cleanup

Signed-off-by: Anatoly Myachev <[email protected]>

modin/core/storage_formats/base/query_compiler.py

@@ -22,7 +22,7 @@
 import abc
 import warnings
 from functools import cached_property
-from typing import TYPE_CHECKING, Hashable, List, Optional
+from typing import Hashable, List, Optional
 
 import numpy as np
 import pandas
@@ -53,10 +53,6 @@
 
 from . import doc_utils
 
-if TYPE_CHECKING:
-    # TODO: should be ModinDataframe
-    from modin.core.dataframe.pandas.dataframe.dataframe import PandasDataframe
-
 
 def _get_axis(axis):
     """
@@ -131,8 +127,6 @@ class BaseQueryCompiler(
     for a list of requirements for subclassing this object.
     """
 
-    _modin_frame: PandasDataframe
-
     def __wrap_in_qc(self, obj):
         """
         Wrap `obj` in query compiler.

modin/core/storage_formats/pandas/query_compiler.py

@@ -945,7 +945,7 @@ def compute_dtypes_fn(dtypes, axis, **kwargs):
             ):
                 return np.object_
             # how to take into account backend here?
-            return "float64"
+            return np.float64
 
         return TreeReduce.register(
             map_fn,
@@ -2141,7 +2141,7 @@ def searchsorted(df):
     dt_month_name = Map.register(_dt_func_map("month_name"), dtypes=np.object_)
     dt_day_name = Map.register(_dt_func_map("day_name"), dtypes=np.object_)
     dt_to_pytimedelta = Map.register(_dt_func_map("to_pytimedelta"), dtypes=np.object_)
-    dt_total_seconds = Map.register(_dt_func_map("total_seconds"), dtypes="float64")
+    dt_total_seconds = Map.register(_dt_func_map("total_seconds"), dtypes=np.float64)
     dt_seconds = Map.register(_dt_prop_map("seconds"), dtypes=np.int64)
     dt_days = Map.register(_dt_prop_map("days"), dtypes=np.int64)
     dt_microseconds = Map.register(_dt_prop_map("microseconds"), dtypes=np.int64)
@@ -2323,7 +2323,7 @@ def map_func(df):  # pragma: no cover
             # Does it work with pyarrow backend?
             df_mask = np.isfinite(df)
 
-            result = np.empty((n_rows, n_cols), dtype="float64")
+            result = np.empty((n_rows, n_cols), dtype=np.float64)
 
             for i in range(n_rows):
                 df_ith_row = df[i]
@@ -2679,7 +2679,7 @@ def quantile_builder(df, **kwargs):
             lambda df: quantile_builder(df, **kwargs),
             new_index=q_index,
             new_columns=new_columns,
-            dtypes="float64",
+            dtypes=np.float64,
         )
         result = self.__constructor__(new_modin_frame)
         return result.transpose() if axis == 1 else result
@@ -2696,7 +2696,7 @@ def rank(self, **kwargs):
                 if not numeric_only
                 else None
             ),
-            dtypes="float64",
+            dtypes=np.float64,
             sync_labels=False,
         )
         return self.__constructor__(new_modin_frame)

modin/pandas/dataframe.py

@@ -1623,7 +1623,7 @@ def prod(
             and numeric_only is False
             and min_count > len(axis_to_apply)
             # Type inference is not so simple for pyarrow
-            and self._query_compiler.get_backend() == "default"
+            and self._query_compiler.get_backend() is not None
         ):
             new_index = self.columns if not axis else self.index
             # >>> pd.DataFrame([1,2,3,4], dtype="int64[pyarrow]").prod(min_count=10)
@@ -2153,7 +2153,7 @@ def sum(
             and numeric_only is False
             and min_count > len(axis_to_apply)
             # Type inference is not so simple for pyarrow
-            and self._query_compiler.get_backend() == "default"
+            and self._query_compiler.get_backend() is not None
         ):
             new_index = self.columns if not axis else self.index
             return Series(

modin/tests/pandas/utils.py

@@ -1088,11 +1088,7 @@ def eval_io_from_str(csv_str: str, unique_filename: str, **kwargs):
 
 
 def create_test_dfs(*args, **kwargs) -> tuple[pd.DataFrame, pandas.DataFrame]:
-    post_fn = kwargs.pop("post_fn", None)
-
-    if post_fn is None:
-        # TODO: REVERT ME
-        post_fn = lambda df: df.convert_dtypes(dtype_backend="pyarrow")  # noqa: E731
+    post_fn = kwargs.pop("post_fn", lambda df: df)
     return tuple(
         map(post_fn, [pd.DataFrame(*args, **kwargs), pandas.DataFrame(*args, **kwargs)])
     )
@@ -1108,10 +1104,6 @@ def create_test_series(vals, sort=False, **kwargs) -> tuple[pd.Series, pandas.Se
     if sort:
         modin_series = modin_series.sort_values().reset_index(drop=True)
         pandas_series = pandas_series.sort_values().reset_index(drop=True)
-
-    # TODO: REVERT ME
-    modin_series = modin_series.convert_dtypes(dtype_backend="pyarrow")
-    pandas_series = pandas_series.convert_dtypes(dtype_backend="pyarrow")
     return modin_series, pandas_series