Once you send a pull request, it is automatically tested on CircleCI. By setting up the CircleCI Local CLI, you can check your code in your local environment before submitting the pull request.
The following circleci job runs code checking:
$ circleci build --job checks
The above job contains following checkers:
If any warnings or errors are emitted, please fix them.
Note that we use comment-style type annotation for compatibility with Python 2.
Please see also our Coding Style Conventions.
When adding a new feature or fixing a bug, you also need to write sufficient test code. We use pytest as the testing framework and unit tests are stored under the tests directory.
You can run your tests as follows:
// Run all the unit tests.
$ pytest
// Run all the unit tests defined in the specified test file.
$ pytest tests/${TARGET_TEST_FILE_NAME}
The following circleci job runs all unit tests in Python 3.7:
// Note that this job will download several hundred megabytes of data to
// install all the packages required for testing,
// and take several tens of minutes to complete all tests.
$ circleci build --job tests-python37
You can run tests and examples for each Python version using the following jobs:
tests-python27
tests-python35
tests-python36
tests-python37
examples-python27
examples-python35
examples-python36
examples-python37
In addition, to check the documents, run:
$ circleci build --job document
When adding a new feature to the framework, you also need to document it in the reference. The documentation source is stored under docs directory and written in reStructuredText format.
To build the documentation, you need to install Sphinx:
$ pip install sphinx sphinx_rtd_theme
Then you can build the documentation in HTML format locally:
$ cd docs
$ make html
HTML files are generated under build/html
directory. Open index.html
with the browser and see
if it is rendered as expected.
Note that docstrings (documentation comments in the source code) are collected from the installed
Optuna module. If you modified docstrings, make sure to install the module (e.g.,
using pip install -e .
) before building the documentation.