Remove unneeded CUDF_EXPORT from some cudf/detail headers (rapidsai#21693)

Removes the `CUDF_EXPORT` decorator from many headers in the `cpp/include/cudf/detail` folder.
All of these have public interfaces and are not longer called outside of libcudf.
Also fixes the doxygen for some of the detail functions as appropriate.

Authors:
  - David Wendt (https://github.com/davidwendt)

Approvers:
  - Bradley Dice (https://github.com/bdice)
  - Nghia Truong (https://github.com/ttnghia)

URL: rapidsai#21693

Author: davidwendt

cpp/include/cudf/detail/aggregation/aggregation.hpp

@@ -18,7 +18,7 @@
 #include <numeric>
 #include <utility>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace detail {
 
 /**
@@ -1375,4 +1375,4 @@ void initialize_with_identity(mutable_table_view const& table,
                               rmm::cuda_stream_view stream);
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/aggregation/result_cache.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2019-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -11,7 +11,7 @@
 
 #include <unordered_map>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace detail {
 struct pair_column_aggregation_equal_to {
   bool operator()(std::pair<column_view, aggregation const&> const& lhs,
@@ -54,4 +54,4 @@ class result_cache {
 };
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/binaryop.hpp

@@ -1,25 +1,22 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2018-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2018-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 #pragma once
 
 #include <cudf/binaryop.hpp>
 #include <cudf/utilities/default_stream.hpp>
-#include <cudf/utilities/export.hpp>
 #include <cudf/utilities/memory_resource.hpp>
 
 #include <rmm/cuda_stream_view.hpp>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 //! Inner interfaces and implementations
 namespace detail {
 
 /**
  * @copydoc cudf::binary_operation(column_view const&, column_view const&,
- * std::string const&, data_type, rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * std::string const&, data_type, rmm::cuda_stream_view, rmm::device_async_resource_ref)
  */
 std::unique_ptr<column> binary_operation(column_view const& lhs,
                                          column_view const& rhs,
@@ -30,9 +27,7 @@ std::unique_ptr<column> binary_operation(column_view const& lhs,
 
 /**
  * @copydoc cudf::binary_operation(scalar const&, column_view const&, binary_operator,
- * data_type, rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * data_type, rmm::cuda_stream_view, rmm::device_async_resource_ref)
  */
 std::unique_ptr<column> binary_operation(scalar const& lhs,
                                          column_view const& rhs,
@@ -43,9 +38,7 @@ std::unique_ptr<column> binary_operation(scalar const& lhs,
 
 /**
  * @copydoc cudf::binary_operation(column_view const&, scalar const&, binary_operator,
- * data_type, rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * data_type, rmm::cuda_stream_view, rmm::device_async_resource_ref)
  */
 std::unique_ptr<column> binary_operation(column_view const& lhs,
                                          scalar const& rhs,
@@ -56,9 +49,7 @@ std::unique_ptr<column> binary_operation(column_view const& lhs,
 
 /**
  * @copydoc cudf::binary_operation(column_view const&, column_view const&,
- * binary_operator, data_type, rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * binary_operator, data_type, rmm::cuda_stream_view, rmm::device_async_resource_ref)
  */
 std::unique_ptr<column> binary_operation(column_view const& lhs,
                                          column_view const& rhs,
@@ -67,4 +58,4 @@ std::unique_ptr<column> binary_operation(column_view const& lhs,
                                          rmm::cuda_stream_view stream,
                                          rmm::device_async_resource_ref mr);
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/concatenate.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2020-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 #pragma once
@@ -8,34 +8,31 @@
 #include <cudf/concatenate.hpp>
 #include <cudf/table/table_view.hpp>
 #include <cudf/utilities/default_stream.hpp>
-#include <cudf/utilities/export.hpp>
 #include <cudf/utilities/memory_resource.hpp>
 #include <cudf/utilities/span.hpp>
 
 #include <rmm/cuda_stream_view.hpp>
 
 #include <vector>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 //! Inner interfaces and implementations
 namespace detail {
 /**
- * @copydoc cudf::concatenate(host_span<column_view const>,rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * @copydoc cudf::concatenate(host_span<column_view
+ * const>,rmm::cuda_stream_view,rmm::device_async_resource_ref)
  */
 std::unique_ptr<column> concatenate(host_span<column_view const> columns_to_concat,
                                     rmm::cuda_stream_view stream,
                                     rmm::device_async_resource_ref mr);
 
 /**
- * @copydoc cudf::concatenate(host_span<table_view const>,rmm::device_async_resource_ref)
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
+ * @copydoc cudf::concatenate(host_span<table_view
+ * const>,rmm::cuda_stream_view,rmm::device_async_resource_ref)
  */
 std::unique_ptr<table> concatenate(host_span<table_view const> tables_to_concat,
                                    rmm::cuda_stream_view stream,
                                    rmm::device_async_resource_ref mr);
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/concatenate_masks.hpp

@@ -1,20 +1,19 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2020-2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 #pragma once
 
