actually user should handle null values in their own datagit status

cthorrez · Nov 1, 2024 · fcc85de · fcc85de
1 parent 6c8bd53
commit fcc85de
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 5 deletions.
diff --git a/riix/eval.py b/riix/eval.py
@@ -35,7 +35,6 @@ def grid_search(
     metric='log_loss',
     minimize_metric=True,
     num_processes=None,
-    seed=0,
     return_all_metrics=False,
 ):
     """Perform grid search and return the best hyperparameters."""

diff --git a/riix/utils/data_utils.py b/riix/utils/data_utils.py
@@ -47,8 +47,8 @@ def __init__(
         # Create a single competitors reference dataframe
         competitors_df = pl.DataFrame(
             {'competitor': pl.concat([
-                df[competitor_cols[0]].cast(pl.Utf8).fill_null(''),
-                df[competitor_cols[1]].cast(pl.Utf8).fill_null('')
+                df[competitor_cols[0]].cast(pl.Utf8),
+                df[competitor_cols[1]].cast(pl.Utf8)
             ]).unique().sort()
         }).lazy().select(
             pl.all(),
@@ -59,8 +59,8 @@ def __init__(
         self.competitor_to_idx = dict(zip(self.competitors, range(self.num_competitors)))
         matchups_df = (df.lazy()
             .select([
-                pl.col(competitor_cols[0]).cast(pl.Utf8).fill_null('').alias('comp1'),
-                pl.col(competitor_cols[1]).cast(pl.Utf8).fill_null('').alias('comp2')
+                pl.col(competitor_cols[0]).cast(pl.Utf8).alias('comp1'),
+                pl.col(competitor_cols[1]).cast(pl.Utf8).alias('comp2')
             ])
             .join(
                 competitors_df,