id
stringlengths
30
32
content
stringlengths
139
2.8k
codereview_new_python_data_10667
def test_repartition(axis, dtype): if dtype == "DataFrame": results = { - None: ([4, 0, 0, 0], [3, 0, 0, 0]), - 0: ([4, 0, 0, 0], [2, 1]), - 1: ([2, 2], [3, 0, 0, 0]), } else: results = { - None: ([4, 0, 0, 0], [1]), - 0: ([4, 0, 0, 0], [1]), 1: ([2, 2], [1]), } # in some cases empty partitions aren't filtered so it's better to use IIUC empty partitions are not filtered out only in cases when there's not enough metadata to do so: https://github.com/modin-project/modin/blob/f137564496ceff581a312de9af5ecbd143c7ee32/modin/core/dataframe/pandas/dataframe/dataframe.py#L471-L478 `._repartition` seems to have all the required metadata in the end, why does this change needed then? def test_repartition(axis, dtype): if dtype == "DataFrame": results = { + None: ([4], [3]), + 0: ([4], [2, 1]), + 1: ([2, 2], [3]), } else: results = { + None: ([4], [1]), + 0: ([4], [1]), 1: ([2, 2], [1]), } # in some cases empty partitions aren't filtered so it's better to use
codereview_new_python_data_10668
def apply_full_axis( Setting it to True disables shuffling data from one partition to another. synchronize : boolean, default: True Synchronize external indexes (`new_index`, `new_columns`) with internal indexes. Returns ------- ```suggestion Synchronize external indexes (`new_index`, `new_columns`) with internal indexes. This could be used when you're certain that the indices in partitions are equal to the provided hints in order to save time on syncing them. ``` def apply_full_axis( Setting it to True disables shuffling data from one partition to another. synchronize : boolean, default: True Synchronize external indexes (`new_index`, `new_columns`) with internal indexes. + This could be used when you're certain that the indices in partitions are equal to + the provided hints in order to save time on syncing them. Returns -------
codereview_new_python_data_10669
def repartition(self, axis=None): new_index=self._modin_frame._index_cache, new_columns=self._modin_frame._columns_cache, keep_partitioning=False, - synchronize=False, ) ) return new_query_compiler ```suggestion sync_labels=False, ``` def repartition(self, axis=None): new_index=self._modin_frame._index_cache, new_columns=self._modin_frame._columns_cache, keep_partitioning=False, + sync_labels=False, ) ) return new_query_compiler
codereview_new_python_data_10670
def check_parameters_support( if read_kwargs.get("skipfooter"): if read_kwargs.get("nrows") or read_kwargs.get("engine") == "c": - return (False, "raise exception by pandas itself") skiprows_supported = True if is_list_like(skiprows_md) and skiprows_md[0] < header_size: ```suggestion return (False, "Exception is raised by pandas itself") ``` def check_parameters_support( if read_kwargs.get("skipfooter"): if read_kwargs.get("nrows") or read_kwargs.get("engine") == "c": + return (False, "Exception is raised by pandas itself") skiprows_supported = True if is_list_like(skiprows_md) and skiprows_md[0] < header_size:
codereview_new_python_data_10671
def _read_csv_check_support( ) if read_csv_kwargs.get("skipfooter") and read_csv_kwargs.get("nrows"): - return (False, "raise exception by pandas itself") for arg, def_value in cls.read_csv_unsup_defaults.items(): if read_csv_kwargs[arg] != def_value: ```suggestion return (False, "Exception is raised by pandas itself") ``` def _read_csv_check_support( ) if read_csv_kwargs.get("skipfooter") and read_csv_kwargs.get("nrows"): + return (False, "Exception is raised by pandas itself") for arg, def_value in cls.read_csv_unsup_defaults.items(): if read_csv_kwargs[arg] != def_value:
codereview_new_python_data_10672
def _read(cls, io, **kwargs): if index_col is not None: column_names = column_names.drop(column_names[index_col]) - if not all(column_names) or kwargs["usecols"]: # some column names are empty, use pandas reader to take the names from it pandas_kw["nrows"] = 1 df = pandas.read_excel(io, **pandas_kw) Previously, there was no processing of `usecols` parameter in dispatcher. However, for the correct calculation of the columns, we can use pandas, just as we do for `read_csv`. def _read(cls, io, **kwargs): if index_col is not None: column_names = column_names.drop(column_names[index_col]) + if not all(column_names) or kwargs.get("usecols"): # some column names are empty, use pandas reader to take the names from it pandas_kw["nrows"] = 1 df = pandas.read_excel(io, **pandas_kw)
codereview_new_python_data_10673
def _read(cls, io, **kwargs): if index_col is not None: column_names = column_names.drop(column_names[index_col]) - if not all(column_names) or kwargs["usecols"]: # some column names are empty, use pandas reader to take the names from it pandas_kw["nrows"] = 1 df = pandas.read_excel(io, **pandas_kw) ```suggestion if not all(column_names) or kwargs.get("usecols"): ``` ? def _read(cls, io, **kwargs): if index_col is not None: column_names = column_names.drop(column_names[index_col]) + if not all(column_names) or kwargs.get("usecols"): # some column names are empty, use pandas reader to take the names from it pandas_kw["nrows"] = 1 df = pandas.read_excel(io, **pandas_kw)
codereview_new_python_data_10674
def modin_df_almost_equals_pandas(modin_df, pandas_df): ) -def try_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator dtypes1, dtypes2 = map( ```suggestion def try_modin_df_almost_equals_compare(df1, df2): ``` def modin_df_almost_equals_pandas(modin_df, pandas_df): ) +def try_modin_df_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator dtypes1, dtypes2 = map(
codereview_new_python_data_10675
def walk( elif depth == 1: if len(value) != 2: raise ValueError( - f"Incorrect rename format. Renamer must consist of exactly two elements, got {len(value)=}." ) func_name, func = value yield key, func, func_name, col_renaming_required ```suggestion f"Incorrect rename format. Renamer must consist of exactly two elements, got: {len(value)}." ``` def walk( elif depth == 1: if len(value) != 2: raise ValueError( + f"Incorrect rename format. Renamer must consist of exactly two elements, got: {len(value)}." ) func_name, func = value yield key, func, func_name, col_renaming_required
codereview_new_python_data_10676
def try_modin_df_almost_equals_compare(df1, df2): if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( is_numeric_dtype(dtype) for dtype in dtypes2 ): - return modin_df_almost_equals_pandas(df1, df2) - else: - return df_equals(df1, df2) def df_is_empty(df): To fix CodeQL warning: ```suggestion modin_df_almost_equals_pandas(df1, df2) ``` def try_modin_df_almost_equals_compare(df1, df2): if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( is_numeric_dtype(dtype) for dtype in dtypes2 ): + modin_df_almost_equals_pandas(df1, df2) + df_equals(df1, df2) def df_is_empty(df):
codereview_new_python_data_10677
def try_modin_df_almost_equals_compare(df1, df2): if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( is_numeric_dtype(dtype) for dtype in dtypes2 ): - return modin_df_almost_equals_pandas(df1, df2) - else: - return df_equals(df1, df2) def df_is_empty(df): `df_equals` this function does not return anything. ```suggestion df_equals(df1, df2) ``` def try_modin_df_almost_equals_compare(df1, df2): if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( is_numeric_dtype(dtype) for dtype in dtypes2 ): + modin_df_almost_equals_pandas(df1, df2) + df_equals(df1, df2) def df_is_empty(df):
codereview_new_python_data_10678
def try_modin_df_almost_equals_compare(df1, df2): is_numeric_dtype(dtype) for dtype in dtypes2 ): modin_df_almost_equals_pandas(df1, df2) df_equals(df1, df2) ```suggestion modin_df_almost_equals_pandas(df1, df2) else: df_equals(df1, df2) ``` def try_modin_df_almost_equals_compare(df1, df2): is_numeric_dtype(dtype) for dtype in dtypes2 ): modin_df_almost_equals_pandas(df1, df2) + else: df_equals(df1, df2)
codereview_new_python_data_10679
def groupby_agg( # Defaulting to pandas in case of an empty frame as we can't process it properly. # Higher API level won't pass empty data here unless the frame has delayed # computations. So we apparently lose some laziness here (due to index access) - # because of the disability to process empty groupby natively. if len(self.columns) == 0 or len(self.index) == 0: return super().groupby_agg( by, agg_func, axis, groupby_kwargs, agg_args, agg_kwargs, how, drop ```suggestion # because of the inability to process empty groupby natively. ``` def groupby_agg( # Defaulting to pandas in case of an empty frame as we can't process it properly. # Higher API level won't pass empty data here unless the frame has delayed # computations. So we apparently lose some laziness here (due to index access) + # because of the inability to process empty groupby natively. if len(self.columns) == 0 or len(self.index) == 0: return super().groupby_agg( by, agg_func, axis, groupby_kwargs, agg_args, agg_kwargs, how, drop
codereview_new_python_data_10680
def modin_df_almost_equals_pandas(modin_df, pandas_df): def try_modin_df_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator - dtypes1, dtypes2 = map( - lambda df: dtype if is_list_like(dtype := df.dtypes) else [dtype], (df1, df2) - ) - if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( - is_numeric_dtype(dtype) for dtype in dtypes2 - ): modin_df_almost_equals_pandas(df1, df2) else: df_equals(df1, df2) ```suggestion dtypes1, dtypes2 = [dtype if is_list_like(dtype := df.dtypes) else [dtype] for df in (df1, df2)] ``` ?.. def modin_df_almost_equals_pandas(modin_df, pandas_df): def try_modin_df_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator + dtypes1, dtypes2 = [ + dtype if is_list_like(dtype := df.dtypes) else [dtype] for df in (df1, df2) + ] + if all(map(is_numeric_dtype, dtypes1)) and all(map(is_numeric_dtype, dtypes2)): modin_df_almost_equals_pandas(df1, df2) else: df_equals(df1, df2)
codereview_new_python_data_10681
def modin_df_almost_equals_pandas(modin_df, pandas_df): def try_modin_df_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator - dtypes1, dtypes2 = map( - lambda df: dtype if is_list_like(dtype := df.dtypes) else [dtype], (df1, df2) - ) - if all(is_numeric_dtype(dtype) for dtype in dtypes1) and all( - is_numeric_dtype(dtype) for dtype in dtypes2 - ): modin_df_almost_equals_pandas(df1, df2) else: df_equals(df1, df2) ```suggestion if all(map(is_numeric_dtype, dtypes1)) and all(map(is_numeric_dtype, dtypes2)): ``` ? def modin_df_almost_equals_pandas(modin_df, pandas_df): def try_modin_df_almost_equals_compare(df1, df2): """Compare two dataframes as nearly equal if possible, otherwise compare as completely equal.""" # `modin_df_almost_equals_pandas` is numeric-only comparator + dtypes1, dtypes2 = [ + dtype if is_list_like(dtype := df.dtypes) else [dtype] for df in (df1, df2) + ] + if all(map(is_numeric_dtype, dtypes1)) and all(map(is_numeric_dtype, dtypes2)): modin_df_almost_equals_pandas(df1, df2) else: df_equals(df1, df2)
