FEAT-#7401: Implement DataFrame/Series.attrs

modin/pandas/base.py

@@ -16,6 +16,8 @@
 from __future__ import annotations
 
 import abc
+import copy
+import functools
 import pickle as pkl
 import re
 import warnings
@@ -105,6 +107,7 @@
     "_ipython_canary_method_should_not_exist_",
     "_ipython_display_",
     "_repr_mimebundle_",
+    "_attrs",
 }
 
 _DEFAULT_BEHAVIOUR = {
@@ -193,6 +196,26 @@ def _get_repr_axis_label_indexer(labels, num_for_repr):
     )
 
 
+def propagate_self_attrs(method):
+    """
+    Wrap a BasePandasDataset/DataFrame/Series method with a function that automatically deep-copies self.attrs if present.
+
+    This annotation should not be used on special methods like concat, str, and groupby, which may need to
+    examine multiple sources to reconcile `attrs`.
+    """
+
+    @functools.wraps(method)
+    def wrapper(self, *args, **kwargs):
+        result = method(self, *args, **kwargs)
+        if isinstance(result, BasePandasDataset) and len(self._attrs):
+            # If the result of the method call is a modin.pandas object and `self.attrs` is
+            # not empty, perform a deep copy of `self.attrs`.
+            result._attrs = copy.deepcopy(self._attrs)
+        return result
+
+    return wrapper
+
+
 @_inherit_docstrings(pandas.DataFrame, apilink=["pandas.DataFrame", "pandas.Series"])
 class BasePandasDataset(ClassLogger):
     """
@@ -208,6 +231,7 @@ class BasePandasDataset(ClassLogger):
     _pandas_class = pandas.core.generic.NDFrame
     _query_compiler: BaseQueryCompiler
     _siblings: list[BasePandasDataset]
+    _attrs: dict
 
     @cached_property
     def _is_dataframe(self) -> bool:
@@ -1125,6 +1149,20 @@ def at(self, axis=None) -> _LocIndexer:  # noqa: PR01, RT01, D200
 
         return _LocIndexer(self)
 
+    def _set_attrs(self, key: Any, value: Any) -> dict:  # noqa: PR01, RT01, D200
+        """
+        Set the dictionary of global attributes of this dataset.
+        """
+        self._attrs[key] = value
+
+    def _get_attrs(self) -> dict:  # noqa: PR01, RT01, D200
+        """
+        Get the dictionary of global attributes of this dataset.
+        """
+        return self._attrs
+
+    attrs: dict = property(_get_attrs, _set_attrs)
+
     def at_time(self, time, asof=False, axis=None) -> Self:  # noqa: PR01, RT01, D200
         """
         Select values at particular time of day (e.g., 9:30AM).
@@ -3221,6 +3259,7 @@ def tail(self, n=5) -> Self:  # noqa: PR01, RT01, D200
             return self.iloc[-n:]
         return self.iloc[len(self) :]
 
+    @propagate_self_attrs
     def take(self, indices, axis=0, **kwargs) -> Self:  # noqa: PR01, RT01, D200
         """
         Return the elements in the given *positional* indices along an axis.

modin/pandas/dataframe.py

@@ -151,8 +151,11 @@ def __init__(
         # Siblings are other dataframes that share the same query compiler. We
         # use this list to update inplace when there is a shallow copy.
         self._siblings = []
+        self._attrs = {}
         if isinstance(data, (DataFrame, Series)):
             self._query_compiler = data._query_compiler.copy()
+            if len(data._attrs):
+                self._attrs = copy.deepcopy(data._attrs)
             if index is not None and any(i not in data.index for i in index):
                 raise NotImplementedError(
                     "Passing non-existant columns or index values to constructor not"
@@ -2636,12 +2639,12 @@ def __setattr__(self, key, value) -> None:
         # - anything in self.__dict__. This includes any attributes that the
         #   user has added to the dataframe with,  e.g., `df.c = 3`, and
         #   any attribute that Modin has added to the frame, e.g.
-        #   `_query_compiler` and `_siblings`
+        #   `_query_compiler`, `_siblings`, and "_attrs"
         # - `_query_compiler`, which Modin initializes before it appears in
         #   __dict__
         # - `_siblings`, which Modin initializes before it appears in __dict__
         #   before it appears in __dict__.
-        if key in ("_query_compiler", "_siblings") or key in self.__dict__:
+        if key in ("_attrs", "_query_compiler", "_siblings") or key in self.__dict__:
             pass
         # we have to check for the key in `dir(self)` first in order not to trigger columns computation
         elif key not in dir(self) and key in self:
@@ -2938,17 +2941,6 @@ def __dataframe_consortium_standard__(
         )
         return convert_to_standard_compliant_dataframe(self, api_version=api_version)
 
-    @property
-    def attrs(self) -> dict:  # noqa: RT01, D200
-        """
-        Return dictionary of global attributes of this dataset.
-        """
-
-        def attrs(df):
-            return df.attrs
-
-        return self._default_to_pandas(attrs)
-
     @property
     def style(self):  # noqa: RT01, D200
         """

modin/pandas/series.py

@@ -114,8 +114,11 @@ def __init__(
         # Siblings are other dataframes that share the same query compiler. We
         # use this list to update inplace when there is a shallow copy.
         self._siblings = []
+        self._attrs = {}
         if isinstance(data, type(self)):
             query_compiler = data._query_compiler.copy()
+            if len(data._attrs):
+                self._attrs = copy.deepcopy(data._attrs)
             if index is not None:
                 if any(i not in data.index for i in index):
                     raise NotImplementedError(
@@ -2264,17 +2267,6 @@ def where(
             level=level,
         )
 
-    @property
-    def attrs(self) -> dict:  # noqa: RT01, D200
-        """
-        Return dictionary of global attributes of this dataset.
-        """
-
-        def attrs(df):
-            return df.attrs
-
-        return self._default_to_pandas(attrs)
-
     @property
     def array(self) -> ExtensionArray:  # noqa: RT01, D200
         """