Add cuDF devcontainers (#14015)

* add devcontainers

* fix tag for CUDA 12.0

* use CUDA 11.8 for now

* default to CUDA 12.0

* install cuda-cupti-dev in conda environment

* remove MODIFY_PREFIX_PATH so the driver is found

* install cuda-nvtx-dev in conda environment

* update conda env

* add MODIFY_PREFIX_PATH back

* temporarily default to my branch with the fix for MODIFY_PREFIX_PATH in conda envs

* remove temporary rapids-cmake pin

* build all RAPIDS archs to take maximum advantage of sccache

* add clangd and nsight vscode customizations

* copy in default clangd config

* remove options for pip vs. conda unless using the launch script

* fix unified mounts

* ensure dirs exist before mounting

* add compile_commands to .gitignore

* allow defining cudf and cudf_kafka include dirs via envvars

* add kvikio

* use volumes for isolated devcontainer source dirs

* update README.md

* update to rapidsai/devcontainers 23.10

* update rapids-build-utils version to 23.10

* add .clangd config file

* update RAPIDS versions in devcontainer files

* ensure the directory for the generated jitify kernels is exists after configuring

* add clang and clang-tools 16

* remove isolated and unified devcontainers, make single the default

* separate CUDA 11.8 and 12.0 devcontainers

* fix version string for requirements.txt

* update conda envs

* clean up envvars, mounts, and build args, add codespaces post-attach command workaround

* consolidate common vscode customizations

* enumerate CUDA 11 packages, include up to CUDA 12.2

* include protoc-wheel when generating requirements.txt

* default to cuda-python for cu11

* separate devcontainer mounts by CUDA version

* add devcontainer build jobs to PR workflow

* use pypi.nvidia.com instead of pypi.ngc.nvidia.com

* fix venvs mount path

* fix lint

* ensure rmm-cuXX is included in pip requirements

* disable libcudf_kakfa build for now

* build dask-cudf

* be more explicit in update-versions.sh, make devcontainer build required in pr jobs

* revert rename devcontainer job

* install librdkafka-dev in pip containers so we can build libcudf_kafka and cudf_kafka

* separate cupy, cudf, and cudf_kafka matrices for CUDA 11 and 12

* add fallback include path for RMM

* fallback to CUDA_PATH if CUDA_HOME is not set

* define envvars in dockerfile

* define envvars for cudf_kafka

* build verbose

* include wheel and setuptools in requirements.txt

* switch workflow to branch-23.10

* update clang-tools version to 16.0.6

* fix update-version.sh

* Use 24.02 branches.

* fix version numbers

* Fix dependencies.yaml.

* Update .devcontainer/Dockerfile

---------

Co-authored-by: Bradley Dice <[email protected]>

.devcontainer/Dockerfile

@@ -0,0 +1,35 @@
+# syntax=docker/dockerfile:1.5
+
+ARG BASE
+ARG PYTHON_PACKAGE_MANAGER=conda
+
+FROM ${BASE} as pip-base
+
+RUN apt update -y \
+ && DEBIAN_FRONTEND=noninteractive apt install -y \
+    librdkafka-dev \
+ && rm -rf /tmp/* /var/tmp/* /var/cache/apt/* /var/lib/apt/lists/*;
+
+ENV DEFAULT_VIRTUAL_ENV=rapids
+
+FROM ${BASE} as conda-base
+
+ENV DEFAULT_CONDA_ENV=rapids
+
+FROM ${PYTHON_PACKAGE_MANAGER}-base
+
+ARG CUDA
+ENV CUDAARCHS="RAPIDS"
+ENV CUDA_VERSION="${CUDA_VERSION:-${CUDA}}"
+
+ARG PYTHON_PACKAGE_MANAGER
+ENV PYTHON_PACKAGE_MANAGER="${PYTHON_PACKAGE_MANAGER}"
+
+ENV PYTHONSAFEPATH="1"
+ENV PYTHONUNBUFFERED="1"
+ENV PYTHONDONTWRITEBYTECODE="1"
+
+ENV SCCACHE_REGION="us-east-2"
+ENV SCCACHE_BUCKET="rapids-sccache-devs"
+ENV VAULT_HOST="https://vault.ops.k8s.rapids.ai"
+ENV HISTFILE="/home/coder/.cache/._bash_history"