codereview_new_python_data_10682
def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) -@pytest.mark.parametrize("converter", [int, float]) -def test___int__and__float__(converter, count_elements): eval_general( *create_test_series(test_data["int_data"]), - lambda df: converter(df[:count_elements]), ) I would rather we test int and float methods separately. We have a test for bool separately as well. def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) +def test___int__(count_elements): + eval_general( + *create_test_series(test_data["float_data"]), + lambda df: int(df[:count_elements]), + ) + + +@pytest.mark.parametrize("count_elements", [0, 1, 10]) +def test___float__(count_elements): eval_general( *create_test_series(test_data["int_data"]), + lambda df: float(df[:count_elements]), )
codereview_new_python_data_10683
def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): eval_general( - *create_test_series(test_data["float_nan_data"]), - lambda df: int(df[:count_elements]), ) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): eval_general( - *create_test_series(test_data["int_data"]), - lambda df: float(df[:count_elements]), ) Actually, I have two comments here. 1) Why do we need to test a series with the number of elements other than 1? 2) If a series consist of one np.nan, we just check exception equality. Should we test both a valid number and np.nan in the series? def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): eval_general( + *create_test_series([1.5] * count_elements), + lambda df: int(df), ) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): eval_general( + *create_test_series([1] * count_elements), + lambda df: float(df), )
codereview_new_python_data_10684
def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): - eval_general( - *create_test_series([1.5] * count_elements), - lambda df: int(df), - ) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): - eval_general( - *create_test_series([1] * count_elements), - lambda df: float(df), - ) @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys) ## Unnecessary lambda This 'lambda' is just a simple wrapper around a callable object. Use that object directly. [Show more details](https://github.com/modin-project/modin/security/code-scanning/496) def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): + eval_general(*create_test_series([1.5] * count_elements), int) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): + eval_general(*create_test_series([1] * count_elements), float) @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
codereview_new_python_data_10685
def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): - eval_general( - *create_test_series([1.5] * count_elements), - lambda df: int(df), - ) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): - eval_general( - *create_test_series([1] * count_elements), - lambda df: float(df), - ) @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys) ## Unnecessary lambda This 'lambda' is just a simple wrapper around a callable object. Use that object directly. [Show more details](https://github.com/modin-project/modin/security/code-scanning/497) def test___gt__(data): @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___int__(count_elements): + eval_general(*create_test_series([1.5] * count_elements), int) @pytest.mark.parametrize("count_elements", [0, 1, 10]) def test___float__(count_elements): + eval_general(*create_test_series([1] * count_elements), float) @pytest.mark.parametrize("data", test_data_values, ids=test_data_keys)
codereview_new_python_data_10686
def show_versions(as_json: Union[str, bool] = False) -> None: print(f"{k:<{maxlen}}: {v}") -def int_to_float32(dtype: np.dtype) -> np.dtype: """ Check if a datatype is a variant of integer. - If dtype is integer function returns float32 datatype if not returns the argument datatype itself Parameters Should this be `int_to_float64`? def show_versions(as_json: Union[str, bool] = False) -> None: print(f"{k:<{maxlen}}: {v}") +def int_to_float64(dtype: np.dtype) -> np.dtype: """ Check if a datatype is a variant of integer. + If dtype is integer function returns float64 datatype if not returns the argument datatype itself Parameters
codereview_new_python_data_10687
def show_versions(as_json: Union[str, bool] = False) -> None: print(f"{k:<{maxlen}}: {v}") -def int_to_float32(dtype: np.dtype) -> np.dtype: """ Check if a datatype is a variant of integer. - If dtype is integer function returns float32 datatype if not returns the argument datatype itself Parameters ```suggestion If dtype is integer function returns float64 datatype if not returns the ``` def show_versions(as_json: Union[str, bool] = False) -> None: print(f"{k:<{maxlen}}: {v}") +def int_to_float64(dtype: np.dtype) -> np.dtype: """ Check if a datatype is a variant of integer. + If dtype is integer function returns float64 datatype if not returns the argument datatype itself Parameters
codereview_new_python_data_10688
def compute_dtypes_common_cast(first, second) -> np.dtype: ----- The dtypes of the operands are supposed to be known. """ - dtypes_first = dict(zip(first.columns, first._modin_frame._dtypes)) - dtypes_second = dict(zip(second.columns, second._modin_frame._dtypes)) columns_first = set(first.columns) columns_second = set(second.columns) common_columns = columns_first.intersection(columns_second) ```suggestion dtypes_first = first._modin_frame._dtypes.to_dict() dtypes_second = second._modin_frame._dtypes.to_dict() ``` or even we can remove this at all and use just this further. ```python dtypes_first = first._modin_frame._dtypes dtypes_first.loc[col] ``` def compute_dtypes_common_cast(first, second) -> np.dtype: ----- The dtypes of the operands are supposed to be known. """ + dtypes_first = first._modin_frame._dtypes.to_dict() + dtypes_second = second._modin_frame._dtypes.to_dict() columns_first = set(first.columns) columns_second = set(second.columns) common_columns = columns_first.intersection(columns_second)
codereview_new_python_data_10689
def values(self): # noqa: RT01, D200 """ import modin.pandas as pd - if isinstance(self.dtype, pandas.core.dtypes.dtypes.ExtensionDtype): return self._default_to_pandas("values") data = self.to_numpy() Modin doesn't have `core` module. def values(self): # noqa: RT01, D200 """ import modin.pandas as pd + if isinstance( + self.dtype, pandas.core.dtypes.dtypes.ExtensionDtype + ) and not isinstance(self.dtype, pd.CategoricalDtype): return self._default_to_pandas("values") data = self.to_numpy()
codereview_new_python_data_10690
def partitioned_file( List with the next elements: int : partition start read byte int : partition end read byte - pandas.DataFrame Dataframe from which metadata can be retrieved. Can be None if `read_callback_kw=None`. """ if read_callback_kw is not None and pre_reading != 0: ```suggestion pandas.DataFrame or None ``` def partitioned_file( List with the next elements: int : partition start read byte int : partition end read byte + pandas.DataFrame or None Dataframe from which metadata can be retrieved. Can be None if `read_callback_kw=None`. """ if read_callback_kw is not None and pre_reading != 0:
codereview_new_python_data_10691
def default_to_pandas(self, pandas_op, *args, **kwargs): """ op_name = getattr(pandas_op, "__name__", str(pandas_op)) ErrorMessage.default_to_pandas(op_name) - args = (a.to_pandas() if isinstance(a, type(self)) else a for a in args) - kwargs = { - k: v.to_pandas() if isinstance(v, type(self)) else v - for k, v in kwargs.items() - } - result = pandas_op(self.to_pandas(), *args, **kwargs) if isinstance(result, pandas.Series): if result.name is None: result.name = MODIN_UNNAMED_SERIES_LABEL let's use more comprehensive approach ```suggestion args = try_cast_to_pandas(args) kwargs = try_cast_to_pandas(kwargs) ``` https://github.com/modin-project/modin/blob/1a57a542ad0413fae2591d4b4928b652effd01db/modin/utils.py#L480 def default_to_pandas(self, pandas_op, *args, **kwargs): """ op_name = getattr(pandas_op, "__name__", str(pandas_op)) ErrorMessage.default_to_pandas(op_name) + args = try_cast_to_pandas(args) + kwargs = try_cast_to_pandas(kwargs) + result = pandas_op(try_cast_to_pandas(self), *args, **kwargs) if isinstance(result, pandas.Series): if result.name is None: result.name = MODIN_UNNAMED_SERIES_LABEL
codereview_new_python_data_10692
def default_to_pandas(self, pandas_op, *args, **kwargs): """ op_name = getattr(pandas_op, "__name__", str(pandas_op)) ErrorMessage.default_to_pandas(op_name) - args = (a.to_pandas() if isinstance(a, type(self)) else a for a in args) - kwargs = { - k: v.to_pandas() if isinstance(v, type(self)) else v - for k, v in kwargs.items() - } - result = pandas_op(self.to_pandas(), *args, **kwargs) if isinstance(result, pandas.Series): if result.name is None: result.name = MODIN_UNNAMED_SERIES_LABEL ```suggestion result = pandas_op(try_cast_to_pandas(self), *args, **kwargs) ``` def default_to_pandas(self, pandas_op, *args, **kwargs): """ op_name = getattr(pandas_op, "__name__", str(pandas_op)) ErrorMessage.default_to_pandas(op_name) + args = try_cast_to_pandas(args) + kwargs = try_cast_to_pandas(kwargs) + result = pandas_op(try_cast_to_pandas(self), *args, **kwargs) if isinstance(result, pandas.Series): if result.name is None: result.name = MODIN_UNNAMED_SERIES_LABEL
codereview_new_python_data_10693
def from_labels(self) -> "PandasDataframe": if "index" not in self.columns else "level_{}".format(0) ] - names = tuple(level_names) if len(level_names) > 1 else level_names[0] - new_dtypes = self.index.to_frame(name=names).dtypes - new_dtypes = pandas.concat([new_dtypes, self.dtypes]) # We will also use the `new_column_names` in the calculation of the internal metadata, so this is a # lightweight way of ensuring the metadata matches. I'm wondering should we do it all the time or only when the `self._dtypes` cache is accessible? def from_labels(self) -> "PandasDataframe": if "index" not in self.columns else "level_{}".format(0) ] + new_dtypes = None + if self._dtypes is not None: + names = tuple(level_names) if len(level_names) > 1 else level_names[0] + new_dtypes = self.index.to_frame(name=names).dtypes + new_dtypes = pandas.concat([new_dtypes, self._dtypes]) # We will also use the `new_column_names` in the calculation of the internal metadata, so this is a # lightweight way of ensuring the metadata matches.
