiner, AggTarget
 from ._check_dependencies import check_dependencies
 from ._column_associations import column_associations
-from ._datetime_encoder import DatetimeEncoder
+from ._datetime_encoder import CircularEncoder, DatetimeEncoder, SplineEncoder
 from ._deduplicate import compute_ngram_distance, deduplicate
 from ._fuzzy_join import fuzzy_join
 from ._gap_encoder import GapEncoder
@@ -56,4 +56,6 @@
     "TextEncoder",
     "StringEncoder",
     "column_associations",
+    "SplineEncoder",
+    "CircularEncoder",
 ]

skrub/_dataframe/_common.py

@@ -97,6 +97,7 @@
     "unique",
     "filter",
     "where",
+    "where_row",
     "sample",
     "head",
     "slice",
@@ -1182,6 +1183,23 @@ def _where_polars(col, mask, other):
     return col.zip_with(mask, pl.Series(other))
 
 
+@dispatch
+def where_row(obj, mask, other):
+    raise NotImplementedError()
+
+
+@where_row.specialize("pandas")
+def _where_row_pandas(obj, mask, other):
+    return obj.apply(pd.Series.where, **{"cond": mask, "other": other})
+
+
+@where_row.specialize("polars")
+def _where_row_polars(obj, mask, other):
+    return obj.with_columns(
+        pl.when(pl.Series(mask)).then(pl.all()).otherwise(pl.Series(other))
+    )
+
+
 @dispatch
 def sample(obj, n, seed=None):
     raise NotImplementedError()

skrub/_dataframe/tests/test_common.py

@@ -778,6 +778,24 @@ def test_where(df_module):
     )
 
 
+def test_where_row(df_module):
+    df = df_module.make_dataframe({"col1": [1, 2, 3], "col2": [1000, 2000, 3000]})
+    out = ns.where_row(
+        df,
+        df_module.make_column("", [False, True, False]),  # mask
+        df_module.make_column(
+            "", [None, None, None]
+        ),  # values to put in on the entire row
+    )
+    right = df_module.make_dataframe(
+        {"col1": [None, 2, None], "col2": [None, 2000, None]}
+    )
+    df_module.assert_frame_equal(
+        ns.pandas_convert_dtypes(out),
+        ns.pandas_convert_dtypes(right),
+    )
+
+
 def test_sample(df_module):
     s = ns.pandas_convert_dtypes(df_module.make_column("", [0, 1, 2]))
     sample = ns.sample(s, 2)