Merge branch 'branch-24.02' into fix-dep-strs-factory-calls

conda/environments/all_cuda-118_arch-x86_64.yaml

@@ -75,7 +75,7 @@ dependencies:
 - pydata-sphinx-theme!=0.14.2
 - pytest
 - pytest-benchmark
-- pytest-cases<3.8.2
+- pytest-cases>=3.8.2
 - pytest-cov
 - pytest-xdist
 - python-confluent-kafka>=1.9.0,<1.10.0a0

conda/environments/all_cuda-120_arch-x86_64.yaml

@@ -72,7 +72,7 @@ dependencies:
 - pydata-sphinx-theme!=0.14.2
 - pytest
 - pytest-benchmark
-- pytest-cases<3.8.2
+- pytest-cases>=3.8.2
 - pytest-cov
 - pytest-xdist
 - python-confluent-kafka>=1.9.0,<1.10.0a0

cpp/benchmarks/CMakeLists.txt

@@ -1,5 +1,5 @@
 # =============================================================================
-# Copyright (c) 2018-2023, NVIDIA CORPORATION.
+# Copyright (c) 2018-2024, NVIDIA CORPORATION.
 #
 # Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
 # in compliance with the License. You may obtain a copy of the License at
@@ -183,6 +183,11 @@ ConfigureNVBench(
   sort/sort_lists.cpp sort/sort_structs.cpp
 )
 
+# ##################################################################################################
+# * structs benchmark
+# --------------------------------------------------------------------------------
+ConfigureNVBench(STRUCT_CREATION_NVBENCH structs/create_structs.cpp)
+
 # ##################################################################################################
 # * quantiles benchmark
 # --------------------------------------------------------------------------------

cpp/benchmarks/structs/create_structs.cpp

@@ -0,0 +1,31 @@
+/*
+ * Copyright (c) 2024, NVIDIA CORPORATION.
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+#include <benchmarks/common/generate_nested_types.hpp>
+
+#include <nvbench/nvbench.cuh>
+
+void nvbench_create_structs(nvbench::state& state)
+{
+  state.exec(nvbench::exec_tag::sync,
+             [&](nvbench::launch& launch) { auto const table_ptr = create_structs_data(state); });
+}
+
+NVBENCH_BENCH(nvbench_create_structs)
+  .set_name("create_structs")
+  .add_int64_power_of_two_axis("NumRows", {10, 18, 26})
+  .add_int64_axis("Depth", {1, 8, 16})
+  .add_int64_axis("Nulls", {0, 1});

cpp/src/structs/utilities.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2020-2023, NVIDIA CORPORATION.
+ * Copyright (c) 2020-2024, NVIDIA CORPORATION.
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -17,6 +17,7 @@
 #include <cudf/column/column_factories.hpp>
 #include <cudf/detail/copy.hpp>
 #include <cudf/detail/null_mask.hpp>
+#include <cudf/detail/nvtx/ranges.hpp>
 #include <cudf/detail/structs/utilities.hpp>
 #include <cudf/detail/unary.hpp>
 #include <cudf/structs/structs_column_view.hpp>
@@ -229,6 +230,7 @@ std::unique_ptr<column> superimpose_nulls_no_sanitize(bitmask_type const* null_m
                                                       rmm::cuda_stream_view stream,
                                                       rmm::mr::device_memory_resource* mr)
 {
+  CUDF_FUNC_RANGE();
   if (input->type().id() == cudf::type_id::EMPTY) {
     // EMPTY columns should not have a null mask,
     // so don't superimpose null mask on empty columns.
@@ -258,19 +260,11 @@ std::unique_ptr<column> superimpose_nulls_no_sanitize(bitmask_type const* null_m
   // If the input is also a struct, repeat for all its children. Otherwise just return.
   if (input->type().id() != cudf::type_id::STRUCT) { return std::move(input); }
 
-  auto const current_mask   = input->view().null_mask();
   auto const new_null_count = input->null_count();  // this was just computed in the step above
   auto content              = input->release();
 
-  // Build new children columns.
-  std::for_each(content.children.begin(),
-                content.children.end(),
-                [current_mask, new_null_count, stream, mr](auto& child) {
-                  child = superimpose_nulls_no_sanitize(
-                    current_mask, new_null_count, std::move(child), stream, mr);
-                });
-
   // Replace the children columns.
+  // make_structs_column recursively calls superimpose_nulls
   return cudf::make_structs_column(num_rows,
                                    std::move(content.children),
                                    new_null_count,

dependencies.yaml

@@ -620,7 +620,7 @@ dependencies:
           - fastavro>=0.22.9
           - hypothesis
           - pytest-benchmark
-          - pytest-cases<3.8.2
+          - pytest-cases>=3.8.2
           - python-snappy>=0.6.0
           - scipy
       - output_types: conda

docs/cudf/source/conf.py

@@ -109,12 +109,12 @@ def clean_definitions(root):
     # All of these in type declarations cause Breathe to choke.
     # For friend, see https://github.com/breathe-doc/breathe/issues/916
     strings_to_remove = ("__forceinline__", "CUDF_HOST_DEVICE", "decltype(auto)", "friend")
-    for field in (".//type", ".//definition"):
-        for type_ in root.findall(field):
-            if type_.text is not None:
-                for string in strings_to_remove:
-                    type_.text = type_.text.replace(string, "")
-
+    for node in root.iter():
+        for string in strings_to_remove:
+            if node.text is not None:
+                node.text = node.text.replace(string, "")
+            if node.tail is not None:
+                node.tail = node.tail.replace(string, "")
 
 def clean_all_xml_files(path):
     for fn in glob.glob(os.path.join(path, "*.xml")):

python/cudf/benchmarks/API/bench_dataframe.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2022-2023, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 """Benchmarks of DataFrame methods."""
 
@@ -178,6 +178,8 @@ def bench_nsmallest(benchmark, dataframe, num_cols_to_sort, n):
     benchmark(dataframe.nsmallest, n, by)
 
 
-@pytest_cases.parametrize_with_cases("dataframe, cond, other", prefix="where")
+@pytest_cases.parametrize_with_cases(
+    "dataframe, cond, other", prefix="where", cases="cases_dataframe"
+)
 def bench_where(benchmark, dataframe, cond, other):
     benchmark(dataframe.where, cond, other)

python/cudf/benchmarks/API/bench_functions.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2022, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 """Benchmarks of free functions that accept cudf objects."""
 
@@ -9,7 +9,9 @@
 from utils import benchmark_with_object
 
 
-@pytest_cases.parametrize_with_cases("objs", prefix="concat")
+@pytest_cases.parametrize_with_cases(
+    "objs", prefix="concat", cases="cases_functions"
+)
 @pytest.mark.parametrize(
     "axis",
     [

python/cudf/benchmarks/API/bench_dataframe_cases.py → python/cudf/benchmarks/API/cases_dataframe.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2022, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 from utils import benchmark_with_object
 

python/cudf/benchmarks/API/bench_functions_cases.py → python/cudf/benchmarks/API/cases_functions.py

@@ -1,4 +1,4 @@
-# Copyright (c) 2022, NVIDIA CORPORATION.
+# Copyright (c) 2022-2024, NVIDIA CORPORATION.
 
 """Test cases for benchmarks in bench_functions.py."""
 

python/cudf/pyproject.toml

@@ -59,7 +59,7 @@ test = [
     "msgpack",
     "pytest",
     "pytest-benchmark",
-    "pytest-cases<3.8.2",
+    "pytest-cases>=3.8.2",
     "pytest-cov",
     "pytest-xdist",
     "python-snappy>=0.6.0",