diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 841a02f72e6..f83d7c5b759 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -55,87 +55,17 @@ implementation of the issue, ask them in the issue instead of the PR. The following instructions are for developers and contributors to cuDF OSS development. These instructions are tested on Linux Ubuntu 16.04 & 18.04. Use these instructions to build cuDF from source and contribute to its development. Other operating systems may be compatible, but are not currently tested. -### Code Formatting - -#### Python - -cuDF uses [Black](https://black.readthedocs.io/en/stable/), -[isort](https://readthedocs.org/projects/isort/), and -[flake8](http://flake8.pycqa.org/en/latest/) to ensure a consistent code format -throughout the project. `Black`, `isort`, and `flake8` can be installed with -`conda` or `pip`: - -```bash -conda install black isort flake8 -``` - -```bash -pip install black isort flake8 -``` -These tools are used to auto-format the Python code, as well as check the Cython -code in the repository. Additionally, there is a CI check in place to enforce -that committed code follows our standards. You can use the tools to -automatically format your python code by running: -```bash -isort --atomic python/**/*.py -black python -``` +### General requirements -and then check the syntax of your Python and Cython code by running: - -```bash -flake8 python -flake8 --config=python/.flake8.cython -``` - -Additionally, many editors have plugins that will apply `isort` and `Black` as -you edit files, as well as use `flake8` to report any style / syntax issues. - -#### C++/CUDA - -cuDF uses [`clang-format`](https://clang.llvm.org/docs/ClangFormat.html) - -In order to format the C++/CUDA files, navigate to the root (`cudf`) directory and run: -``` -python3 ./cpp/scripts/run-clang-format.py -inplace -``` - -Additionally, many editors have plugins or extensions that you can set up to automatically run `clang-format` either manually or on file save. - -#### Pre-commit hooks - -Optionally, you may wish to setup [pre-commit hooks](https://pre-commit.com/) -to automatically run `isort`, `Black`, `flake8` and `clang-format` when you make a git commit. -This can be done by installing `pre-commit` via `conda` or `pip`: - -```bash -conda install -c conda-forge pre_commit -``` - -```bash -pip install pre-commit -``` - -and then running: - -```bash -pre-commit install -``` - -from the root of the cuDF repository. Now `isort`, `Black`, `flake8` and `clang-format` will be -run each time you commit changes. - -### Get libcudf Dependencies - -Compiler requirements: +Compilers: * `gcc` version 9.3+ * `nvcc` version 11.0+ * `cmake` version 3.20.1+ -CUDA/GPU requirements: +CUDA/GPU: * CUDA 11.0+ * NVIDIA driver 450.80.02+ @@ -143,11 +73,7 @@ CUDA/GPU requirements: You can obtain CUDA from [https://developer.nvidia.com/cuda-downloads](https://developer.nvidia.com/cuda-downloads). -## Script to build cuDF from source - -### Build from Source - -To install cuDF from source, ensure the dependencies are met and follow the steps below: +### Create the build Environment - Clone the repository and submodules ```bash @@ -166,86 +92,147 @@ conda activate cudf_dev ``` - For other CUDA versions, check the corresponding cudf_dev_cuda*.yml file in conda/environments -- Build and install `libcudf` after its dependencies. CMake depends on the `nvcc` executable being on your path or defined in `$CUDACXX`. +### Build cuDF from source + +- A `build.sh` script is provided in `$CUDF_HOME`. Running the script with no additional arguments will install the `libcudf`, `cudf` and `dask_cudf` libraries. By default, the libraries are installed to the `$CONDA_PREFIX` directory. To install into a different location, set the location in `$INSTALL_PREFIX`. Finally, note that the script depends on the `nvcc` executable being on your path, or defined in `$CUDACXX`. +```bash +cd $CUDF_HOME + +# Choose one of the following commands, depending on whether +# you want to build and install the libcudf C++ library only, +# or include the cudf and/or dask_cudf Python libraries: + +./build.sh # libcudf, cudf and dask_cudf +./build.sh libcudf # libcudf only +./build.sh libcudf cudf # libcudf and cudf only +``` +- Other libraries like `cudf-kafka` and `custreamz` can be installed with this script. For the complete list of libraries as well as details about the script usage, run the `help` command: +```bash +./build.sh --help +``` + +### Build, install and test cuDF libraries for contributors + +The general workflow is provided below. Please, also see the last section about [code formatting](###code-formatting). + +#### `libcudf` (C++) + +If you're only interested in building the library (and not the unit tests): + +```bash +cd $CUDF_HOME +./build.sh libcudf +``` +If, in addition, you want to build tests: + +```bash +./build.sh libcudf tests +``` +To run the tests: + ```bash -$ cd $CUDF_HOME/cpp # navigate to C/C++ CUDA source root directory -$ mkdir build # make a build directory -$ cd build # enter the build directory +make test +``` -# CMake options: -# -DCMAKE_INSTALL_PREFIX set to the install path for your libraries or $CONDA_PREFIX if you're using Anaconda, i.e. -DCMAKE_INSTALL_PREFIX=/install/path or -DCMAKE_INSTALL_PREFIX=$CONDA_PREFIX -$ cmake .. -DCMAKE_INSTALL_PREFIX=$CONDA_PREFIX # configure cmake ... -$ make -j # compile the libraries librmm.so, libcudf.so ... '-j' will start a parallel job using the number of physical cores available on your system -$ make install # install the libraries librmm.so, libcudf.so to the CMAKE_INSTALL_PREFIX +#### `cudf` (Python) + +- First, build the `libcudf` C++ library following the steps above + +- To build and install in edit/develop `cudf` python package: +```bash +cd $CUDF_HOME/python/cudf +python setup.py build_ext --inplace +python setup.py develop ``` -- As a convenience, a `build.sh` script is provided in `$CUDF_HOME`. To execute the same build commands above, run the script as shown below. Note that the libraries will be installed to the location set in `$INSTALL_PREFIX` if set (i.e. `export INSTALL_PREFIX=/install/path`), otherwise to `$CONDA_PREFIX`. +- To run `cudf` tests : ```bash -$ cd $CUDF_HOME -$ ./build.sh # To build both C++ and Python cuDF versions with their dependencies +cd $CUDF_HOME/python +py.test -v cudf/cudf/tests ``` -- To build only the C++ component with the script + +#### `dask-cudf` (Python) + +- First, build the `libcudf` C++ and `cudf` Python libraries following the steps above + +- To install in edit/develop mode the `dask-cudf` python package: ```bash -$ ./build.sh libcudf # Build only the cuDF C++ components and install them to $INSTALL_PREFIX if set, otherwise $CONDA_PREFIX +cd $CUDF_HOME/python/dask_cudf +python setup.py build_ext --inplace +python setup.py develop ``` -- To run tests (Optional): +- To run `dask_cudf` tests : ```bash -$ make test +cd $CUDF_HOME/python +py.test -v dask_cudf ``` -- Build the `cudf` python package, in the `python/cudf` folder: + +#### `libcudf_kafka` (C++) + +If you're only interested in building the library (and not the unit tests): + ```bash -$ cd $CUDF_HOME/python/cudf -$ python setup.py build_ext --inplace -$ python setup.py install +cd $CUDF_HOME +./build.sh libcudf_kafka ``` +If, in addition, you want to build tests: -- Like the `libcudf` build step above, `build.sh` can also be used to build the `cudf` python package, as shown below: ```bash -$ cd $CUDF_HOME -$ ./build.sh cudf +./build.sh libcudf_kafka tests ``` +To run the tests: -- Additionally to build the `dask-cudf` python package, in the `python/dask_cudf` folder: ```bash -$ cd $CUDF_HOME/python/dask_cudf -$ python setup.py install +make test ``` -- The `build.sh` script can also be used to build the `dask-cudf` python package, as shown below: +#### `cudf-kafka` (Python) + +- First, build the `libcudf` and `libcudf_kafka` following the steps above + +- To install in edit/develop mode the `cudf-kafka` python package: ```bash -$ cd $CUDF_HOME -$ ./build.sh dask_cudf +cd $CUDF_HOME/python/cudf_kafka +python setup.py build_ext --inplace +python setup.py develop ``` -- To run Python tests (Optional): +#### `custreamz` (Python) + +- First, build `libcudf`, `libcudf_kafka`, and `cudf_kafka` following the steps above + +- To install in edit/develop mode the `custreamz` python package: ```bash -$ cd $CUDF_HOME/python -$ py.test -v cudf # run cudf test suite -$ py.test -v dask_cudf # run dask_cudf test suite +cd $CUDF_HOME/python/custreamz +python setup.py build_ext --inplace +python setup.py develop ``` -- Other `build.sh` options: +- To run `custreamz` tests : ```bash -$ cd $CUDF_HOME -$ ./build.sh clean # remove any prior build artifacts and configuration (start over) -$ ./build.sh libcudf -v # compile and install libcudf with verbose output -$ ./build.sh libcudf -g # compile and install libcudf for