[REVIEW] Refactor the Buffer class

python/cudf/cudf/_lib/column.pyx

@@ -9,7 +9,7 @@ import rmm
 import cudf
 import cudf._lib as libcudf
 from cudf.api.types import is_categorical_dtype, is_list_dtype, is_struct_dtype
-from cudf.core.buffer import Buffer
+from cudf.core.buffer import Buffer, as_buffer, buffer_from_pointer
 
 from cpython.buffer cimport PyObject_CheckBuffer
 from libc.stdint cimport uintptr_t
@@ -233,27 +233,27 @@ cdef class Column:
                 if isinstance(value, Column):
                     value = value.data_array_view
                 value = cp.asarray(value).view('|u1')
-            mask = Buffer(value)
+            mask = as_buffer(value)
             if mask.size < required_num_bytes:
                 raise ValueError(error_msg.format(str(value.size)))
             if mask.size < mask_size:
                 dbuf = rmm.DeviceBuffer(size=mask_size)
                 dbuf.copy_from_device(value)
-                mask = Buffer(dbuf)
+                mask = as_buffer(dbuf)
         elif hasattr(value, "__array_interface__"):
             value = np.asarray(value).view("u1")[:mask_size]
             if value.size < required_num_bytes:
                 raise ValueError(error_msg.format(str(value.size)))
             dbuf = rmm.DeviceBuffer(size=mask_size)
             dbuf.copy_from_host(value)
-            mask = Buffer(dbuf)
+            mask = as_buffer(dbuf)
         elif PyObject_CheckBuffer(value):
             value = np.asarray(value).view("u1")[:mask_size]
             if value.size < required_num_bytes:
                 raise ValueError(error_msg.format(str(value.size)))
             dbuf = rmm.DeviceBuffer(size=mask_size)
             dbuf.copy_from_host(value)
-            mask = Buffer(dbuf)
+            mask = as_buffer(dbuf)
         else:
             raise TypeError(
                 "Expected a Buffer-like object or None for mask, got "
@@ -455,11 +455,11 @@ cdef class Column:
         cdef column_contents contents = move(c_col.get()[0].release())
 
         data = DeviceBuffer.c_from_unique_ptr(move(contents.data))
-        data = Buffer(data)
+        data = as_buffer(data)
 
         if has_nulls:
             mask = DeviceBuffer.c_from_unique_ptr(move(contents.null_mask))
-            mask = Buffer(mask)
+            mask = as_buffer(mask)
             null_count = c_col.get()[0].null_count()
         else:
             mask = None
@@ -511,18 +511,18 @@ cdef class Column:
 
         if data_ptr:
             if data_owner is None:
-                data = Buffer(
+                data = as_buffer(
                     rmm.DeviceBuffer(ptr=data_ptr,
                                      size=(size+offset) * dtype.itemsize)
                 )
             else:
-                data = Buffer(
-                    data=data_ptr,
+                data = buffer_from_pointer(
+                    ptr=data_ptr,
                     size=(base_size) * dtype.itemsize,
                     owner=data_owner
                 )
         else:
-            data = Buffer(
+            data = as_buffer(
                 rmm.DeviceBuffer(ptr=data_ptr, size=0)
             )
 
@@ -552,15 +552,15 @@ cdef class Column:
                     # result:
                     mask = None
                 else:
-                    mask = Buffer(
+                    mask = as_buffer(
                         rmm.DeviceBuffer(
                             ptr=mask_ptr,
                             size=bitmask_allocation_size_bytes(size+offset)
                         )
                     )
             else:
-                mask = Buffer(
-                    data=mask_ptr,
+                mask = buffer_from_pointer(
+                    ptr=mask_ptr,
                     size=bitmask_allocation_size_bytes(base_size),
                     owner=mask_owner
                 )

python/cudf/cudf/_lib/concat.pyx

@@ -1,4 +1,4 @@
-# Copyright (c) 2020, NVIDIA CORPORATION.
+# Copyright (c) 2020-2022, NVIDIA CORPORATION.
 
 from libcpp cimport bool
 from libcpp.memory cimport make_unique, unique_ptr
@@ -19,7 +19,7 @@ from cudf._lib.utils cimport (
     table_view_from_table,
 )
 
