CONTRIBUTING.md

Contribution guide

▶️ Requirements Python3.6 • AWS credentials • Git • DVC

git clone [email protected]:ai/bayesian-optimization.git
cd bayesian-optimization
pip install -r requirements.txt

📖 Build the documentation

make -C doc html

Output files are located in _build/html. Continuous integration will automatically deployed a static version at this adress

🤖 Test your code

# Static analysis
invoke codestyle
invoke lint

# Unit tests
invoke tests core
invoke tests statespace
invoke tests ...

# Integration tests

# Data tests
invoke tests data --dvc

Before committing, you may ensure that all test status are green. In any case, commiting will trigger an automatic test suite execution. Merging into master and deploying releases should be done after validation. Adding or removing features from the codebase should be accompanied with an appropriate test suite update.

🚚 Data sources

DVC is the way to go to add large files to the repository without burden it. Data sources are copied to a S3 cloud storage solution whereas only file hashes will be commited into dvc files. The commands are similar to Git, as shown in the documentation. The pull and push commands need valid AWS credentials to connect to S3 to be working. Ask @m.janvier for yours.

Add a new data source

dvc add path/to/file.extension
# The command generates the .dvc file
git add path/to/file.extension.dvc
git commit -m 'add data file'
dvc push

Download data sources

git checkout <commit>
dvc pull

The downloaded data sources will be the ones versionned at the commit you're currently at.