Thank you for your interest in contributing to Ecole! 🌟 Contributions are more diverse than contributing new features. Improving the documentation, reporting and reproducing bugs, discussing the direction of Ecole in the discussions, helping others use Ecole.
Not all code contributions are relevant for Ecole. It does not mean that the idea is not good. We try to balance added value with maintenance and complexity of a growing codebase. For that reason, it is a good idea to communicate with the developpers first to be sure that we agree on what should be added/modified.
Be sure to open an issue before sending a pull request.
Tour of the codebase¶
libecoleis the Ecole C++ library. Ecole is mostly written in C++ so this is where you will find most features, rewards, observations…
pythoncontains some Python and C++ code to create bindings to Python. Ecole uses PyBind to create the binding, these are all the C++ files in this directory. Sometimes, you may find Python code as well. This is either because a feature is more naturally implemented in Python, or because we have accepted an early contribution that is not yet ported to C++.
docsis the Sphinx documentation written in reStructuredText.
examplesare practical examples showcasing how to use Ecole for certain tasks.
Dependencies with Conda¶
All dependencies required for building Ecole (including SCIP) can be resolved using a
Install everything in a development environment (named
conda env create -n ecole -f dev/conda.yaml
conda activate ecole conda config --append channels conda-forge conda config --set channel_priority flexible
This environment contains tools to build ecole and scip, format code, test, generate documentation etc. These are more than the dependencies to only use Ecole.
CMake is a meta-build tool, used for configuring other build tools
(e.g. Make) or IDE’s.
The whole build of Ecole can be done with CMake.
A one-time configuration is necessary for CMake to find dependencies, detect system
CMake is made available in the
ecole environment created earlier.
For the following, this environment always needs to be activated.
In the Ecole source repository, configure using
cmake -B build/ -D ECOLE_DEVELOPER=ON
This is the time to pass optional build options, such as the build type and compiler
choice. For instance
-D CMAKE_BUILD_TYPE=Debug can be added to compile with debug
-D ECOLE_DEVELOPER=ON changes the default settings (such as the build
type, static analysis, etc.) for added convenience.
Only the default settings are changed, this mode does not override any explicit setting.
Then, build Ecole with
cmake --build build/ [--parallel [count]]
The Python package can finally be installed from the build directory
python -m pip install -I build/python
Be sure to eliminate all warnings. They will be considered as errors in the PR.
If you encounter problems with your compiler (because it is too old for instance), you can use the ones from ananconda.
conda install cxx-compiler
And start again the configuration of Ecole.
rm -r build/ && cmake -B build/ -D ECOLE_DEVELOPER=ON
Running the tests¶
The C++ tests are build with Catch2.
It produces a single standalone executable under
For test options, consult
Python tests are build with PyTest. As any Python project, running the tests requires that Python finds the correct Ecole package. It is therefore not recommended to manually install Ecole between every tests, as it leaves room for running tests with an out of date installation. The CMake configuration creates a Python virtual environment and installs the Ecole Python package inside automatically. To run the tests (or another operation with Python Ecole), use the virtual environment
./build/venv/bin/python -m pytest python/tests/
Generating the documentation¶
The documentation is a CMake target that is not built by default.
-D ECOLE DEVELOPER=ON, the documentation can be generated by providing
the target to the CMake build command:
cmake --build build/ --target ecole-sphinx
The generated HTML files are located under
build/doc/sphinx/html/index.html can be opened in your browser to
visualize the documentation.
The quality and conventions of the code are enforced automatically with various tools, for instance to format the layout of the code and fix some C++ error-prone patterns.
Some C++ tools need access to a compilation database.
This is a file called
compile_commands.json that is created automatically by CMake when using
The file needs to be accessible at the root directory of the project, so you should symlink it like
so (assuming you have configured CMake in the
build directory as shown before).
ln -s build/compile_commands.json
This file is also read by clangd, a C++ language server (already installed in the conda environment). To get code completion, compile errors, go-to-definition and more, you can install a language server protocol plugin for your editor.
The tools are configured to run with pre-commit, that is they can be added to run automatically when making a commit, pushing, or on demand. To have the tools run automatically, install the pre-commit hooks using
The tools are configured to run light tests only on the files that were changed during the commit, so they should not run for long. Installing the pre-commit hooks to run the tools is recommended. Similar tests will be run online and pull requests will fail if the tools have not been run.
pre-commit hooks, commits will be rejected by
git if the tests ran by the tools fail.
If the tools can fix the issue for you, you will find some modifications that you can add to
Sometimes when working locally, it can be useful not to run the tools.
You can tell
git to ignore the
pre-commit hooks by passing the
--no-verify to any
git command making commit, including
git commit --no-verify
Other times you may wish to run the tools on all files unconditionally. This can be done using
pre-commit run --all-files
Running the continuous integration locally¶
To run a job locally, install Docker and the
CircleCI command line interface.
To reproduce a job, say
build_py38, on your machine, execute
circleci local execute --job build_py38
When things fail¶
If you cannot eliminate some warnings, code checks, errors, do not hesistate to ask questions in the Github Discussions.
When you cannot figure things out, it’s OK to send a failing pull request. We wish to grow as a community, and help others improve, not exclude and belittle. 🌈