preprocessing.py 文件源码

def transform(self, X: pd.DataFrame, y=None) -> pd.DataFrame:
        is_dask = isinstance(X, dd.DataFrame)
        if is_dask:
            X = X.categorize()

        X = X.copy() if hasattr(X, 'copy') else X
        categories = self.cat_cols_
        for k in categories:
            cat = (categories.get(k, None)
                   if hasattr(categories, 'get')
                   else None)
            ordered = self.ordered.get(k, False)
            # can't use Categorical constructor since dask compat
            if not is_dask:
                X[k] = pd.Categorical(X[k])
            if cat:
                X[k] = X[k].cat.set_categories(cat)
            if ordered:
                X[k] = X[k].cat.as_ordered()
        return X