Setup a Project

Prerequisites

Make sure you have git and python >= 3.7 installed.

Start a new project

Create a virtual environment :

python -m venv ~/venvs/flyte
source ~/venvs/flyte/bin/activate

Then install Flytekit on it:

pip install flytekit

Then run pyflyte init <project_name>, where <project_name> is the directory that will be created containing the scaffolding for a flyte-ready project. For this guide we’re going call it my_flyte_project:

pyflyte init my_flyte_project
cd my_flyte_project

The my_flyte_project directory comes with a sample workflow, which can be found under flyte/workflows/example.py. The structure below shows the most important files and how a typical Flyte app should be laid out.

A typical Flyte app should have these files

my_flyte_project
├── Dockerfile
├── docker_build_and_tag.sh
├── flyte
│         ├── __init__.py
│         └── workflows
│             ├── __init__.py
│             └── example.py
└── requirements.txt

Note

• You can use pip-compile to build your requirements file.

• The Dockerfile that comes with this is not GPU ready, but is a simple Dockerfile that should work for most of your apps.

Run the Workflow Locally

The workflow can be run locally, simply by running it as a Python script:

python flyte/workflows/example.py

Note

The workflow needs to be invoked after the if __name__ == "__main__" entrypoint at the bottom of flyte/workflows/example.py.

Expected output:

Running my_wf() hello world

Conclusion

We’ve successfully created a Flyte project and executed your workflows on a python runtime environment! 🎉 Next, let’s learn how to deploy to the cloud.