 #include <cudf/column/column_device_view.cuh>
 #include <cudf/column/column_view.hpp>
-#include <cudf/utilities/export.hpp>
 #include <cudf/utilities/memory_resource.hpp>
 #include <cudf/utilities/span.hpp>
 
 #include <rmm/cuda_stream_view.hpp>
 #include <rmm/device_buffer.hpp>
 #include <rmm/mr/device_memory_resource.hpp>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 //! Inner interfaces and implementations
 namespace detail {
 
@@ -59,4 +58,4 @@ rmm::device_buffer concatenate_masks(host_span<column_view const> views,
                                      rmm::device_async_resource_ref mr);
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/contiguous_split.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2023-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -12,13 +12,11 @@
 
 #include <rmm/cuda_stream_view.hpp>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace detail {
 
 /**
  * @copydoc cudf::contiguous_split
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
  **/
 std::vector<packed_table> contiguous_split(cudf::table_view const& input,
                                            std::vector<size_type> const& splits,
@@ -27,8 +25,6 @@ std::vector<packed_table> contiguous_split(cudf::table_view const& input,
 
 /**
  * @copydoc cudf::pack
- *
- * @param stream Optional CUDA stream on which to execute kernels
  **/
 packed_columns pack(cudf::table_view const& input,
                     rmm::cuda_stream_view stream,
@@ -114,4 +110,4 @@ std::vector<uint8_t> pack_metadata(table_view const& table,
                                    metadata_builder& builder);
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/datetime.hpp

@@ -1,17 +1,18 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2021-2025, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
 #pragma once
 
+#include <cudf/datetime.hpp>
 #include <cudf/types.hpp>
 #include <cudf/utilities/default_stream.hpp>
 #include <cudf/utilities/memory_resource.hpp>
 
 #include <memory>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace datetime {
 namespace detail {
 /**
@@ -77,4 +78,4 @@ std::unique_ptr<cudf::column> extract_quarter(cudf::column_view const& column,
 
 }  // namespace detail
 }  // namespace datetime
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/fill.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2019-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -14,13 +14,11 @@
 
 #include <memory>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace detail {
 
 /**
  * @copydoc cudf::fill_in_place
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
  */
 void fill_in_place(mutable_column_view& destination,
                    size_type begin,
@@ -30,8 +28,6 @@ void fill_in_place(mutable_column_view& destination,
 
 /**
  * @copydoc cudf::fill
- *
- * @param stream CUDA stream used for device memory operations and kernel launches.
  */
 std::unique_ptr<column> fill(column_view const& input,
                              size_type begin,
@@ -41,4 +37,4 @@ std::unique_ptr<column> fill(column_view const& input,
                              rmm::device_async_resource_ref mr);
 
 }  // namespace detail
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/groupby.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2019-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 #pragma once
@@ -14,7 +14,7 @@
 #include <memory>
 #include <utility>
 
-namespace CUDF_EXPORT cudf {
+namespace cudf {
 namespace groupby::detail::hash {
 /**
  * @brief Indicates if a set of aggregation requests can be satisfied with a
@@ -35,4 +35,4 @@ std::pair<std::unique_ptr<table>, std::vector<aggregation_result>> groupby(
   rmm::cuda_stream_view stream,
   rmm::device_async_resource_ref mr);
 }  // namespace groupby::detail::hash
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf

cpp/include/cudf/detail/groupby/group_replace_nulls.hpp

@@ -1,5 +1,5 @@
 /*
- * SPDX-FileCopyrightText: Copyright (c) 2021-2024, NVIDIA CORPORATION.
+ * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION.
  * SPDX-License-Identifier: Apache-2.0
  */
 
@@ -13,18 +13,19 @@
 #include <cudf/utilities/span.hpp>
 
 #include <rmm/exec_policy.hpp>
-namespace CUDF_EXPORT cudf {
+
+namespace cudf {
 namespace groupby {
 namespace detail {
 
 /**
  * @brief Internal API to replace nulls with preceding/following non-null values in @p value
  *
- * @param[in] grouped_value A column whose null values will be replaced.
- * @param[in] group_labels Group labels for @p grouped_value, corresponding to group keys.
- * @param[in] replace_policy Specify the position of replacement values relative to null values.
+ * @param grouped_value A column whose null values will be replaced.
+ * @param group_labels Group labels for @p grouped_value, corresponding to group keys.
+ * @param replace_policy Specify the position of replacement values relative to null values.
  * @param stream CUDA stream used for device memory operations and kernel launches.
- * @param[in] mr Device memory resource used to allocate device memory of the returned column.
+ * @param mr Device memory resource used to allocate device memory of the returned column.
  */
 std::unique_ptr<column> group_replace_nulls(cudf::column_view const& grouped_value,
                                             device_span<size_type const> group_labels,
@@ -34,4 +35,4 @@ std::unique_ptr<column> group_replace_nulls(cudf::column_view const& grouped_val
 
 }  // namespace detail
 }  // namespace groupby
-}  // namespace CUDF_EXPORT cudf
+}  // namespace cudf