.devcontainer/README.md

@@ -0,0 +1,35 @@
+# cuDF Development Containers
+
+This directory contains [devcontainer configurations](https://containers.dev/implementors/json_reference/) for using VSCode to [develop in a container](https://code.visualstudio.com/docs/devcontainers/containers) via the `Remote Containers` [extension](https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-containers) or [GitHub Codespaces](https://github.com/codespaces).
+
+This container is a turnkey development environment for building and testing the cuDF C++ and Python libraries.
+
+## Table of Contents
+
+* [Prerequisites](#prerequisites)
+* [Host bind mounts](#host-bind-mounts)
+* [Launch a Dev Container](#launch-a-dev-container)
+
+## Prerequisites
+
+* [VSCode](https://code.visualstudio.com/download)
+* [VSCode Remote Containers extension](https://marketplace.visualstudio.com/items?itemName=ms-vscode-remote.remote-containers)
+
+## Host bind mounts
+
+By default, the following directories are bind-mounted into the devcontainer:
+
+* `${repo}:/home/coder/cudf`
+* `${repo}/../.aws:/home/coder/.aws`
+* `${repo}/../.local:/home/coder/.local`
+* `${repo}/../.cache:/home/coder/.cache`
+* `${repo}/../.conda:/home/coder/.conda`
+* `${repo}/../.config:/home/coder/.config`
+
+This ensures caches, configurations, dependencies, and your commits are persisted on the host across container runs.
+
+## Launch a Dev Container
+
+To launch a devcontainer from VSCode, open the cuDF repo and select the "Reopen in Container" button in the bottom right:<br/><img src="https://user-images.githubusercontent.com/178183/221771999-97ab29d5-e718-4e5f-b32f-2cdd51bba25c.png"/>
+
+Alternatively, open the VSCode command palette (typically `cmd/ctrl + shift + P`) and run the "Rebuild and Reopen in Container" command.

.devcontainer/cuda11.8-conda/devcontainer.json

@@ -0,0 +1,37 @@
+{
+  "build": {
+    "context": "${localWorkspaceFolder}/.devcontainer",
+    "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
+    "args": {
+      "CUDA": "11.8",
+      "PYTHON_PACKAGE_MANAGER": "conda",
+      "BASE": "rapidsai/devcontainers:24.02-cpp-llvm16-cuda11.8-mambaforge-ubuntu22.04"
+    }
+  },
+  "hostRequirements": {"gpu": "optional"},
+  "features": {
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.2": {}
+  },
+  "overrideFeatureInstallOrder": [
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
+  ],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config,conda/pkgs,conda/${localWorkspaceFolderBasename}-cuda11.8-envs}"],
+  "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
+  "workspaceFolder": "/home/coder",
+  "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/cudf,type=bind,consistency=consistent",
+  "mounts": [
+    "source=${localWorkspaceFolder}/../.aws,target=/home/coder/.aws,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.conda/pkgs,target=/home/coder/.conda/pkgs,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.conda/${localWorkspaceFolderBasename}-cuda11.8-envs,target=/home/coder/.conda/envs,type=bind,consistency=consistent"
+  ],
+  "customizations": {
+    "vscode": {
+      "extensions": [
+        "ms-python.flake8",
+        "nvidia.nsight-vscode-edition"
+      ]
+    }
+  }
+}

.devcontainer/cuda11.8-pip/devcontainer.json

@@ -0,0 +1,36 @@
+{
+  "build": {
+    "context": "${localWorkspaceFolder}/.devcontainer",
+    "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
+    "args": {
+      "CUDA": "11.8",
+      "PYTHON_PACKAGE_MANAGER": "pip",
+      "BASE": "rapidsai/devcontainers:24.02-cpp-llvm16-cuda11.8-ubuntu22.04"
+    }
+  },
+  "hostRequirements": {"gpu": "optional"},
+  "features": {
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.2": {}
+  },
+  "overrideFeatureInstallOrder": [
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
+  ],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config/pip,local/share/${localWorkspaceFolderBasename}-cuda11.8-venvs}"],
+  "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
+  "workspaceFolder": "/home/coder",
+  "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/cudf,type=bind,consistency=consistent",
+  "mounts": [
+    "source=${localWorkspaceFolder}/../.aws,target=/home/coder/.aws,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.local/share/${localWorkspaceFolderBasename}-cuda11.8-venvs,target=/home/coder/.local/share/venvs,type=bind,consistency=consistent"
+  ],
+  "customizations": {
+    "vscode": {
+      "extensions": [
+        "ms-python.flake8",
+        "nvidia.nsight-vscode-edition"
+      ]
+    }
+  }
+}

