Add license to stub fuile, add comment about syncing impl and stubs

Add comments stating the need to keep implementation and Python stub
file in sync to both files. In the stub file to comment documents
use of mypy's stubgen to generate stubs and calls to compare that against
current stubs. It also calls out the need to keep docstrings and
doctring examples in sync with implementation.

Author: oleksandr-pavlyk

python/cuda/nvbench/__init__.pyi

@@ -1,4 +1,29 @@
-# from __future__ import annotations
+# Copyright 2025 NVIDIA Corporation
+#
+#  Licensed under the Apache License, Version 2.0 with the LLVM exception
+#  (the "License"); you may not use this file except in compliance with
+#  the License.
+#
+#  You may obtain a copy of the License at
+#
+#      http://llvm.org/foundation/relicensing/LICENSE.txt
+#
+#  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.
+
+# ============================================
+# PLEASE KEEP IN SYNC WITH py_nvbench.cpp FILE
+# ============================================
+# Please be sure to keep these type hints and docstring in sync
+# with the pybind11 bindings in ``../../src/py_nvbench.cpp``
+
+# Use mypy's stubgen to auto-generate stubs using
+# ``stubgen -m cuda.nvbench._nvbench`` and compare
+# stubs in generated out/cuda/nvbench/_nvbench.pyi
+# with definitions given here.
 
 from collections.abc import Callable, Sequence
 from typing import Optional, Self, Union
@@ -22,11 +47,11 @@ class CudaStream:
             import cuda.nvbench as nvbench
 
             def bench(state: nvbench.State):
-                dev = core.Device(state.getDevice())
+                dev = core.Device(state.get_device())
                 dev.set_current()
                 # converts CudaString to core.Stream
                 # using __cuda_stream__ protocol
-                dev.create_stream(state.getStream())
+                dev.create_stream(state.get_stream())
         """
         ...
 

python/src/py_nvbench.cpp

@@ -209,6 +209,12 @@ constinit std::unique_ptr<GlobalBenchmarkRegistry, py::nodelete> global_registry
 
 } // end of anonymous namespace
 
+// ==========================================
+// PLEASE KEEP IN SYNC WITH __init__.pyi FILE
+// ==========================================
+// If you modify these bindings, please be sure to update the
+// corresponding type hints in ``../cuda/nvbench/__init__.pyi``
+
 PYBIND11_MODULE(_nvbench, m)
 {
   // == STEP 1
@@ -223,9 +229,6 @@ PYBIND11_MODULE(_nvbench, m)
   // It is reinitialized before running all benchmarks to set devices to use
   nvbench::benchmark_manager::get().initialize();
 
-  // TODO: Use cuModuleGetLoadingMode(&mode) to confirm that (mode == CU_MODULE_EAGER_LOADING)
-  // and issue warning otherwise
-
   // == STEP 2
   // Define CudaStream class
   //    ATTN: nvbench::cuda_stream is move-only class