codereview_new_python_data_10694
def from_labels(self) -> "PandasDataframe": if "index" not in self.columns else "level_{}".format(0) ] - names = tuple(level_names) if len(level_names) > 1 else level_names[0] - new_dtypes = self.index.to_frame(name=names).dtypes - new_dtypes = pandas.concat([new_dtypes, self.dtypes]) # We will also use the `new_column_names` in the calculation of the internal metadata, so this is a # lightweight way of ensuring the metadata matches. ```suggestion new_dtypes = None if self._dtypes is not None: names = tuple(level_names) if len(level_names) > 1 else level_names[0] new_dtypes = self.index.to_frame(name=names).dtypes new_dtypes = pandas.concat([new_dtypes, self._dtypes]) ``` def from_labels(self) -> "PandasDataframe": if "index" not in self.columns else "level_{}".format(0) ] + new_dtypes = None + if self._dtypes is not None: + names = tuple(level_names) if len(level_names) > 1 else level_names[0] + new_dtypes = self.index.to_frame(name=names).dtypes + new_dtypes = pandas.concat([new_dtypes, self._dtypes]) # We will also use the `new_column_names` in the calculation of the internal metadata, so this is a # lightweight way of ensuring the metadata matches.
codereview_new_python_data_10695
def test_mean_with_datetime(by_func): def test_groupby_mad_warn(): - modin_df = pd.DataFrame(test_groupby_data) md_grp = modin_df.groupby(by=modin_df.columns[0]) msg = "The 'mad' method is deprecated and will be removed in a future version." - with pytest.warns(FutureWarning, match=msg): - md_grp.mad() def test_groupby_backfill_warn(): I wish we would test against pandas, too, to make sure we warn the same thing def test_mean_with_datetime(by_func): def test_groupby_mad_warn(): + modin_df, pandas_df = create_test_dfs(test_groupby_data) md_grp = modin_df.groupby(by=modin_df.columns[0]) + pd_grp = pandas_df.groupby(by=pandas_df.columns[0]) msg = "The 'mad' method is deprecated and will be removed in a future version." + for grp_obj in (md_grp, pd_grp): + with pytest.warns(FutureWarning, match=msg): + grp_obj.mad() def test_groupby_backfill_warn():
codereview_new_python_data_10696
def values(self): # noqa: RT01, D200 Return Series as ndarray or ndarray-like depending on the dtype. """ data = self.to_numpy() - if isinstance(self.dtype, pandas.CategoricalDtype): - data = pandas.Categorical(data, dtype=self.dtype) return data def add(self, other, level=None, fill_value=None, axis=0): # noqa: PR01, RT01, D200 Can we use "our own" categorical class somehow? I know that `modin.pandas.Categorical` is just an alias for the pandas class, but maybe it would change in the future and so then we wouldn't need to rename this all over the project. ```suggestion if isinstance(self.dtype, pd.CategoricalDtype): data = pd.Categorical(data, dtype=self.dtype) ``` def values(self): # noqa: RT01, D200 Return Series as ndarray or ndarray-like depending on the dtype. """ data = self.to_numpy() + if isinstance(self.dtype, pd.CategoricalDtype): + data = pd.Categorical(data, dtype=self.dtype) return data def add(self, other, level=None, fill_value=None, axis=0): # noqa: PR01, RT01, D200
codereview_new_python_data_10697
import sys from ipykernel import kernelspec default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd -def new_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments -kernelspec.make_ipkernel_cmd = new_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi" ```suggestion def custom_make_ipkernel_cmd(): mpiexec_args = ["mpiexec", "-n", "1"] default_args = default_make_ipkernel_cmd() return mpiexec_args + default_args ``` import sys from ipykernel import kernelspec + default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd +def custom_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments +kernelspec.make_ipkernel_cmd = custom_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi"
codereview_new_python_data_10698
import sys from ipykernel import kernelspec default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd -def new_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments -kernelspec.make_ipkernel_cmd = new_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi" ```suggestion kernelspec.make_ipkernel_cmd = custom_make_ipkernel_cmd ``` import sys from ipykernel import kernelspec + default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd +def custom_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments +kernelspec.make_ipkernel_cmd = custom_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi"
codereview_new_python_data_10699
import sys from ipykernel import kernelspec default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd -def new_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments -kernelspec.make_ipkernel_cmd = new_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi" Add a licence. import sys from ipykernel import kernelspec + default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd +def custom_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments +kernelspec.make_ipkernel_cmd = custom_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi"
codereview_new_python_data_10700
import sys from ipykernel import kernelspec default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd -def new_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments -kernelspec.make_ipkernel_cmd = new_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi" ```suggestion default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd ``` import sys from ipykernel import kernelspec + default_make_ipkernel_cmd = kernelspec.make_ipkernel_cmd +def custom_make_ipkernel_cmd( mod="ipykernel_launcher", executable=None, extra_arguments=None ): mpi_arguments = ["mpiexec", "-n", "1"] arguments = default_make_ipkernel_cmd(mod, executable, extra_arguments) return mpi_arguments + arguments +kernelspec.make_ipkernel_cmd = custom_make_ipkernel_cmd if __name__ == "__main__": kernel_name = "python3mpi"
codereview_new_python_data_10701
from nbconvert.preprocessors import ExecutePreprocessor test_dataset_path = "taxi.csv" -kernel_name = ( - os.environ["MODIN_KERNEL_NAME"] if "MODIN_KERNEL_NAME" in os.environ else None -) -ep = ExecutePreprocessor( - timeout=600, kernel_name=kernel_name if kernel_name else "python3" -) download_taxi_dataset = f"""import os import urllib.request ```suggestion # the kernel name "python3mpi" must match the one # that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` # for `Unidist` engine kernel_name = "python3" if cfg.Engine.get() != "Unidist" else "python3mpi" ep = ExecutePreprocessor(timeout=600, kernel_name=kernel_name) ``` from nbconvert.preprocessors import ExecutePreprocessor test_dataset_path = "taxi.csv" +# the kernel name "python3mpi" must match the one +# that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` +# for `Unidist` engine +kernel_name = "python3" if cfg.Engine.get() != "Unidist" else "python3mpi" +ep = ExecutePreprocessor(timeout=600, kernel_name=kernel_name) download_taxi_dataset = f"""import os import urllib.request
codereview_new_python_data_10702
_execute_notebook, test_dataset_path, download_taxi_dataset, - change_kernel, ) # the kernel name "python3mpi" must match the one # that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` # for `Unidist` engine -change_kernel(kernel_name="python3mpi") local_notebooks_dir = "examples/tutorial/jupyter/execution/pandas_on_unidist/local" I prefer to use my suggestion in order not to create the executor twice. _execute_notebook, test_dataset_path, download_taxi_dataset, + set_kernel, ) # the kernel name "python3mpi" must match the one # that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` # for `Unidist` engine +set_kernel(kernel_name="python3mpi") local_notebooks_dir = "examples/tutorial/jupyter/execution/pandas_on_unidist/local"
codereview_new_python_data_10703
def custom_make_ipkernel_cmd(*args, **kwargs): """ - Build modifyied Popen command list for launching an IPython kernel with mpi. - Returns ------- array - A Popen command list Notes ----- ```suggestion Build modified Popen command list for launching an IPython kernel with MPI. ``` def custom_make_ipkernel_cmd(*args, **kwargs): """ + Build modified Popen command list for launching an IPython kernel with MPI. + + Parameters + ---------- + *args : iterable + Additional positional arguments to be passed in `default_make_ipkernel_cmd`. + **kwargs : dict + Additional keyword arguments to be passed in `default_make_ipkernel_cmd`. + Returns ------- array + A Popen command list. Notes -----
codereview_new_python_data_10704
def custom_make_ipkernel_cmd(*args, **kwargs): """ - Build modifyied Popen command list for launching an IPython kernel with mpi. - Returns ------- array - A Popen command list Notes ----- ```suggestion Parameters ---------- *args : iterable Additional positional arguments to be passed in `default_make_ipkernel_cmd`. **kwargs : dict Additional keyword arguments to be passed in `default_make_ipkernel_cmd`. Returns ------- array A Popen command list. ``` def custom_make_ipkernel_cmd(*args, **kwargs): """ + Build modified Popen command list for launching an IPython kernel with MPI. + + Parameters + ---------- + *args : iterable + Additional positional arguments to be passed in `default_make_ipkernel_cmd`. + **kwargs : dict + Additional keyword arguments to be passed in `default_make_ipkernel_cmd`. + Returns ------- array + A Popen command list. Notes -----
codereview_new_python_data_10705
_execute_notebook, test_dataset_path, download_taxi_dataset, - change_kernel, ) # the kernel name "python3mpi" must match the one # that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` # for `Unidist` engine -change_kernel(kernel_name="python3mpi") local_notebooks_dir = "examples/tutorial/jupyter/execution/pandas_on_unidist/local" ```suggestion set_kernel, ``` _execute_notebook, test_dataset_path, download_taxi_dataset, + set_kernel, ) # the kernel name "python3mpi" must match the one # that is set up in `examples/tutorial/jupyter/execution/pandas_on_unidist/setup_kernel.py` # for `Unidist` engine +set_kernel(kernel_name="python3mpi") local_notebooks_dir = "examples/tutorial/jupyter/execution/pandas_on_unidist/local"
codereview_new_python_data_10706
def call_progress_bar(result_parts, line_no): elif modin_engine == "Unidist": from unidist import wait else: - raise RuntimeError( f"ProgressBar feature is not supported for {modin_engine} engine." ) We probably need to enable a test for this engine? def call_progress_bar(result_parts, line_no): elif modin_engine == "Unidist": from unidist import wait else: + raise NotImplementedError( f"ProgressBar feature is not supported for {modin_engine} engine." )
codereview_new_python_data_10707
# ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. -"""The module for working with displaying progress bars for Modin execution engines.""" import os import time @modin-project/modin-core, does anyone remember why the file has the suffix `aqp`? @noah-kuo, tagging you as the original author. # ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. +""" +The module for working with displaying progress bars for Modin execution engines. + +Modin Automatic Query Progress (AQP). +""" import os import time
codereview_new_python_data_10708
def check_partition_column(partition_column, cols): if k == partition_column: if v == "int": return - raise InvalidPartitionColumn( - "partition_column must be int, and not {0}".format(v) - ) raise InvalidPartitionColumn( - "partition_column {0} not found in the query".format(partition_column) ) ```suggestion f"partition_column must be int, and not {v}" ``` def check_partition_column(partition_column, cols): if k == partition_column: if v == "int": return + raise InvalidPartitionColumn(f"partition_column must be int, and not {v}") raise InvalidPartitionColumn( + f"partition_column {partition_column} not found in the query" )
codereview_new_python_data_10709
def test_map(data, na_values): # Index into list objects df_equals( - modin_series_lists.map(lambda list: list[0]), - pandas_series_lists.map(lambda list: list[0]), ) ```suggestion modin_series_lists.map(lambda lst: lst[0]), pandas_series_lists.map(lambda lst: lst[0]), ``` better not shadow builtins; also why change? def test_map(data, na_values): # Index into list objects df_equals( + modin_series_lists.map(lambda lst: lst[0]), + pandas_series_lists.map(lambda lst: lst[0]), )
codereview_new_python_data_10710
def groupby_agg( how="axis_wise", drop=False, ): - # Defaulting to pandas in case of an empty frame if len(self.columns) == 0 or len(self.index) == 0: return super().groupby_agg( by, agg_func, axis, groupby_kwargs, agg_args, agg_kwargs, how, drop we don't default to pandas at the front end in case of a lazy query compiler and so there are still cases where empty data could leak into qc def groupby_agg( how="axis_wise", drop=False, ): + # Defaulting to pandas in case of an empty frame as we can't process it properly. + # Higher API level won't pass empty data here unless the frame has delayed + # computations. So we apparently lose some laziness here (due to index access) + # because of the disability to process empty groupby natively. if len(self.columns) == 0 or len(self.index) == 0: return super().groupby_agg( by, agg_func, axis, groupby_kwargs, agg_args, agg_kwargs, how, drop
codereview_new_python_data_10711
def test_index_of_empty_frame(): md_df, pd_df = create_test_dfs( {}, index=pandas.Index([], name="index name"), columns=["a", "b"] ) - assert md_df.empty and md_df.empty - df_equals(md_df.index, md_df.index) # Test on an empty frame produced by Modin's logic data = test_data_values[0] uh? why double-check `md_df.empty`?.. should the second one be `pd_df.empty`? def test_index_of_empty_frame(): md_df, pd_df = create_test_dfs( {}, index=pandas.Index([], name="index name"), columns=["a", "b"] ) + assert md_df.empty and pd_df.empty + df_equals(md_df.index, pd_df.index) # Test on an empty frame produced by Modin's logic data = test_data_values[0]
codereview_new_python_data_10712
def test_index_of_empty_frame(): md_df, pd_df = create_test_dfs( {}, index=pandas.Index([], name="index name"), columns=["a", "b"] ) - assert md_df.empty and md_df.empty - df_equals(md_df.index, md_df.index) # Test on an empty frame produced by Modin's logic data = test_data_values[0] same, should it be comparing `md_df` and `pd_df`? def test_index_of_empty_frame(): md_df, pd_df = create_test_dfs( {}, index=pandas.Index([], name="index name"), columns=["a", "b"] ) + assert md_df.empty and pd_df.empty + df_equals(md_df.index, pd_df.index) # Test on an empty frame produced by Modin's logic data = test_data_values[0]
codereview_new_python_data_10713
from .arr import * from .math import * from .constants import * ## 'import *' may pollute namespace Import pollutes the enclosing namespace, as the imported module [modin.numpy.arr](1) does not define '__all__'. [Show more details](https://github.com/modin-project/modin/security/code-scanning/470) +# Licensed to Modin Development Team under one or more contributor license agreements. +# See the NOTICE file distributed with this work for additional information regarding +# copyright ownership. The Modin Development Team licenses this file to you under the +# Apache License, Version 2.0 (the "License"); you may not use this file except in +# compliance with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software distributed under +# the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF +# ANY KIND, either express or implied. See the License for the specific language +# governing permissions and limitations under the License. + from .arr import * from .math import * from .constants import *
codereview_new_python_data_10714
from .arr import * from .math import * from .constants import * ## 'import *' may pollute namespace Import pollutes the enclosing namespace, as the imported module [modin.numpy.math](1) does not define '__all__'. [Show more details](https://github.com/modin-project/modin/security/code-scanning/471) +# Licensed to Modin Development Team under one or more contributor license agreements. +# See the NOTICE file distributed with this work for additional information regarding +# copyright ownership. The Modin Development Team licenses this file to you under the +# Apache License, Version 2.0 (the "License"); you may not use this file except in +# compliance with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software distributed under +# the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF +# ANY KIND, either express or implied. See the License for the specific language +# governing permissions and limitations under the License. + from .arr import * from .math import * from .constants import *
codereview_new_python_data_10715
from .arr import * from .math import * from .constants import * ## 'import *' may pollute namespace Import pollutes the enclosing namespace, as the imported module [modin.numpy.constants](1) does not define '__all__'. [Show more details](https://github.com/modin-project/modin/security/code-scanning/472) +# Licensed to Modin Development Team under one or more contributor license agreements. +# See the NOTICE file distributed with this work for additional information regarding +# copyright ownership. The Modin Development Team licenses this file to you under the +# Apache License, Version 2.0 (the "License"); you may not use this file except in +# compliance with the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software distributed under +# the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF +# ANY KIND, either express or implied. See the License for the specific language +# governing permissions and limitations under the License. + from .arr import * from .math import * from .constants import *
codereview_new_python_data_10716
def to_numpy( Convert the `BasePandasDataset` to a NumPy array. """ from modin.config import ExperimentalNumPyAPI if ExperimentalNumPyAPI.get(): from ..numpy.arr import array return array(_query_compiler=self._query_compiler, _ndim=2) - return self._query_compiler.to_numpy( dtype=dtype, copy=copy, this should be imported on top of the file, not in the middle of nowhere def to_numpy( Convert the `BasePandasDataset` to a NumPy array. """ from modin.config import ExperimentalNumPyAPI + if ExperimentalNumPyAPI.get(): from ..numpy.arr import array return array(_query_compiler=self._query_compiler, _ndim=2) + return self._query_compiler.to_numpy( dtype=dtype, copy=copy,
codereview_new_python_data_10717
def to_numpy( Convert the `BasePandasDataset` to a NumPy array. """ from modin.config import ExperimentalNumPyAPI if ExperimentalNumPyAPI.get(): from ..numpy.arr import array return array(_query_compiler=self._query_compiler, _ndim=2) - return self._query_compiler.to_numpy( dtype=dtype, copy=copy, I would say this should be put into `modin/pandas/dataframe.py`, not in its base class; `BaseDataset` should contain only common methods for DF and Series, and this piece differs def to_numpy( Convert the `BasePandasDataset` to a NumPy array. """ from modin.config import ExperimentalNumPyAPI + if ExperimentalNumPyAPI.get(): from ..numpy.arr import array return array(_query_compiler=self._query_compiler, _ndim=2) + return self._query_compiler.to_numpy( dtype=dtype, copy=copy,
codereview_new_python_data_10718
def to_numpy( Return the NumPy ndarray representing the values in this Series or Index. """ from modin.config import ExperimentalNumPyAPI if not ExperimentalNumPyAPI.get(): return ( super(Series, self) import should be done on top of the file def to_numpy( Return the NumPy ndarray representing the values in this Series or Index. """ from modin.config import ExperimentalNumPyAPI + if not ExperimentalNumPyAPI.get(): return ( super(Series, self)
codereview_new_python_data_10719
class TestReadFromPostgres(EnvironmentVariable, type=bool): varname = "MODIN_TEST_READ_FROM_POSTGRES" default = False class ExperimentalNumPyAPI(EnvironmentVariable, type=bool): """Set to true to use Modin's experimental NumPy API.""" varname = "MODIN_EXPERIMENTAL_NUMPY_API" default = False class ReadSqlEngine(EnvironmentVariable, type=str): """Engine to run `read_sql`.""" Requires changes in https://modin.readthedocs.io/en/stable/flow/modin/config.html? class TestReadFromPostgres(EnvironmentVariable, type=bool): varname = "MODIN_TEST_READ_FROM_POSTGRES" default = False + class ExperimentalNumPyAPI(EnvironmentVariable, type=bool): """Set to true to use Modin's experimental NumPy API.""" varname = "MODIN_EXPERIMENTAL_NUMPY_API" default = False + class ReadSqlEngine(EnvironmentVariable, type=str): """Engine to run `read_sql`."""
codereview_new_python_data_10720
def where(condition, x=None, y=None): - if condition: return x - if not condition: return y if hasattr(condition, "where"): return condition.where(x=x, y=y) This seems incorrect, since `where` is supposed to check the `condition` array element-wise. Even if the argument here were a modin.numpy array that has the `where` attribute, this check happens too early for that to matter. ```python >>> import numpy >>> numpy.where([1, 0], [1, 1], [2, 2]) array([1, 2]) >>> import modin.numpy as np >>> np.where([1,0], [1,1],[2,2]) [1, 1] ``` def where(condition, x=None, y=None): + if condition is True: return x + if condition is False: return y if hasattr(condition, "where"): return condition.where(x=x, y=y)
codereview_new_python_data_10721
newaxis, pi, ) Do we need to define `__all__` here? newaxis, pi, ) + +__all__ = [ + "Inf", + "Infinity", + "NAN", + "NINF", + "NZERO", + "NaN", + "PINF", + "PZERO", + "e", + "euler_gamma", + "inf", + "infty", + "nan", + "newaxis", + "pi", +]
codereview_new_python_data_10722
# governing permissions and limitations under the License. """Module houses array creation methods for Modin's NumPy API.""" import numpy from modin.error_message import ErrorMessage from .arr import array ## Cyclic import Import of module [modin.numpy.arr](1) begins an import cycle. [Show more details](https://github.com/modin-project/modin/security/code-scanning/537) # governing permissions and limitations under the License. """Module houses array creation methods for Modin's NumPy API.""" + import numpy + from modin.error_message import ErrorMessage from .arr import array
codereview_new_python_data_10723
# governing permissions and limitations under the License. """Module houses array shaping methods for Modin's NumPy API.""" import numpy from modin.error_message import ErrorMessage ## Cyclic import Import of module [modin.numpy.arr](1) begins an import cycle. [Show more details](https://github.com/modin-project/modin/security/code-scanning/555) # governing permissions and limitations under the License. """Module houses array shaping methods for Modin's NumPy API.""" + import numpy from modin.error_message import ErrorMessage
codereview_new_python_data_10724
# governing permissions and limitations under the License. """Module houses array shaping methods for Modin's NumPy API.""" import numpy from modin.error_message import ErrorMessage ```suggestion """Module houses array shaping methods for Modin's NumPy API.""" import numpy ``` # governing permissions and limitations under the License. """Module houses array shaping methods for Modin's NumPy API.""" + import numpy from modin.error_message import ErrorMessage
codereview_new_python_data_10725
# governing permissions and limitations under the License. """Module houses array creation methods for Modin's NumPy API.""" import numpy from modin.error_message import ErrorMessage from .arr import array ```suggestion """Module houses array creation methods for Modin's NumPy API.""" import numpy from modin.error_message import ErrorMessage from .arr import array ``` # governing permissions and limitations under the License. """Module houses array creation methods for Modin's NumPy API.""" + import numpy + from modin.error_message import ErrorMessage from .arr import array
codereview_new_python_data_10726
# the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF # ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. """Module houses ``array`` class, that is distributed version of ``numpy.array``.""" from math import prod ```suggestion """Module houses ``array`` class, that is distributed version of ``numpy.array``.""" ``` # the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF # ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. + """Module houses ``array`` class, that is distributed version of ``numpy.array``.""" from math import prod
codereview_new_python_data_10727
def initialize_ray( ray_init_kwargs = { "num_cpus": CpuCount.get(), "num_gpus": GpuCount.get(), - "include_dashboard": True, "ignore_reinit_error": True, "object_store_memory": object_store_memory, "_redis_password": redis_password, Why? Also, this is not related to the issue. def initialize_ray( ray_init_kwargs = { "num_cpus": CpuCount.get(), "num_gpus": GpuCount.get(), + "include_dashboard": False, "ignore_reinit_error": True, "object_store_memory": object_store_memory, "_redis_password": redis_password,
codereview_new_python_data_10728
from .partition import PandasOnRayDataframePartition from .partition_manager import PandasOnRayDataframePartitionManager -from .virtual_partition import ( - PandasOnRayDataframeVirtualPartition, - PandasOnRayDataframeColumnPartition, - PandasOnRayDataframeRowPartition, -) __all__ = [ "PandasOnRayDataframePartition", "PandasOnRayDataframePartitionManager", - "PandasOnRayDataframeVirtualPartition", - "PandasOnRayDataframeColumnPartition", - "PandasOnRayDataframeRowPartition", ] same as for dask and also not related to the issue from .partition import PandasOnRayDataframePartition from .partition_manager import PandasOnRayDataframePartitionManager __all__ = [ "PandasOnRayDataframePartition", "PandasOnRayDataframePartitionManager", ]
codereview_new_python_data_10729
def initialize_ray( ray_init_kwargs = { "num_cpus": CpuCount.get(), "num_gpus": GpuCount.get(), - "include_dashboard": True, "ignore_reinit_error": True, "object_store_memory": object_store_memory, "_redis_password": redis_password, ```suggestion "include_dashboard": False, ``` def initialize_ray( ray_init_kwargs = { "num_cpus": CpuCount.get(), "num_gpus": GpuCount.get(), + "include_dashboard": False, "ignore_reinit_error": True, "object_store_memory": object_store_memory, "_redis_password": redis_password,
codereview_new_python_data_10730
# We have to explicitly mock subclass implementations of wait_partitions. if engine == "Ray": wait_method = ( - "modin.core.execution.ray.implementations." - + "pandas_on_ray.partitioning." + "PandasOnRayDataframePartitionManager.wait_partitions" ) elif engine == "Dask": Let's make this change in a separate PR where we define public interface for ray engine (e.g., expose the partition classes). # We have to explicitly mock subclass implementations of wait_partitions. if engine == "Ray": wait_method = ( + "modin.core.execution.ray.implementations.pandas_on_ray." + + "partitioning.partition_manager." + "PandasOnRayDataframePartitionManager.wait_partitions" ) elif engine == "Dask":
codereview_new_python_data_10731
if Engine.get() == "Ray": import ray - from modin.core.execution.ray.implementations.pandas_on_ray.partitioning import ( PandasOnRayDataframePartition, ) - from modin.core.execution.ray.implementations.pandas_on_ray.partitioning import ( PandasOnRayDataframeColumnPartition, PandasOnRayDataframeRowPartition, ) same for ray if Engine.get() == "Ray": import ray + from modin.core.execution.ray.implementations.pandas_on_ray.partitioning.partition import ( PandasOnRayDataframePartition, ) + from modin.core.execution.ray.implementations.pandas_on_ray.partitioning.virtual_partition import ( PandasOnRayDataframeColumnPartition, PandasOnRayDataframeRowPartition, )
codereview_new_python_data_10732
def fn( method = kwargs.get("method") if isinstance(result, pandas.Series): - result = result.to_frame(method or MODIN_UNNAMED_SERIES_LABEL) if not as_index: if isinstance(by, pandas.Series): this was actually one of those cases where we still were using "0" column name rather than `MODIN_UNNAMED_SERIES_LABEL` ```python >>> pd.Series([1, 2, 3]).to_frame().columns RangeIndex(start=0, stop=1, step=1) ``` def fn( method = kwargs.get("method") if isinstance(result, pandas.Series): + result = result.to_frame(MODIN_UNNAMED_SERIES_LABEL) if not as_index: if isinstance(by, pandas.Series):
codereview_new_python_data_10733
def fn( method = kwargs.get("method") if isinstance(result, pandas.Series): - result = result.to_frame(method or MODIN_UNNAMED_SERIES_LABEL) if not as_index: if isinstance(by, pandas.Series): So `method` can be used as name for `to_frame`. Is it a mistake or not? def fn( method = kwargs.get("method") if isinstance(result, pandas.Series): + result = result.to_frame(MODIN_UNNAMED_SERIES_LABEL) if not as_index: if isinstance(by, pandas.Series):
codereview_new_python_data_10734
"TimeReindexMethod", "TimeFillnaMethodDataframe", "TimeDropDuplicatesDataframe", - "TimeSimpleReshape", "TimeReplace", # IO benchmarks "TimeReadCsvSkiprows", ```suggestion "TimeStack", "TimeUnstack", ``` "TimeReindexMethod", "TimeFillnaMethodDataframe", "TimeDropDuplicatesDataframe", + "TimeStack", + "TimeUnstack", "TimeReplace", # IO benchmarks "TimeReadCsvSkiprows",
codereview_new_python_data_10735
def do_relabel(obj_to_relabel): agg_kwargs=kwargs, how="axis_wise", ) - return result if not do_relabel else do_relabel(result) agg = aggregate ```suggestion return do_relabel(result) if do_relabel else result ``` positive conditions are read easier def do_relabel(obj_to_relabel): agg_kwargs=kwargs, how="axis_wise", ) + return do_relabel(result) if do_relabel else result agg = aggregate
codereview_new_python_data_10736
class FactoryDispatcher(object): @classmethod def get_factory(cls) -> factories.BaseFactory: """Get current factory.""" - Engine.subscribe(cls._update_factory) - StorageFormat.subscribe(cls._update_factory) return cls.__factory @classmethod Formerly, these were in global scope at the end of this file. I moved them here because we want to subscribe after the user has possibly added extra engines. class FactoryDispatcher(object): @classmethod def get_factory(cls) -> factories.BaseFactory: """Get current factory.""" + if cls.__factory is None: + Engine.subscribe(cls._update_factory) + StorageFormat.subscribe(cls._update_factory) return cls.__factory @classmethod
codereview_new_python_data_10737
class FactoryDispatcher(object): @classmethod def get_factory(cls) -> factories.BaseFactory: """Get current factory.""" - Engine.subscribe(cls._update_factory) - StorageFormat.subscribe(cls._update_factory) return cls.__factory @classmethod This would subscribe multiple times, while originally it was one callback per module import (hence one callback per lifetime); maybe `Parameter.subscribe()` should be modified to only add a callback once? As now it just adds them to a list to maintain a FIFO order. class FactoryDispatcher(object): @classmethod def get_factory(cls) -> factories.BaseFactory: """Get current factory.""" + if cls.__factory is None: + Engine.subscribe(cls._update_factory) + StorageFormat.subscribe(cls._update_factory) return cls.__factory @classmethod
codereview_new_python_data_10738
def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) -class TimeSetCategories: - - params = [get_benchmark_shapes("TimeSetCategories")] - param_names = ["shape"] - def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) -class TimeRemoveCategories: - params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] - def setup(self, shape): - rows = shape[0] - arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] - self.ts = IMPL.Series(arr).astype("category") - execute(self.ts) - def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2])) ```suggestion class BaseCategories: def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) class TimeSetCategories(BaseCategories): params = [get_benchmark_shapes("TimeSetCategories")] param_names = ["shape"] def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) ``` def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) +class BaseCategories: def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) + +class TimeSetCategories(BaseCategories): + params = [get_benchmark_shapes("TimeSetCategories")] + param_names = ["shape"] + def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) +class TimeRemoveCategories(BaseCategories): params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2]))
codereview_new_python_data_10739
def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) -class TimeSetCategories: - - params = [get_benchmark_shapes("TimeSetCategories")] - param_names = ["shape"] - def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) -class TimeRemoveCategories: - params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] - def setup(self, shape): - rows = shape[0] - arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] - self.ts = IMPL.Series(arr).astype("category") - execute(self.ts) - def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2])) ```suggestion class TimeRemoveCategories(BaseCategories): params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2])) ``` def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) +class BaseCategories: def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) + +class TimeSetCategories(BaseCategories): + params = [get_benchmark_shapes("TimeSetCategories")] + param_names = ["shape"] + def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) +class TimeRemoveCategories(BaseCategories): params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2]))
codereview_new_python_data_10740
def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) -class TimeSetCategories: - - params = [get_benchmark_shapes("TimeSetCategories")] - param_names = ["shape"] - def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) -class TimeRemoveCategories: - params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] - def setup(self, shape): - rows = shape[0] - arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] - self.ts = IMPL.Series(arr).astype("category") - execute(self.ts) - def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2])) In Modin, the same mechanism is used for these functions. It is enough to test at least one of these. ```suggestion def time_timedelta_seconds(self, shape): execute(self.series.dt.seconds) ``` def time_timedelta_nanoseconds(self, shape): execute(self.series.dt.nanoseconds) +class BaseCategories: def setup(self, shape): rows = shape[0] arr = [f"s{i:04d}" for i in np.random.randint(0, rows // 10, size=rows)] self.ts = IMPL.Series(arr).astype("category") execute(self.ts) + +class TimeSetCategories(BaseCategories): + params = [get_benchmark_shapes("TimeSetCategories")] + param_names = ["shape"] + def time_set_categories(self, shape): execute(self.ts.cat.set_categories(self.ts.cat.categories[::2])) +class TimeRemoveCategories(BaseCategories): params = [get_benchmark_shapes("TimeRemoveCategories")] param_names = ["shape"] def time_remove_categories(self, shape): execute(self.ts.cat.remove_categories(self.ts.cat.categories[::2]))
codereview_new_python_data_10741
def _repartition(self, axis: Optional[int] = None): DataFrame or Series The repartitioned dataframe or series, depending on the original type. """ - if axis not in (0, 1, None): - raise NotImplementedError if StorageFormat.get() == "Hdk": # Hdk uses only one partition, it makes Can you pass the error message to the exception to be more informative? def _repartition(self, axis: Optional[int] = None): DataFrame or Series The repartitioned dataframe or series, depending on the original type. """ + allowed_axis_values = (0, 1, None) + if axis not in allowed_axis_values: + raise NotImplementedError( + f"passed `axis` parameter: {axis}, but should be one of {allowed_axis_values}" + ) if StorageFormat.get() == "Hdk": # Hdk uses only one partition, it makes
codereview_new_python_data_10742
def _repartition(self, axis: Optional[int] = None): Parameters ---------- - axis : int, optional Returns ------- ```suggestion axis : {0, 1}, optional ``` def _repartition(self, axis: Optional[int] = None): Parameters ---------- + axis : {0, 1}, optional Returns -------
codereview_new_python_data_10743
def get_indices(cls, axis, partitions, index_func=None): new_idx = [idx.apply(func) for idx in target[0]] if len(target) else [] new_idx = cls.get_objects_from_partitions(new_idx) # filter empty indexes - new_idx = list(filter(lambda idx: len(idx), new_idx)) - # TODO FIX INFORMATION LEAK!!!!1!!1!! - total_idx = new_idx[0].append(new_idx[1:]) if new_idx else new_idx return total_idx, new_idx @classmethod ## Unnecessary lambda This 'lambda' is just a simple wrapper around a callable object. Use that object directly. [Show more details](https://github.com/modin-project/modin/security/code-scanning/445) def get_indices(cls, axis, partitions, index_func=None): new_idx = [idx.apply(func) for idx in target[0]] if len(target) else [] new_idx = cls.get_objects_from_partitions(new_idx) # filter empty indexes + total_idx = list(filter(len, new_idx)) + if len(total_idx) > 0: + # TODO FIX INFORMATION LEAK!!!!1!!1!! + total_idx = total_idx[0].append(total_idx[1:]) return total_idx, new_idx @classmethod
codereview_new_python_data_10744
class TimeDropDuplicatesDataframe: param_names = ["shape"] def setup(self, shape): - N = shape[0] // 10 K = 10 - key1 = tm.makeStringIndex(N).values.repeat(K) - key2 = tm.makeStringIndex(N).values.repeat(K) - self.df = IMPL.DataFrame( - {"key1": key1, "key2": key2, "value": np.random.randn(N * K)} - ) execute(self.df) def time_drop_dups(self, shape): `self.df` should have amount of columns equals `shape[1]` class TimeDropDuplicatesDataframe: param_names = ["shape"] def setup(self, shape): + rows,cols = shape + N = rows // 10 K = 10 + self.df = IMPL.DataFrame() + for col in range(cols-1): # dataframe would have cols-1 keys(strings) and one value(int) column + self.df["key"+str(col+1)] = tm.makeStringIndex(N).values.repeat(K) + self.df["value"] = np.random.randn(N * K) execute(self.df) def time_drop_dups(self, shape):
codereview_new_python_data_10745
