Add instructions on how to run clang-tidy locally.

Adds dedicated conda environment.

CONTRIBUTING.md

@@ -115,6 +115,46 @@ please see the `.pre-commit-config.yaml` file.
   of files are up-to-date and in the correct format.
 - `codespell`: Checks for spelling mistakes
 
+### Clang-tidy
+
+In order to maintain high-quality code, cuML uses not only pre-commit hooks
+featuring various formatters and linters but also the clang-tidy tool.
+Clang-tidy is designed to detect potential issues within the C and C++ code. It
+is typically run as part of our continuous integration (CI) process.
+
+While it's generally unnecessary for contributors to run clang-tidy locally,
+there might be cases where you would want to do so. There are two primary
+methods to run clang-tidy on your local machine: using Docker or Conda.
+
+* **Docker**
+
+    1. Navigate to the repository root directory.
+    2. Run the following Docker command:
+
+        ```bash
+        docker run --rm --pull always \
+            --mount type=bind,source="$(pwd)",target=/opt/repo --workdir /opt/repo \
+            rapidsai/ci:latest /opt/repo/ci/run_clang_tidy.sh
+        ```
+
+
+* **Conda**
+
+    1. Navigate to the repository root directory.
+    2. Create and activate the needed conda environment:
+        ```bash
+        conda env create --force -n cuml-clang-tidy -f conda/environments/clang_tidy_cuda-118_arch-x86_64.yaml
+        conda activate cuml-clang-tidy
+        ```
+    3. Generate the compile command database with
+        ```bash
+        ./build.sh --configure-only libcuml
+        ```
+    3. Run clang-tidy with the following command:
+        ```bash
+        python cpp/scripts/run-clang-tidy.py --config pyproject.toml
+        ```
+
 ### Managing PR labels
 
 Each PR must be labeled according to whether it is a "breaking" or "non-breaking" change (using Github labels). This is used to highlight changes that users should know about when upgrading.

ci/run_clang_tidy.sh

@@ -19,4 +19,4 @@ set +u && conda activate clang_tidy && set -u
 
 rapids-logger "Run clang-tidy"
 
-python cpp/scripts/run-clang-tidy.py
+python cpp/scripts/run-clang-tidy.py --config pyproject.toml

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

@@ -0,0 +1,37 @@
+# This file is generated by `rapids-dependency-file-generator`.
+# To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`.
+channels:
+- rapidsai
+- rapidsai-nightly
+- dask/label/dev
+- conda-forge
+- nvidia
+dependencies:
+- c-compiler
+- clang-tools==15.0.7
+- clang==15.0.7
+- cmake>=3.26.4
+- cudatoolkit=11.8
+- cxx-compiler
+- gcc_linux-64=11.*
+- gmock>=1.13.0
+- gtest>=1.13.0
+- libcublas-dev=11.11.3.6
+- libcublas=11.11.3.6
+- libcufft-dev=10.9.0.58
+- libcufft=10.9.0.58
+- libcumlprims==23.8.*
+- libcurand-dev=10.3.0.86
+- libcurand=10.3.0.86
+- libcusolver-dev=11.4.1.48
+- libcusolver=11.4.1.48
+- libcusparse-dev=11.7.5.86
+- libcusparse=11.7.5.86
+- libraft-headers==23.8.*
+- libraft==23.8.*
+- librmm==23.8.*
+- ninja
+- nvcc_linux-64=11.8
+- sysroot_linux-64==2.17
+- tomli
+name: clang_tidy_cuda-118_arch-x86_64

dependencies.yaml

@@ -27,7 +27,10 @@ files:
       - checks
       - py_version
   clang_tidy:
-    output: none
+    output: conda
+    matrix:
+      cuda: ["11.8"]
+      arch: [x86_64]
     includes:
       - clang_tidy
       - common_build