fix: LabelEncoder params consistent with Sklearn (#60)

* fix: LabelEncoder params consistent with Sklearn * fix:add LabelTransformer * fix: address comments for base LabelTransformer * fix: type for params

Author: ashleyxuu

bigframes/ml/base.py

@@ -195,3 +195,23 @@ def fit_transform(
  y: Optional[Union[bpd.DataFrame, bpd.Series]] = None,
  ) -> bpd.DataFrame:
  return self.fit(X, y).transform(X)
+
+
+class LabelTransformer(BaseEstimator):
+ """A BigQuery DataFrames Label Transformer base class that transforms data.
+
+ Also the transformers can be attached to a pipeline with a predictor."""
+
+ @abc.abstractmethod
+ def fit(self, y):
+ pass
+
+ @abc.abstractmethod
+ def transform(self, y):
+ pass
+
+ def fit_transform(
+ self,
+ y: Union[bpd.DataFrame, bpd.Series],
+ ) -> bpd.DataFrame:
+ return self.fit(y).transform(y)

bigframes/ml/preprocessing.py

@@ -315,7 +315,7 @@ def transform(self, X: Union[bpd.DataFrame, bpd.Series]) -> bpd.DataFrame:
 
 
 class LabelEncoder(
- base.Transformer,
+ base.LabelTransformer,
  third_party.bigframes_vendored.sklearn.preprocessing._label.LabelEncoder,
 ):
  # BQML max value https://cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-one-hot-encoder#syntax
@@ -401,16 +401,15 @@ def _parse_from_sql(cls, sql: str) -> tuple[LabelEncoder, str]:
 
  def fit(
  self,
- X: Union[bpd.DataFrame, bpd.Series],
- y=None, # ignored
+ y: Union[bpd.DataFrame, bpd.Series],
  ) -> LabelEncoder:
- (X,) = utils.convert_to_dataframe(X)
+ (y,) = utils.convert_to_dataframe(y)
 
- compiled_transforms = self._compile_to_sql(X.columns.tolist())
+ compiled_transforms = self._compile_to_sql(y.columns.tolist())
  transform_sqls = [transform_sql for transform_sql, _ in compiled_transforms]
 
  self._bqml_model = self._bqml_model_factory.create_model(
- X,
+ y,
  options={"model_type": "transform_only"},
  transforms=transform_sqls,
  )
@@ -419,13 +418,13 @@ def fit(
  self._output_names = [name for _, name in compiled_transforms]
  return self
 
- def transform(self, X: Union[bpd.DataFrame, bpd.Series]) -> bpd.DataFrame:
+ def transform(self, y: Union[bpd.DataFrame, bpd.Series]) -> bpd.DataFrame:
  if not self._bqml_model:
  raise RuntimeError("Must be fitted before transform")
 
- (X,) = utils.convert_to_dataframe(X)
+ (y,) = utils.convert_to_dataframe(y)
 
- df = self._bqml_model.transform(X)
+ df = self._bqml_model.transform(y)
  return typing.cast(
  bpd.DataFrame,
  df[self._output_names],

tests/system/small/ml/test_preprocessing.py

@@ -357,9 +357,9 @@ def test_one_hot_encoder_different_data(penguins_df_default_index, new_penguins_
 
 def test_label_encoder_default_params(new_penguins_df):
  encoder = bigframes.ml.preprocessing.LabelEncoder()
- encoder.fit(new_penguins_df[["species", "sex"]])
+ encoder.fit(new_penguins_df["species"])
 
- result = encoder.transform(new_penguins_df).to_pandas()
+ result = encoder.transform(new_penguins_df["species"]).to_pandas()
 
  # TODO: bug? feature columns seem to be in nondeterministic random order
  # workaround: sort columns by name. Can't repro it in pantheon, so could
@@ -368,11 +368,6 @@ def test_label_encoder_default_params(new_penguins_df):
 
  expected = pd.DataFrame(
  {
- "labelencoded_sex": [
- 2,
- 1,
- 1,
- ],
  "labelencoded_species": [
  1,
  1,
@@ -389,7 +384,7 @@ def test_label_encoder_default_params(new_penguins_df):
 def test_label_encoder_default_params_fit_transform(new_penguins_df):
  encoder = bigframes.ml.preprocessing.LabelEncoder()
 
- result = encoder.fit_transform(new_penguins_df[["species", "sex"]]).to_pandas()
+ result = encoder.fit_transform(new_penguins_df[["species"]]).to_pandas()
 
  # TODO: bug? feature columns seem to be in nondeterministic random order
  # workaround: sort columns by name. Can't repro it in pantheon, so could
@@ -398,11 +393,6 @@ def test_label_encoder_default_params_fit_transform(new_penguins_df):
 
  expected = pd.DataFrame(
  {
- "labelencoded_sex": [
- 2,
- 1,
- 1,
- ],
  "labelencoded_species": [
  1,
  1,
@@ -444,7 +434,7 @@ def test_label_encoder_series_default_params(new_penguins_df):
 
 def test_label_encoder_params(new_penguins_df):
  encoder = bigframes.ml.preprocessing.LabelEncoder(100, 2)
- encoder.fit(new_penguins_df[["species", "sex"]])
+ encoder.fit(new_penguins_df[["species"]])
 
  result = encoder.transform(new_penguins_df).to_pandas()
 
@@ -455,11 +445,6 @@ def test_label_encoder_params(new_penguins_df):
 
  expected = pd.DataFrame(
  {
- "labelencoded_sex": [
- 0,
- 0,
- 0,
- ],
  "labelencoded_species": [
  0,
  0,
@@ -475,7 +460,7 @@ def test_label_encoder_params(new_penguins_df):
 
 def test_label_encoder_different_data(penguins_df_default_index, new_penguins_df):
  encoder = bigframes.ml.preprocessing.LabelEncoder()
- encoder.fit(penguins_df_default_index[["species", "sex"]])
+ encoder.fit(penguins_df_default_index[["species"]])
 
  result = encoder.transform(new_penguins_df).to_pandas()
 
@@ -486,11 +471,6 @@ def test_label_encoder_different_data(penguins_df_default_index, new_penguins_df
 
  expected = pd.DataFrame(
  {
- "labelencoded_sex": [
- 3,
- 2,
- 2,
- ],
  "labelencoded_species": [
  1,
  1,

third_party/bigframes_vendored/sklearn/preprocessing/_label.py

@@ -28,23 +28,23 @@ class LabelEncoder(BaseEstimator):
  Default None, set limit to 1,000,000.
  """
 
- def fit(self, X):
- """Fit LabelEncoder to X.
+ def fit(self, y):
+ """Fit label encoder.
 
  Args:
- X (bigframes.dataframe.DataFrame or bigframes.series.Series):
+ y (bigframes.dataframe.DataFrame or bigframes.series.Series):
  The DataFrame or Series with training data.
 
  Returns:
  LabelEncoder: Fitted encoder.
  """
  raise NotImplementedError(constants.ABSTRACT_METHOD_ERROR_MESSAGE)
 
- def transform(self, X):
- """Transform X using label encoding.
+ def transform(self, y):
+ """Transform y using label encoding.
 
  Args:
- X (bigframes.dataframe.DataFrame or bigframes.series.Series):
+ y (bigframes.dataframe.DataFrame or bigframes.series.Series):
  The DataFrame or Series to be transformed.
 
  Returns: