FIX-modin-project#7134: Use a separate docstring class for BasePandas…

…Dataset. (modin-project#7353)

Signed-off-by: sfc-gh-mvashishtha <mahesh.vashishtha@snowflake.com>

modin/config/envvars.py

@@ -331,6 +331,24 @@ class CpuCount(EnvironmentVariable, type=int):
     """How many CPU cores to use during initialization of the Modin engine."""
 
     varname = "MODIN_CPUS"
+
+    @classmethod
+    def _put(cls, value: int) -> None:
+        """
+        Put specific value if CpuCount wasn't set by a user yet.
+
+        Parameters
+        ----------
+        value : int
+            Config value to set.
+
+        Notes
+        -----
+        This method is used to set CpuCount from cluster resources internally
+        and should not be called by a user.
+        """
+        if cls.get_value_source() == ValueSource.DEFAULT:
+            cls.put(value)
 
     @classmethod
     def _get_default(cls) -> int:

modin/core/execution/dask/common/utils.py

@@ -74,3 +74,4 @@ def _disable_warnings():
 
     num_cpus = len(client.ncores())
     NPartitions._put(num_cpus)
+    CpuCount._put(num_cpus)

modin/core/execution/ray/common/utils.py

@@ -151,6 +151,7 @@ def initialize_ray(
 
     num_cpus = int(ray.cluster_resources()["CPU"])
     NPartitions._put(num_cpus)
+    CpuCount._put(num_cpus)
 
     # TODO(https://github.com/ray-project/ray/issues/28216): remove this
     # workaround once Ray gives a better way to suppress task errors.

modin/core/execution/unidist/common/utils.py

@@ -42,6 +42,7 @@ def initialize_unidist():
 
     num_cpus = sum(v["CPU"] for v in unidist.cluster_resources().values())
     modin_cfg.NPartitions._put(num_cpus)
+    modin_cfg.CpuCount._put(num_cpus)
 
 
 def deserialize(obj):  # pragma: no cover

modin/tests/config/docs_module/__init__.py

@@ -11,7 +11,7 @@
 # ANY KIND, either express or implied. See the License for the specific language
 # governing permissions and limitations under the License.
 
-from .classes import DataFrame, Series
+from .classes import BasePandasDataset, DataFrame, Series
 from .functions import read_csv
 
-__all__ = ["DataFrame", "Series", "read_csv"]
+__all__ = ["BasePandasDataset", "DataFrame", "Series", "read_csv"]

modin/tests/config/docs_module/classes.py

@@ -22,3 +22,11 @@ class Series:
     def isna(self):
         """This is a test of the documentation module for Series."""
         return
+
+
+class BasePandasDataset:
+    """This is a test of the documentation module for BasePandasDataSet."""
+
+    def apply():
+        """This is a test of the documentation module for BasePandasDataSet.apply."""
+        return

modin/tests/config/test_envvars.py

@@ -20,6 +20,7 @@
 import modin.pandas as pd
 from modin.config.envvars import _check_vars
 from modin.config.pubsub import _UNSET, ExactStr
+from modin.pandas.base import BasePandasDataset
 
 
 def reset_vars(*vars: tuple[cfg.Parameter]):
@@ -89,13 +90,20 @@ def test_overrides(self):
         cfg.DocModule.put("modin.tests.config.docs_module")
 
         # Test for override
+        assert BasePandasDataset.__doc__ == (
+            "This is a test of the documentation module for BasePandasDataSet."
+        )
+        assert BasePandasDataset.apply.__doc__ == (
+            "This is a test of the documentation module for BasePandasDataSet.apply."
+        )
         assert (
             pd.DataFrame.apply.__doc__
             == "This is a test of the documentation module for DataFrame."
         )
         # Test for pandas doc when method is not defined on the plugin module
         assert pandas.DataFrame.isna.__doc__ in pd.DataFrame.isna.__doc__
         assert pandas.DataFrame.isnull.__doc__ in pd.DataFrame.isnull.__doc__
+        assert BasePandasDataset.astype.__doc__ in pd.DataFrame.astype.__doc__
         # Test for override
         assert (
             pd.Series.isna.__doc__

modin/utils.py

@@ -462,7 +462,18 @@ def _inherit_docstrings_in_place(
     if doc_module != DocModule.default and "pandas" in str(
         getattr(parent, "__module__", "")
     ):
-        parent = getattr(imported_doc_module, getattr(parent, "__name__", ""), parent)
+        parent_name = (
+            # DocModule should use the class BasePandasDataset to override the
+            # docstrings of BasePandasDataset, even if BasePandasDataset
+            # normally inherits docstrings from a different `parent`.
+            "BasePandasDataset"
+            if getattr(cls_or_func, "__name__", "") == "BasePandasDataset"
+            # For other classes, override docstrings with the class that has the
+            # same name as the `parent` class, e.g. DataFrame inherits
+            # docstrings from doc_module.DataFrame.
+            else getattr(parent, "__name__", "")
+        )
+        parent = getattr(imported_doc_module, parent_name, parent)
     if parent != default_parent:
         # Reset API link in case the docs are overridden.
         apilink = None