-from cudf.core.buffer import Buffer
+from cudf.core.buffer import as_buffer
 
 from rmm._lib.device_buffer cimport DeviceBuffer, device_buffer
 
@@ -31,7 +31,7 @@ cpdef concat_masks(object columns):
     with nogil:
         c_result = move(libcudf_concatenate_masks(c_views))
         c_unique_result = make_unique[device_buffer](move(c_result))
-    return Buffer(DeviceBuffer.c_from_unique_ptr(move(c_unique_result)))
+    return as_buffer(DeviceBuffer.c_from_unique_ptr(move(c_unique_result)))
 
 
 cpdef concat_columns(object columns):

python/cudf/cudf/_lib/copying.pyx

@@ -14,7 +14,7 @@ from libcpp.vector cimport vector
 from rmm._lib.device_buffer cimport DeviceBuffer
 
 import cudf
-from cudf.core.buffer import Buffer
+from cudf.core.buffer import Buffer, buffer_from_pointer
 
 from cudf._lib.column cimport Column
 
@@ -718,7 +718,11 @@ cdef class _CPackedColumns:
         header = {}
         frames = []
 
-        gpu_data = Buffer(self.gpu_data_ptr, self.gpu_data_size, self)
+        gpu_data = buffer_from_pointer(
+            ptr=self.gpu_data_ptr,
+            size=self.gpu_data_size,
+            owner=self
+        )
         data_header, data_frames = gpu_data.serialize()
         header["data"] = data_header
         frames.extend(data_frames)

python/cudf/cudf/_lib/null_mask.pyx

@@ -17,7 +17,7 @@ from cudf._lib.cpp.null_mask cimport (
 )
 from cudf._lib.cpp.types cimport mask_state, size_type
 
-from cudf.core.buffer import Buffer
+from cudf.core.buffer import as_buffer
 
 
 class MaskState(Enum):
@@ -47,7 +47,7 @@ def copy_bitmask(Column col):
         up_db = make_unique[device_buffer](move(db))
 
     rmm_db = DeviceBuffer.c_from_unique_ptr(move(up_db))
-    buf = Buffer(rmm_db)
+    buf = as_buffer(rmm_db)
     return buf
 
 
@@ -93,5 +93,5 @@ def create_null_mask(size_type size, state=MaskState.UNINITIALIZED):
         up_db = make_unique[device_buffer](move(db))
 
     rmm_db = DeviceBuffer.c_from_unique_ptr(move(up_db))
-    buf = Buffer(rmm_db)
+    buf = as_buffer(rmm_db)
     return buf

python/cudf/cudf/_lib/transform.pyx

@@ -6,7 +6,7 @@ from numba.np import numpy_support
 import cudf
 from cudf._lib.types import SUPPORTED_NUMPY_TO_LIBCUDF_TYPES
 from cudf.core._internals.expressions import parse_expression
-from cudf.core.buffer import Buffer
+from cudf.core.buffer import as_buffer
 from cudf.utils import cudautils
 
 from cython.operator cimport dereference
@@ -52,7 +52,7 @@ def bools_to_mask(Column col):
         up_db = move(cpp_out.first)
 
     rmm_db = DeviceBuffer.c_from_unique_ptr(move(up_db))
-    buf = Buffer(rmm_db)
+    buf = as_buffer(rmm_db)
     return buf
 
 
@@ -88,7 +88,7 @@ def nans_to_nulls(Column input):
         return None
 
     buffer = DeviceBuffer.c_from_unique_ptr(move(c_buffer))
-    buffer = Buffer(buffer)
+    buffer = as_buffer(buffer)
     return buffer
 
 

python/cudf/cudf/core/abc.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2020-2021, NVIDIA CORPORATION.
+# Copyright (c) 2020-2022, NVIDIA CORPORATION.
 """Common abstract base classes for cudf."""
 
 import sys
@@ -130,7 +130,7 @@ def device_deserialize(cls, header, frames):
         """
         typ = pickle.loads(header["type-serialized"])
         frames = [
-            cudf.core.buffer.Buffer(f) if c else memoryview(f)
+            cudf.core.buffer.as_buffer(f) if c else memoryview(f)
             for c, f in zip(header["is-cuda"], frames)
         ]
         assert all(