def setup(self, shape): execute(self.df) def time_drop_dups(self, shape): - execute(self.df.drop_duplicates(["key1", "key2"])) def time_drop_dups_inplace(self, shape): - self.df.drop_duplicates(["key1", "key2"], inplace=True) execute(self.df) Let's perform the operation on all columns: ```suggestion execute(self.df.drop_duplicates()) ``` def setup(self, shape): execute(self.df) def time_drop_dups(self, shape): + execute(self.df.drop_duplicates(self.df.columns[:-1])) def time_drop_dups_inplace(self, shape): + self.df.drop_duplicates(self.df.columns[:-1], inplace=True) execute(self.df)
codereview_new_python_data_10746
def setup(self, shape): execute(self.df) def time_drop_dups(self, shape): - execute(self.df.drop_duplicates(["key1", "key2"])) def time_drop_dups_inplace(self, shape): - self.df.drop_duplicates(["key1", "key2"], inplace=True) execute(self.df) Same ```suggestion self.df.drop_duplicates(inplace=True) ``` def setup(self, shape): execute(self.df) def time_drop_dups(self, shape): + execute(self.df.drop_duplicates(self.df.columns[:-1])) def time_drop_dups_inplace(self, shape): + self.df.drop_duplicates(self.df.columns[:-1], inplace=True) execute(self.df)
codereview_new_python_data_10747
def _wrap_aggregation( DataFrame or Series Returns the same type as `self._df`. """ - if not isinstance(numeric_only, NumericOnly): - numeric_only = NumericOnly(numeric_only) agg_args = tuple() if agg_args is None else agg_args agg_kwargs = dict() if agg_kwargs is None else agg_kwargs I wonder if we can unconditionally cast to `NumericOnly(numeric_only)`, even if it already is a proper enum... def _wrap_aggregation( DataFrame or Series Returns the same type as `self._df`. """ + numeric_only = NumericOnly(numeric_only) agg_args = tuple() if agg_args is None else agg_args agg_kwargs = dict() if agg_kwargs is None else agg_kwargs
codereview_new_python_data_10748
class BasePandasDataset(ClassLogger): # but lives in "pandas" namespace. _pandas_class = pandas.core.generic.NDFrame - # TODO(https://github.com/modin-project/modin/issues/4821): - # make this cache_readonly - @property def _is_dataframe(self) -> bool: """ Tell whether this is a dataframe. `cache_readonly` is accessible from `pandas.util`, why cannot we just import it and use instead of adding a TODO? It's not specific to any Python version (as I'm _not_ suggesting to use `cached_property` from `functools`, as I find pandas version safer even if a little slower) class BasePandasDataset(ClassLogger): # but lives in "pandas" namespace. _pandas_class = pandas.core.generic.NDFrame + @pandas.util.cache_readonly def _is_dataframe(self) -> bool: """ Tell whether this is a dataframe.
codereview_new_python_data_10749
def broadcast_item( index_values = obj.index[row_lookup] if not index_values.equals(item.index): axes_to_reindex["index"] = index_values - if need_columns_reindex and isinstance(item, (pandas.DataFrame, DataFrame)): column_values = obj.columns[col_lookup] if not column_values.equals(item.columns): axes_to_reindex["columns"] = column_values this looks super-weird, should we really support both pandas and Modin versions of `DataFrame` here? def broadcast_item( index_values = obj.index[row_lookup] if not index_values.equals(item.index): axes_to_reindex["index"] = index_values + if need_columns_reindex and isinstance(item, DataFrame): column_values = obj.columns[col_lookup] if not column_values.equals(item.columns): axes_to_reindex["columns"] = column_values
codereview_new_python_data_10750
def new_col_adder(df, partition_id): NotImplementedError, match="Dynamic repartitioning is currently only supported for DataFrames with 1 partition.", ): - _ = pipeline.compute_batch() def test_fan_out(self): """Check that the fan_out argument is appropriately handled.""" ```suggestion pipeline.compute_batch() ``` Why assign anything at all? def new_col_adder(df, partition_id): NotImplementedError, match="Dynamic repartitioning is currently only supported for DataFrames with 1 partition.", ): + pipeline.compute_batch() def test_fan_out(self): """Check that the fan_out argument is appropriately handled."""
codereview_new_python_data_10751
def test_astype(data): "data", [["A", "A", "B", "B", "A"], [1, 1, 2, 1, 2, 2, 3, 1, 2, 1, 2]] ) def test_astype_categorical(data): - modin_df, pandas_df = pd.Series(data), pandas.Series(data) modin_result = modin_df.astype("category") pandas_result = pandas_df.astype("category") ```suggestion modin_df, pandas_df = create_test_series(data) ``` def test_astype(data): "data", [["A", "A", "B", "B", "A"], [1, 1, 2, 1, 2, 2, 3, 1, 2, 1, 2]] ) def test_astype_categorical(data): + modin_df, pandas_df = create_test_series(data) modin_result = modin_df.astype("category") pandas_result = pandas_df.astype("category")
codereview_new_python_data_10752
def initialize_ray( What password to use when connecting to Redis. If not specified, ``modin.config.RayRedisPassword`` is used. """ extra_init_kw = {"runtime_env": {"env_vars": {"__MODIN_AUTOIMPORT_PANDAS__": "1"}}} if not ray.is_initialized() or override_is_cluster: - # TODO(https://github.com/ray-project/ray/issues/28216): remove this - # workaround once Ray gives a better way to suppress task errors. - # Ideally we would not set global environment variables. - # If user has explicitly set _RAY_IGNORE_UNHANDLED_ERRORS_VAR, don't - # don't override its value. - if _RAY_IGNORE_UNHANDLED_ERRORS_VAR not in os.environ: - os.environ[_RAY_IGNORE_UNHANDLED_ERRORS_VAR] = "1" cluster = override_is_cluster or IsRayCluster.get() redis_address = override_redis_address or RayRedisAddress.get() redis_password = ( What if ray is already initialized by the user? Do we want to supress errors in remote tasks in that case? def initialize_ray( What password to use when connecting to Redis. If not specified, ``modin.config.RayRedisPassword`` is used. """ + # TODO(https://github.com/ray-project/ray/issues/28216): remove this + # workaround once Ray gives a better way to suppress task errors. + # Ideally we would not set global environment variables. + # If user has explicitly set _RAY_IGNORE_UNHANDLED_ERRORS_VAR, don't + # don't override its value. + if _RAY_IGNORE_UNHANDLED_ERRORS_VAR not in os.environ: + os.environ[_RAY_IGNORE_UNHANDLED_ERRORS_VAR] = "1" extra_init_kw = {"runtime_env": {"env_vars": {"__MODIN_AUTOIMPORT_PANDAS__": "1"}}} if not ray.is_initialized() or override_is_cluster: cluster = override_is_cluster or IsRayCluster.get() redis_address = override_redis_address or RayRedisAddress.get() redis_password = (
codereview_new_python_data_10753
def execute( return partitions = df._query_compiler._modin_frame._partitions.flatten() if len(partitions) > 0 and hasattr(partitions[0], "wait"): - all(map(lambda partition: partition.wait() or True, partitions)) return # compatibility with old Modin versions I would suggest using either a simple for-loop or a special partition manager's method, I don't see reasons why we should use this complicated `all(map(...))` approach ```suggestion df._query_compiler._modin_frame._partition_mgr_cls.wait_partitions(partitions) ``` btw, the further code in this function still uses the incorrect logic def execute( return partitions = df._query_compiler._modin_frame._partitions.flatten() if len(partitions) > 0 and hasattr(partitions[0], "wait"): + df._query_compiler._modin_frame._partition_mgr_cls.wait_partitions(partitions) return # compatibility with old Modin versions
codereview_new_python_data_10754
def execute( df._query_compiler._modin_frame._execute() return partitions = df._query_compiler._modin_frame._partitions.flatten() - if len(partitions) > 0 and hasattr(partitions[0], "wait"): - df._query_compiler._modin_frame._partition_mgr_cls.wait_partitions( - partitions - ) return # compatibility with old Modin versions we probably should change the condition then to something like: ```python if len(partitions) and hasattr(partition_mgr_cls, "wait_partitions"): ``` def execute( df._query_compiler._modin_frame._execute() return partitions = df._query_compiler._modin_frame._partitions.flatten() + mgr_cls = df._query_compiler._modin_frame._partition_mgr_cls + if len(partitions) and hasattr(mgr_cls, "wait_partitions"): + mgr_cls.wait_partitions(partitions) return # compatibility with old Modin versions
codereview_new_python_data_10755
def test_dict(self): assert mdt == "category" assert isinstance(mdt, pandas.CategoricalDtype) assert pandas.api.types.is_categorical_dtype(mdt) if type(mdt) != pandas.CategoricalDtype: # This is a lazy proxy. - # Make sure the table is not materialized after - # the instance check and comparison with string. assert mdt._table is not None assert mdt == pdt assert pdt == mdt - assert str(mdt) == str(pdt) assert repr(mdt) == repr(pdt) can we put `str` case in the "lazy" section of the test? We would like to print `df.dtypes` that trigger `str(dtype)` for every without materialization def test_dict(self): assert mdt == "category" assert isinstance(mdt, pandas.CategoricalDtype) assert pandas.api.types.is_categorical_dtype(mdt) + assert str(mdt) == str(pdt) if type(mdt) != pandas.CategoricalDtype: # This is a lazy proxy. + # Make sure the table is not materialized yet. assert mdt._table is not None assert mdt == pdt assert pdt == mdt assert repr(mdt) == repr(pdt)
