arca 0.2.1

A library for running Python functions (callables) from git repositories in various states of isolation with integrating caching.

Arca

Arca is a library for running Python functions (callables) from git repositories in various states of isolation. Arca can also cache the results of these callables using dogpile.cache.

Getting started

Glossary

• Arca - name of the library. When written as Arca, the main interface class is being referenced.

• Task - definition of the function (callable), consists of a reference to the object and arguments.

• Backend - a way of running tasks.

Installation

Requirements

• Python >= 3.6

Requirements for certain backends:

• Docker (for Docker Backend and Vagrant Backend)

• Vagrant (for the Vagrant Backend)

Installation

To install the last stable version:

python -m pip install arca

If you want to use the Docker backend:

python -m  pip install arca[docker]

Or if you want to use the Vagrant backend:

python -m pip install arca[vagrant]

Or if you wish to install the upstream version:

python -m pip install git+https://github.com/mikicz/arca.git#egg=arca
python -m pip install git+https://github.com/mikicz/arca.git#egg=arca[docker]
python -m pip install git+https://github.com/mikicz/arca.git#egg=arca[vagrant]

Example

To run a Hello World example you’ll only need the arca.Arca and arca.Task classes. Task is used for defining the task that’s supposed to be run in the repositories. Arca takes care of all the settings and provides the basic API for running the tasks.

Let’s say we have the following file, called hello_world.py, in a repository https://example.com/hello_word.git, on branch master.

def say_hello():
   return "Hello World!"

To call the function using Arca, the following example would do so:

from arca import Arca, Task

task = Task("hello_world:say_hello")
arca = Arca()

result = arca.run("https://example.com/hello_word.git", "master", task)
print(result.output)

The code would print Hello World!. result would be a arca.Result instance which currently only has one attribute, output, with the output of the function call. If the task fails, arca.exceptions.BuildError would be raised.

By default, the Current Environment Backend is used to run tasks, which uses the current Python, launching the code in a subprocess. You can learn about backends here.

Further reading

You can read the full documentation on Read The Docs.

Running tests

To run tests you’ll need the optional requirements, Docker and Vagrant. Once you have them and they can be used by the current user you just need to run:

python setup.py test

This will launch the tests and a PEP8 check. The tests will take some time since building the custom docker images is also tested and vagrant, in general, takes a long time to set up.

Contributing

I am developing this library as my bachelor thesis and will be not accepting any PRs at the moment.

Changes

0.2.1 (2018-06-11)

Updated dogpile.cache to 0.6.5, enabling compatability with Python 3.7. Updated the Docker backend to be able to run on Python betas.

0.2.0 (2018-05-09)

This release has multiple backwards incompatible changes against the previous release.

Changes:

• Using extras to install Docker and Vagrant dependencies

  • pip install arca[docker] or pip install arca[vagrant] has to be used

• Automatically using cloned repositories as reference for newly cloned branches

• Using Debian as the default base image in the Docker backend:

  • apk_dependencies changed to apt_dependencies, now installing using apt-get

• Vagrant backend only creates one VM, instead of multiple – see its documentation

• Added timeout to tasks, 5 seconds by default. Can be set using the argument timeout for Task.

• Added timeout to installing requirements, 300 seconds by default. Can be set using the requirements_timeout configuration option for backends.

0.1.1 (2018-04-23)

Updated gitpython to 2.1.9

0.1.0 (2018-04-18)

Initial release

Changes:

• Updated PyPI description and metadata

0.1.0a0 (2018-04-13)

Initial alfa release