Ensure cudf internals use pylibcudf in pure Python mode

python/cudf/cudf/_lib/column.pxd

@@ -7,7 +7,6 @@ from libcpp.memory cimport unique_ptr
 
 from rmm._lib.device_buffer cimport device_buffer
 
-from cudf._lib cimport pylibcudf
 from cudf._lib.cpp.column.column cimport column
 from cudf._lib.cpp.column.column_view cimport column_view, mutable_column_view
 from cudf._lib.cpp.types cimport size_type
@@ -30,7 +29,7 @@ cdef class Column:
     cdef column_view _view(self, size_type null_count) except *
     cdef column_view view(self) except *
     cdef mutable_column_view mutable_view(self) except *
-    cpdef pylibcudf.Column to_pylibcudf(self, mode: Literal["read", "write"])
+    cpdef to_pylibcudf(self, mode: Literal["read", "write"])
 
     @staticmethod
     cdef Column from_unique_ptr(

python/cudf/cudf/_lib/column.pyx

@@ -11,9 +11,6 @@ import rmm
 import cudf
 import cudf._lib as libcudf
 from cudf._lib import pylibcudf
-
-from cudf._lib cimport pylibcudf
-
 from cudf.api.types import is_categorical_dtype, is_datetime64tz_dtype
 from cudf.core.buffer import (
     Buffer,
@@ -40,6 +37,7 @@ from cudf._lib.types cimport (
 )
 
 from cudf._lib.null_mask import bitmask_allocation_size_bytes
+from cudf._lib.types import dtype_from_pylibcudf_column
 
 cimport cudf._lib.cpp.types as libcudf_types
 cimport cudf._lib.cpp.unary as libcudf_unary
@@ -446,7 +444,7 @@ cdef class Column:
     # underlying buffers as exposed before this function can itself be exposed
     # publicly.  User requests to convert to pylibcudf must assume that the
     # data may be modified afterwards.
-    cpdef pylibcudf.Column to_pylibcudf(self, mode: Literal["read", "write"]):
+    cpdef to_pylibcudf(self, mode: Literal["read", "write"]):
         """Convert this Column to a pylibcudf.Column.
 
         This function will generate a pylibcudf Column pointing to the same
@@ -476,9 +474,9 @@ cdef class Column:
         else:
             col = self
 
-        cdef pylibcudf.DataType dtype = dtype_to_pylibcudf_type(col.dtype)
+        dtype = dtype_to_pylibcudf_type(col.dtype)
 
-        cdef pylibcudf.gpumemoryview data = None
+        data = None
         if col.base_data is not None:
             cai = cuda_array_interface_wrapper(
                 ptr=col.base_data.get_ptr(mode=mode),
@@ -487,7 +485,7 @@ cdef class Column:
             )
             data = pylibcudf.gpumemoryview(cai)
 
-        cdef pylibcudf.gpumemoryview mask = None
+        mask = None
         if self.nullable:
             # TODO: Are we intentionally use self's mask instead of col's?
             # Where is the mask stored for categoricals?
@@ -586,7 +584,7 @@ cdef class Column:
     #  TODO: Actually support exposed data pointers.
     @staticmethod
     def from_pylibcudf(
-        pylibcudf.Column col, bint data_ptr_exposed=False
+        col, bint data_ptr_exposed=False
     ):
         """Create a Column from a pylibcudf.Column.
 
@@ -607,19 +605,20 @@ cdef class Column:
         pylibcudf.Column
             A new pylibcudf.Column referencing the same data.
         """
-        # TODO: Rewrite utility for dtype conversion to not need a column view.
-        dtype = dtype_from_column_view(col.view())
+        dtype = dtype_from_pylibcudf_column(col)
 
         return cudf.core.column.build_column(
-            data=as_buffer(col.data.obj) if col.data is not None else None,
+            data=as_buffer(col.data().obj) if col.data() is not None else None,
             dtype=dtype,
-            size=col.size,
-            mask=as_buffer(col.mask.obj) if col.mask is not None else None,
-            offset=col.offset,
-            null_count=col.null_count,
+            size=col.size(),
+            mask=as_buffer(
+                col.null_mask().obj
+            ) if col.null_mask() is not None else None,
+            offset=col.offset(),
+            null_count=col.null_count(),
             children=tuple([
                 Column.from_pylibcudf(child)
-                for child in col.children
+                for child in col.children()
             ])
         )
 

python/cudf/cudf/_lib/copying.pyx

@@ -11,9 +11,9 @@ from libcpp.vector cimport vector
 from rmm._lib.device_buffer cimport DeviceBuffer
 
 import cudf
+from cudf._lib import pylibcudf
 from cudf.core.buffer import Buffer, acquire_spill_lock, as_buffer
 