codereview_new_python_data_10756
def test_dict(self): assert pandas.api.types.is_categorical_dtype(mdt) assert str(mdt) == str(pdt) - if type(mdt) != pandas.CategoricalDtype: - # This is a lazy proxy. - # Make sure the table is not materialized yet. - assert mdt._table is not None assert mdt == pdt assert pdt == mdt why do we need this if-conditioning if we intend to test a lazy proxy here? ```suggestion assert type(mdt) == LazyProxyCategoricalDtype # This is a lazy proxy. # Make sure the table is not materialized yet. assert mdt._table is not None ``` def test_dict(self): assert pandas.api.types.is_categorical_dtype(mdt) assert str(mdt) == str(pdt) + # Make sure the lazy proxy dtype is not materialized yet. + assert type(mdt) != pandas.CategoricalDtype + assert mdt._table is not None assert mdt == pdt assert pdt == mdt
codereview_new_python_data_10757
# ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. -"""Utilities for internal use by te hdk_on_native module.""" import pandas ```suggestion """Utilities for internal use by the ``HdkOnNativeDataframe``.""" ``` # ANY KIND, either express or implied. See the License for the specific language # governing permissions and limitations under the License. +"""Utilities for internal use by the ``HdkOnNativeDataframe``.""" import pandas
codereview_new_python_data_10758
def ml(train_final, test_final): evals=watchlist, feval=func_loss, early_stopping_rounds=10, - verbose_eval=1000, ) yp = clf.predict(dvalid) ```suggestion verbose_eval=None, ``` Decided to leave as is to not issue a lot of warnings which might be kind of confusing to users. def ml(train_final, test_final): evals=watchlist, feval=func_loss, early_stopping_rounds=10, + verbose_eval=None, ) yp = clf.predict(dvalid)
codereview_new_python_data_10759
def ml(train_final, test_final): evals=watchlist, feval=func_loss, early_stopping_rounds=10, - verbose_eval=1000, ) yp = clf.predict(dvalid) ```suggestion verbose_eval=None, ``` Decided to leave as is to not issue a lot of warnings which might be kind of confusing to users. def ml(train_final, test_final): evals=watchlist, feval=func_loss, early_stopping_rounds=10, + verbose_eval=None, ) yp = clf.predict(dvalid)
codereview_new_python_data_10760
def test_info(data, verbose, max_cols, memory_usage, null_counts): assert modin_info[1:] == pandas_info[1:] -def test_info_default_cols(): - # Covers https://github.com/modin-project/modin/issues/5137 - with io.StringIO() as first, io.StringIO() as second: - data = np.random.randint(0, 100, (10, 10)) - eval_general( - pd.DataFrame(data), - pandas.DataFrame(data), - operation=lambda df, **kwargs: df.info(**kwargs), - buf=lambda df: second if isinstance(df, pandas.DataFrame) else first, - ) - modin_info = first.getvalue().splitlines() - pandas_info = second.getvalue().splitlines() - - assert modin_info[0] == str(pd.DataFrame) - assert pandas_info[0] == str(pandas.DataFrame) - assert modin_info[1:] == pandas_info[1:] - - @pytest.mark.parametrize("axis", axis_values, ids=axis_keys) @pytest.mark.parametrize("skipna", bool_arg_values, ids=bool_arg_keys) @pytest.mark.parametrize("numeric_only", bool_arg_values, ids=bool_arg_keys) Should we delete this test case? def test_info(data, verbose, max_cols, memory_usage, null_counts): assert modin_info[1:] == pandas_info[1:] @pytest.mark.parametrize("axis", axis_values, ids=axis_keys) @pytest.mark.parametrize("skipna", bool_arg_values, ids=bool_arg_keys) @pytest.mark.parametrize("numeric_only", bool_arg_values, ids=bool_arg_keys)
codereview_new_python_data_10761
def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) - s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2) ## Statement has no effect This statement has no effect. [Show more details](https://github.com/modin-project/modin/security/code-scanning/385) def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) + _ = s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2)
codereview_new_python_data_10762
def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) - s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2) Maybe make the change to disable CodeQL warning? ```suggestion _ = s1 + s2 ``` def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) + _ = s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2)
codereview_new_python_data_10763
def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) - s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2) why not do something like `s1.add(s2)` or something l like that? def test_add_does_not_change_original_series_name(): s2 = pd.Series(2, name=2) original_s1 = s1.copy(deep=True) original_s2 = s2.copy(deep=True) + _ = s1 + s2 df_equals(s1, original_s1) df_equals(s2, original_s2)
codereview_new_python_data_10764
def test_add_custom_class(): ) def test_non_commutative_multiply(): # This test checks that mul and rmul do different things when # multiplication is not commutative, e.g. for adding a string to a string. # For context see https://github.com/modin-project/modin/issues/5238 modin_df, pandas_df = create_test_dfs([1], dtype=int) integer = NonCommutativeMultiplyInteger(2) - # It's tricky to get the non commutative integer class implementation - # right, so before we do the actual test, check that the operation - # we care about is really not commmutative in pandas. - assert not (integer * pandas_df).equals(pandas_df * integer) eval_general(modin_df, pandas_df, lambda s: integer * s) eval_general(modin_df, pandas_df, lambda s: s * integer) this should be a separate test that implementation of `NonCommutativeMultiplyInteger` is correct def test_add_custom_class(): ) +def test_non_commutative_multiply_pandas(): + # The non commutative integer class implementation is tricky. Check that + # multiplying such an integer with a pandas dataframe is really not + # commutative. + pandas_df = pd.DataFrame([[1]], dtype=int) + integer = NonCommutativeMultiplyInteger(2) + assert not (integer * pandas_df).equals(pandas_df * integer) + + def test_non_commutative_multiply(): # This test checks that mul and rmul do different things when # multiplication is not commutative, e.g. for adding a string to a string. # For context see https://github.com/modin-project/modin/issues/5238 modin_df, pandas_df = create_test_dfs([1], dtype=int) integer = NonCommutativeMultiplyInteger(2) eval_general(modin_df, pandas_df, lambda s: integer * s) eval_general(modin_df, pandas_df, lambda s: s * integer)
codereview_new_python_data_10765
def test_non_commutative_add_string_to_series(data): eval_general(*create_test_series(data), lambda s: s + "string") def test_non_commutative_multiply(): # This test checks that mul and rmul do different things when # multiplication is not commutative, e.g. for adding a string to a string. # For context see https://github.com/modin-project/modin/issues/5238 modin_series, pandas_series = create_test_series(1, dtype=int) integer = NonCommutativeMultiplyInteger(2) - # It's tricky to get the non commutative integer class implementation - # right, so before we do the actual test, check that the operation - # we care about is really not commmutative in pandas. - assert not (integer * pandas_series).equals(pandas_series * integer) eval_general(modin_series, pandas_series, lambda s: integer * s) eval_general(modin_series, pandas_series, lambda s: s * integer) again, this should be a separate test that checks implementation of non-commutative int def test_non_commutative_add_string_to_series(data): eval_general(*create_test_series(data), lambda s: s + "string") +def test_non_commutative_multiply_pandas(): + # The non commutative integer class implementation is tricky. Check that + # multiplying such an integer with a pandas series is really not + # commutative. + pandas_series = pd.DataFrame([[1]], dtype=int) + integer = NonCommutativeMultiplyInteger(2) + assert not (integer * pandas_series).equals(pandas_series * integer) + + def test_non_commutative_multiply(): # This test checks that mul and rmul do different things when # multiplication is not commutative, e.g. for adding a string to a string. # For context see https://github.com/modin-project/modin/issues/5238 modin_series, pandas_series = create_test_series(1, dtype=int) integer = NonCommutativeMultiplyInteger(2) eval_general(modin_series, pandas_series, lambda s: integer * s) eval_general(modin_series, pandas_series, lambda s: s * integer)
codereview_new_python_data_10766
def compare(self, other, **kwargs): return self.__constructor__( self._modin_frame.broadcast_apply_full_axis( 0, - lambda l, r: pandas.DataFrame.compare(l, r, **kwargs), other._modin_frame, ) ) ```suggestion lambda l, r: pandas.DataFrame.compare(l, other=r, **kwargs), ``` for clarity? def compare(self, other, **kwargs): return self.__constructor__( self._modin_frame.broadcast_apply_full_axis( 0, + lambda l, r: pandas.DataFrame.compare(l, other=r, **kwargs), other._modin_frame, ) )