.devcontainer/cuda12.0-conda/devcontainer.json

@@ -0,0 +1,37 @@
+{
+  "build": {
+    "context": "${localWorkspaceFolder}/.devcontainer",
+    "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
+    "args": {
+      "CUDA": "12.0",
+      "PYTHON_PACKAGE_MANAGER": "conda",
+      "BASE": "rapidsai/devcontainers:24.02-cpp-mambaforge-ubuntu22.04"
+    }
+  },
+  "hostRequirements": {"gpu": "optional"},
+  "features": {
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.2": {}
+  },
+  "overrideFeatureInstallOrder": [
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
+  ],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config,conda/pkgs,conda/${localWorkspaceFolderBasename}-cuda12.0-envs}"],
+  "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
+  "workspaceFolder": "/home/coder",
+  "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/cudf,type=bind,consistency=consistent",
+  "mounts": [
+    "source=${localWorkspaceFolder}/../.aws,target=/home/coder/.aws,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.conda/pkgs,target=/home/coder/.conda/pkgs,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.conda/${localWorkspaceFolderBasename}-cuda12.0-envs,target=/home/coder/.conda/envs,type=bind,consistency=consistent"
+  ],
+  "customizations": {
+    "vscode": {
+      "extensions": [
+        "ms-python.flake8",
+        "nvidia.nsight-vscode-edition"
+      ]
+    }
+  }
+}

.devcontainer/cuda12.0-pip/devcontainer.json

@@ -0,0 +1,36 @@
+{
+  "build": {
+    "context": "${localWorkspaceFolder}/.devcontainer",
+    "dockerfile": "${localWorkspaceFolder}/.devcontainer/Dockerfile",
+    "args": {
+      "CUDA": "12.0",
+      "PYTHON_PACKAGE_MANAGER": "pip",
+      "BASE": "rapidsai/devcontainers:24.02-cpp-llvm16-cuda12.0-ubuntu22.04"
+    }
+  },
+  "hostRequirements": {"gpu": "optional"},
+  "features": {
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils:24.2": {}
+  },
+  "overrideFeatureInstallOrder": [
+    "ghcr.io/rapidsai/devcontainers/features/rapids-build-utils"
+  ],
+  "initializeCommand": ["/bin/bash", "-c", "mkdir -m 0755 -p ${localWorkspaceFolder}/../.{aws,cache,config/pip,local/share/${localWorkspaceFolderBasename}-cuda12.0-venvs}"],
+  "postAttachCommand": ["/bin/bash", "-c", "if [ ${CODESPACES:-false} = 'true' ]; then . devcontainer-utils-post-attach-command; . rapids-post-attach-command; fi"],
+  "workspaceFolder": "/home/coder",
+  "workspaceMount": "source=${localWorkspaceFolder},target=/home/coder/cudf,type=bind,consistency=consistent",
+  "mounts": [
+    "source=${localWorkspaceFolder}/../.aws,target=/home/coder/.aws,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.cache,target=/home/coder/.cache,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.config,target=/home/coder/.config,type=bind,consistency=consistent",
+    "source=${localWorkspaceFolder}/../.local/share/${localWorkspaceFolderBasename}-cuda12.0-venvs,target=/home/coder/.local/share/venvs,type=bind,consistency=consistent"
+  ],
+  "customizations": {
+    "vscode": {
+      "extensions": [
+        "ms-python.flake8",
+        "nvidia.nsight-vscode-edition"
+      ]
+    }
+  }
+}

.github/workflows/pr.yaml

@@ -25,6 +25,7 @@ jobs:
       - wheel-tests-cudf
       - wheel-build-dask-cudf
       - wheel-tests-dask-cudf
+      - devcontainer
       - unit-tests-cudf-pandas
       - pandas-tests
       #- pandas-tests-diff
@@ -130,6 +131,14 @@ jobs:
       matrix_filter: map(select(.ARCH == "amd64" and .PY_VER == "3.10" and (.CUDA_VER == "11.8.0" or .CUDA_VER == "12.0.1")))
       build_type: pull-request
       script: ci/test_wheel_dask_cudf.sh
+  devcontainer:
+    secrets: inherit
+    uses: rapidsai/shared-action-workflows/.github/workflows/[email protected]
+    with:
+      build_command: |
+        sccache -z;
+        build-all -DBUILD_BENCHMARKS=ON -DNVBench_ENABLE_CUPTI=OFF --verbose;
+        sccache -s;
   unit-tests-cudf-pandas:
     needs: wheel-build-cudf
     secrets: inherit

