Getting started with atomicapp

atomicapp is a reference implementation of the Nulecule Specification. It can be used to bootstrap container applications and to install and run them. atomicapp is designed to be run in a container context. Examples using this tool may be found in the Nulecule library.

If you want to know the internals of atomicapp, how it works, etc., or contribute to it's development, this post is for you.

Setup

Install python virtualenv utils

virtualenv is a tool to create isolated Python environments. virtualenv creates a folder which contains all the necessary executables to use the packages that a Python project would need. virtualenvwrapper is a set of extensions to easily create, manage and destroy virtualenvs. I personally prefer using virtualenvwrapper for my Python developement work.

sudo dnf install -y python-virtualenvwrapper  

Restart your shell after the above is installed.

Getting the code

git clone https://github.com/projecatomic/atomicapp.git  

Setup for development

cd atomicapp  
mkvirtualenv atomicapp  
python setup.py develop  
echo "alias atomicapp=~/.virtualenvs/atomicapp/bin/atomicapp" >> ~/.virtualenvs/atomicapp/bin/postactivate  
echo "alias sudo='sudo '" >> ~/.virtualenvs/atomicapp/bin/postactivate  
echo "unalias atomicapp && unalias sudo" >> ~/.virtualenvs/atomicapp/bin/postdeactivate  

Understanding the code

atomicapp  
├── cli
│   ├── __init__.py
│   └── main.py
├── constants.py
├── __init__.py
├── nulecule
│   ├── base.py
│   ├── container.py
│   ├── exceptions.py
│   ├── __init__.py
│   ├── lib.py
│   └── main.py
├── plugin.py
├── providers
│   ├── docker.py
│   ├── external
│   ├── __init__.py
│   ├── kubernetes.py
│   ├── marathon.py
│   ├── openshift.py
│   └── README.md
├── requirements.py
└── utils.py

CLI

The entry point for atomicapp CLI is atomicapp/cli/main.py. The CLI args and options are added in atomicapp.cli.main.CLI class. The CLI commands are handled by functions named as cli_<command_name> in atomicapp.cli.main.

Nulecule

atomicapp.nulecule is the crux of atomicapp as it implements the Nulecule specification. The CLI interacts with NuleculeManager in atomicapp.nulecule.main to execute various commands. NuleculeManager provides the entrypoints to various functionalites: unpack, genanswers, fetch, run, stop, clean, to manage the life cycle of a Nulecule application.

atomicapp.nulecule.base implements Nulecule to represent a Nulecule application from Nulecule SPEC file, and NuleculeComponent to represent an item in the graph attribute of the Nulecule application. atomicapp.nulecule.base.NuleculeComponent interacts with the underlying providers using atomicapp.providers in it's deploy() and undeploy() methods.

Providers

atomicapp.providers implements interfaces for atomicapp to interact with the different providers: docker, kubernetes, openshift, etc. A provider class usually has three important methods:

  • init: to do initialization required by the provider
  • run: Run artifacts on provider
  • stop: Stop artifacts on provider

Conclusion

The above should have given you a brief overview of the atomicapp and should be good enough to help you get started with atomicapp development. In case you have any query, please feel free to ping us on #nulecule on Freenode or create an issue at https://github.com/projecatomic/atomicapp/issues/.

Check the next post: Storing IRC logs for Fedora Hubs »

Share on:
comments powered by Disqus