get-started.rst

Get Started

Installation

Bionic can be installed using pip:

pip install 'bionic[standard]'

The bionic[standard] package includes the core framework as well as the most commonly-used dependencies. There are several other subpackages offering different dependencies, documented below<extra-packages>.

You will probably also want to install Graphviz, which Bionic uses to generate visualizations of its workflow graph. Unfortunately Graphviz is not written in Python and can't be installed by pip. On Mac OS X, you can use Homebrew to install it:

brew install graphviz

If you want your data to be automatically cached to Google Cloud Storage, you'll also need to have the Google Cloud SDK installed, have access to a GCS bucket, and install the bionic[gcp] subpackage.

Finally, installing LibYAML will improve performance for some workloads. LibYAML is also available via Homebrew:

brew install libyaml

Bionic supports Python 3.6 and above.

Extra Packages

The default bionic PyPI package installs only the minimal dependencies for building and running flows. However, many other dependency configurations are available. Most users will want the bionic[standard] package, which supports common integrations like Matplotlib, as well as graph visualization.

The full set of subpackages is as follows:

Subpackage	Installation Command	Enables
dask	pip install 'bionic[dask]'	the @dask decorator
----------	------------------------------------	--------------------------------
dev	pip install 'bionic[dev]'	every feature; testing; building documentation
----------	------------------------------------	--------------------------------
dill	pip install 'bionic[dill]'	the @dillable decorator
----------	------------------------------------	--------------------------------
examples	pip install 'bionic[examples]'	the tutorial example code
----------	------------------------------------	--------------------------------
full	pip install 'bionic[full]'	every non-development feature
----------	------------------------------------	--------------------------------
gcp	pip install 'bionic[gcp]'	caching to GCS
----------	------------------------------------	--------------------------------
geopandas	pip install 'bionic[geopandas]'	the @geodataframe decorator
----------	------------------------------------	--------------------------------
image	pip install 'bionic[image]'	automatic de/serialization of PIL.Image objects
----------	------------------------------------	--------------------------------
matplotlib	pip install 'bionic[matplotlib]'	the @pyplot decorator
----------	------------------------------------	--------------------------------
parallel	pip install 'bionic[parallel]'	parallel execution
----------	------------------------------------	--------------------------------
standard	pip install 'bionic[standard]'	graph visualization; Image handling; @pyplot
----------	------------------------------------	--------------------------------
viz	pip install 'bionic[viz]'	graph visualization

Tutorials

These two worked examples illustrate the basic mechanics of Bionic.

tutorials/hello_world.ipynb tutorials/ml_workflow.ipynb