.gitignore

@@ -173,3 +173,7 @@ jupyter_execute
 
 # cibuildwheel
 /wheelhouse
+
+# clang tooling
+compile_commands.json
+.clangd/

.pre-commit-config.yaml

@@ -145,7 +145,7 @@ repos:
                     ^CHANGELOG.md$
                   )
       - repo: https://github.com/rapidsai/dependency-file-generator
-        rev: v1.5.1
+        rev: v1.7.1
         hooks:
             - id: rapids-dependency-file-generator
               args: ["--clean"]

ci/release/update-version.sh

@@ -115,3 +115,9 @@ NEXT_FULL_JAVA_TAG="${NEXT_SHORT_TAG}.${PATCH_PEP440}-SNAPSHOT"
 sed_runner "s|<version>.*-SNAPSHOT</version>|<version>${NEXT_FULL_JAVA_TAG}</version>|g" java/pom.xml
 sed_runner "s/branch-.*/branch-${NEXT_SHORT_TAG}/g" java/ci/README.md
 sed_runner "s/cudf-.*-SNAPSHOT/cudf-${NEXT_FULL_JAVA_TAG}/g" java/ci/README.md
+
+# .devcontainer files
+find .devcontainer/ -type f -name devcontainer.json -print0 | while IFS= read -r -d '' filename; do
+    sed_runner "s@rapidsai/devcontainers:[0-9.]*@rapidsai/devcontainers:${NEXT_SHORT_TAG}@g" "${filename}"
+    sed_runner "s@rapidsai/devcontainers/features/rapids-build-utils:[0-9.]*@rapidsai/devcontainers/features/rapids-build-utils:${NEXT_SHORT_TAG_PEP440}@" "${filename}"
+done

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

@@ -14,6 +14,8 @@ dependencies:
 - botocore>=1.24.21
 - c-compiler
 - cachetools
+- clang-tools=16.0.6
+- clang==16.0.6
 - cmake>=3.26.4
 - cramjam
 - cubinlinker

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

@@ -14,6 +14,8 @@ dependencies:
 - botocore>=1.24.21
 - c-compiler
 - cachetools
+- clang-tools=16.0.6
+- clang==16.0.6
 - cmake>=3.26.4
 - cramjam
 - cuda-cudart-dev

cpp/.clangd

@@ -0,0 +1,65 @@
+# https://clangd.llvm.org/config
+
+# Apply a config conditionally to all C files
+If:
+  PathMatch: .*\.(c|h)$
+
+---
+
+# Apply a config conditionally to all C++ files
+If:
+  PathMatch: .*\.(c|h)pp
+
+---
+
+# Apply a config conditionally to all CUDA files
+If:
+  PathMatch: .*\.cuh?
+CompileFlags:
+  Add:
+    - "-x"
+    - "cuda"
+    # No error on unknown CUDA versions
+    - "-Wno-unknown-cuda-version"
+    # Allow variadic CUDA functions
+    - "-Xclang=-fcuda-allow-variadic-functions"
+Diagnostics:
+  Suppress:
+    - "variadic_device_fn"
+    - "attributes_not_allowed"
+
+---
+
+# Tweak the clangd parse settings for all files
+CompileFlags:
+  Add:
+    # report all errors
+    - "-ferror-limit=0"
+    - "-fmacro-backtrace-limit=0"
+    - "-ftemplate-backtrace-limit=0"
+    # Skip the CUDA version check
+    - "--no-cuda-version-check"
+  Remove:
+    # remove gcc's -fcoroutines
+    - -fcoroutines
+    # remove nvc++ flags unknown to clang
+    - "-gpu=*"
+    - "-stdpar*"
+    # remove nvcc flags unknown to clang
+    - "-arch*"
+    - "-gencode*"
+    - "--generate-code*"
+    - "-ccbin*"
+    - "-t=*"
+    - "--threads*"
+    - "-Xptxas*"
+    - "-Xcudafe*"
+    - "-Xfatbin*"
+    - "-Xcompiler*"
+    - "--diag-suppress*"
+    - "--diag_suppress*"
+    - "--compiler-options*"
+    - "--expt-extended-lambda"
+    - "--expt-relaxed-constexpr"
+    - "-forward-unknown-to-host-compiler"
+    - "-Werror=cross-execution-space-call"

cpp/cmake/Modules/JitifyPreprocessKernels.cmake

@@ -1,5 +1,5 @@
 # =============================================================================
-# Copyright (c) 2021-2022, NVIDIA CORPORATION.
+# Copyright (c) 2021-2023, 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
@@ -54,6 +54,10 @@ function(jit_preprocess_files)
   )
 endfunction()
 
+if(NOT (EXISTS "${CUDF_GENERATED_INCLUDE_DIR}/include"))
+  make_directory("${CUDF_GENERATED_INCLUDE_DIR}/include")
+endif()
+
 jit_preprocess_files(
   SOURCE_DIRECTORY ${CUDF_SOURCE_DIR}/src FILES binaryop/jit/kernel.cu transform/jit/kernel.cu
   rolling/jit/kernel.cu