-from cudf._lib cimport pylibcudf
 from cudf._lib.column cimport Column
 
 from cudf._lib.scalar import as_device_scalar
@@ -174,7 +174,7 @@ def gather(
     Column gather_map,
     bool nullify=False
 ):
-    cdef pylibcudf.Table tbl = pylibcudf.copying.gather(
+    tbl = pylibcudf.copying.gather(
         pylibcudf.Table([col.to_pylibcudf(mode="read") for col in columns]),
         gather_map.to_pylibcudf(mode="read"),
         pylibcudf.copying.OutOfBoundsPolicy.NULLIFY if nullify

python/cudf/cudf/_lib/pylibcudf/column.pxd

@@ -15,16 +15,36 @@ cdef class Column:
     # TODO: Should we document these attributes? Should we mark them readonly?
     cdef:
         # Core data
-        DataType data_type
-        size_type size
-        gpumemoryview data
-        gpumemoryview mask
-        size_type null_count
-        size_type offset
+        DataType _data_type
+        size_type _size
+        gpumemoryview _data
+        gpumemoryview _mask
+        size_type _null_count
+        size_type _offset
         # children: List[Column]
-        list children
+        list _children
+        size_type _num_children
 
     cdef column_view view(self) nogil
 
     @staticmethod
     cdef Column from_libcudf(unique_ptr[column] libcudf_col)
+
+    cpdef DataType type(self)
+    cpdef Column child(self, size_type index)
+    cpdef size_type num_children(self)
+    cpdef size_type size(self)
+    cpdef size_type null_count(self)
+    cpdef size_type offset(self)
+    cpdef gpumemoryview data(self)
+    cpdef gpumemoryview null_mask(self)
+    cpdef list children(self)
+
+    cpdef list_view(self)
+
+
+cdef class ListColumnView:
+    """Accessor for methods of a Column that are specific to lists."""
+    cdef Column _column
+    cpdef child(self)
+    cpdef offsets(self)

python/cudf/cudf/_lib/pylibcudf/column.pyx

@@ -9,7 +9,7 @@ from cudf._lib.cpp.column.column cimport column, column_contents
 from cudf._lib.cpp.types cimport size_type
 
 from .gpumemoryview cimport gpumemoryview
-from .types cimport DataType
+from .types cimport DataType, TypeId
 from .utils cimport int_to_bitmask_ptr, int_to_void_ptr
 
 
@@ -45,13 +45,14 @@ cdef class Column:
         gpumemoryview mask, size_type null_count, size_type offset,
         list children
     ):
-        self.data_type = data_type
-        self.size = size
-        self.data = data
-        self.mask = mask
-        self.null_count = null_count
-        self.offset = offset
-        self.children = children
+        self._data_type = data_type
+        self._size = size
+        self._data = data
+        self._mask = mask
+        self._null_count = null_count
+        self._offset = offset
+        self._children = children
+        self._num_children = len(children)
 
     cdef column_view view(self) nogil:
         """Generate a libcudf column_view to pass to libcudf algorithms.
@@ -63,17 +64,17 @@ cdef class Column:
         cdef const void * data = NULL
         cdef const bitmask_type * null_mask = NULL
 
-        if self.data is not None:
-            data = int_to_void_ptr(self.data.ptr)
-        if self.mask is not None:
-            null_mask = int_to_bitmask_ptr(self.mask.ptr)
+        if self._data is not None:
+            data = int_to_void_ptr(self._data.ptr)
+        if self._mask is not None:
+            null_mask = int_to_bitmask_ptr(self._mask.ptr)
 
         # TODO: Check if children can ever change. If not, this could be
         # computed once in the constructor and always be reused.
         cdef vector[column_view] c_children
         with gil:
-            if self.children is not None:
-                for child in self.children:
+            if self._children is not None:
+                for child in self._children:
                     # Need to cast to Column here so that Cython knows that
                     # `view` returns a typed object, not a Python object. We
                     # cannot use a typed variable for `child` because cdef
@@ -86,8 +87,8 @@ cdef class Column:
                     c_children.push_back((<Column> child).view())
 
         return column_view(
-            self.data_type.c_obj, self.size, data, null_mask,
-            self.null_count, self.offset, c_children
+            self._data_type.c_obj, self._size, data, null_mask,
+            self._null_count, self._offset, c_children
         )
 