debug -$ PARALLEL_LEVEL=4 ./build.sh libcudf # compile and install libcudf limiting parallel build jobs to 4 (make -j4) -$ ./build.sh libcudf -n # compile libcudf but do not install +cd $CUDF_HOME/python +py.test -v custreamz ``` -Done! You are ready to develop for the cuDF OSS project. +#### `cudf` (Java): + +- First, build the `libcudf` C++ library following the steps above + +- Then, refer to [Java README](https://github.com/rapidsai/cudf/blob/branch-21.10/java/README.md) + + +Done! You are ready to develop for the cuDF OSS project. But please go to [code formatting](###code-formatting) to ensure that you contributing code follows the expected format. ## Debugging cuDF ### Building Debug mode from source -Follow the [above instructions](#build-from-source) to build from source and add `-DCMAKE_BUILD_TYPE=Debug` to the `cmake` step. +Follow the [above instructions](####build-cudf-from-source) to build from source and add `-g` to the `./build.sh` command. For example: ```bash -$ cmake .. -DCMAKE_INSTALL_PREFIX=/install/path -DCMAKE_BUILD_TYPE=Debug # configure cmake ... use -DCMAKE_INSTALL_PREFIX=$CONDA_PREFIX if you're using Anaconda +./build.sh libcudf -g ``` This builds `libcudf` in Debug mode which enables some `assert` safety checks and includes symbols in the library for debugging. @@ -289,6 +276,7 @@ You can then use `cuda-dbg` to debug into the kernels in that source file. Before submitting a pull request, you can do a local build and test on your machine that mimics our gpuCI environment using the `ci/local/build.sh` script. For detailed information on usage of this script, see [here](ci/local/README.md). + ## Automated Build in Docker Container A Dockerfile is provided with a preconfigured conda environment for building and installing cuDF from source based off of the main branch. @@ -303,11 +291,11 @@ A Dockerfile is provided with a preconfigured conda environment for building and From cudf project root run the following, to build with defaults: ```bash -$ docker build --tag cudf . +docker build --tag cudf . ``` After the container is built run the container: ```bash -$ docker run --runtime=nvidia -it cudf bash +docker run --runtime=nvidia -it cudf bash ``` Activate the conda environment `cudf` to use the newly built cuDF and libcudf libraries: ``` @@ -337,6 +325,71 @@ flag. Below is a list of the available arguments and their purpose: | `CYTHON_VERSION` | 0.29 | Not supported | set Cython version | | `PYTHON_VERSION` | 3.7 | 3.8 | set python version | + +### Code Formatting + + +#### Python + +cuDF uses [Black](https://black.readthedocs.io/en/stable/), +[isort](https://readthedocs.org/projects/isort/), and +[flake8](http://flake8.pycqa.org/en/latest/) to ensure a consistent code format +throughout the project. They have been installed during the `cudf_dev` environment creation. + +These tools are used to auto-format the Python code, as well as check the Cython +code in the repository. Additionally, there is a CI check in place to enforce +that committed code follows our standards. You can use the tools to +automatically format your python code by running: + +```bash +isort --atomic python/**/*.py +black python +``` + +and then check the syntax of your Python and Cython code by running: + +```bash +flake8 python +flake8 --config=python/.flake8.cython +``` + +Additionally, many editors have plugins that will apply `isort` and `Black` as +you edit files, as well as use `flake8` to report any style / syntax issues. + +#### C++/CUDA + +cuDF uses [`clang-format`](https://clang.llvm.org/docs/ClangFormat.html) + +In order to format the C++/CUDA files, navigate to the root (`cudf`) directory and run: +``` +python3 ./cpp/scripts/run-clang-format.py -inplace +``` + +Additionally, many editors have plugins or extensions that you can set up to automatically run `clang-format` either manually or on file save. + +#### Pre-commit hooks + +Optionally, you may wish to setup [pre-commit hooks](https://pre-commit.com/) +to automatically run `isort`, `Black`, `flake8` and `clang-format` when you make a git commit. +This can be done by installing `pre-commit` via `conda` or `pip`: + +```bash +conda install -c conda-forge pre_commit +``` + +```bash +pip install pre-commit +``` + +and then running: + +```bash +pre-commit install +``` + +from the root of the cuDF repository. Now `isort`, `Black`, `flake8` and `clang-format` will be +run each time you commit changes. + --- ## Attribution