     @staticmethod
@@ -133,3 +134,60 @@ cdef class Column:
             0,
             children,
         )
+
+    cpdef DataType type(self):
+        """The type of data in the column."""
+        return self._data_type
+
+    cpdef Column child(self, size_type index):
+        """Get a child column of this column.
+
+        Parameters
+        ----------
+        index : size_type
+            The index of the child column to get.
+
+        Returns
+        -------
+        Column
+            The child column.
+        """
+        return self._children[index]
+
+    cpdef size_type num_children(self):
+        """The number of children of this column."""
+        return self._num_children
+
+    cpdef list_view(self):
+        return ListColumnView(self)
+
+    cpdef gpumemoryview data(self):
+        return self._data
+
+    cpdef gpumemoryview null_mask(self):
+        return self._mask
+
+    cpdef size_type size(self):
+        return self._size
+
+    cpdef size_type offset(self):
+        return self._offset
+
+    cpdef size_type null_count(self):
+        return self._null_count
+
+    cpdef list children(self):
+        return self._children
+
+cdef class ListColumnView:
+    """Accessor for methods of a Column that are specific to lists."""
+    def __init__(self, Column col):
+        if col.type().id() != TypeId.LIST:
+            raise TypeError("Column is not a list type")
+        self._column = col
+
+    cpdef child(self):
+        return self._column.child(1)
+
+    cpdef offsets(self):
+        return self._column.child(1)

python/cudf/cudf/_lib/pylibcudf/gpumemoryview.pxd

@@ -2,5 +2,8 @@
 
 
 cdef class gpumemoryview:
-    cdef Py_ssize_t ptr
-    cdef object obj
+    # TODO: Eventually probably want to make this opaque, but for now it's fine
+    # to treat this object as something like a POD struct
+    cdef readonly:
+        Py_ssize_t ptr
+        object obj

python/cudf/cudf/_lib/pylibcudf/table.pxd

@@ -8,9 +8,11 @@ from cudf._lib.cpp.table.table_view cimport table_view
 
 cdef class Table:
     # List[pylibcudf.Column]
-    cdef object columns
+    cdef list _columns
 
     cdef table_view view(self) nogil
 
     @staticmethod
     cdef Table from_libcudf(unique_ptr[table] libcudf_tbl)
+
+    cpdef list columns(self)

python/cudf/cudf/_lib/pylibcudf/table.pyx

@@ -21,7 +21,7 @@ cdef class Table:
         The columns in this table.
     """
     def __init__(self, list columns):
-        self.columns = columns
+        self._columns = columns
 
     cdef table_view view(self) nogil:
         """Generate a libcudf table_view to pass to libcudf algorithms.
@@ -31,11 +31,11 @@ cdef class Table:
         (even direct pylibcudf Cython users).
         """
         # TODO: Make c_columns a class attribute that is updated along with
-        # self.columns whenever new columns are added or columns are removed.
+        # self._columns whenever new columns are added or columns are removed.
         cdef vector[column_view] c_columns
 
         with gil:
-            for col in self.columns:
+            for col in self._columns:
                 c_columns.push_back((<Column> col).view())
 
         return table_view(c_columns)
@@ -57,3 +57,6 @@ cdef class Table:
             Column.from_libcudf(move(c_columns[i]))
             for i in range(c_columns.size())
         ])
+
+    cpdef list columns(self):
+        return self._columns

python/cudf/cudf/_lib/types.pxd

@@ -4,7 +4,6 @@ from libc.stdint cimport int32_t
 from libcpp cimport bool
 
 cimport cudf._lib.cpp.types as libcudf_types
-from cudf._lib cimport pylibcudf
 from cudf._lib.cpp.column.column_view cimport column_view
 from cudf._lib.cpp.lists.lists_column_view cimport lists_column_view
 
@@ -18,5 +17,5 @@ ctypedef bool underlying_type_t_null_policy
 cdef dtype_from_column_view(column_view cv)
 
 cdef libcudf_types.data_type dtype_to_data_type(dtype) except *
-cpdef pylibcudf.DataType dtype_to_pylibcudf_type(dtype)
+cpdef dtype_to_pylibcudf_type(dtype)
 cdef bool is_decimal_type_id(libcudf_types.type_id tid) except *