xal 0.2

Execution abstraction layer for high-level system scripts.

xal is a Python library which provides an high-level API to interact with system resources (files, commands, …) and low-level execution via third-parties (stdlib, Fabric, Salt, …).

The concept is you open a session in system, then you run commands within the session:

• session is specific, it holds the execution context, it knows the low-level implementation.
• commands use a generic API. You could run the same commands in another session.

Tip

“xal” is the acronym of “eXecution Abstraction Layer”.

Example

Let’s initialize a session on local system:

>>> import xal
>>> local_session = xal.LocalSession()
>>> local_session.client.connect()
True

In this session, we can manage files:

>>> path = local_session.path('hello-xal.txt')
>>> path.exists()
False
>>> written = path.open('w').write(u'Hello world!')
>>> path.exists()
True
>>> print path.open().read()
Hello world!
>>> path.unlink()
>>> path.exists()
False

We can also execute sh commands:

>>> result = local_session.sh.run(u"echo 'Goodbye!'")
>>> print result.stdout
Goodbye!
<BLANKLINE>

Now let’s make a function that does the same. It takes the session as input argument:

>>> def hello(session):
...     path = session.path('hello-xal.txt')
...     path.open('w').write(u"Hello world!")
...     print path.open().read()
...     path.unlink()
...     print session.sh.run(u"echo 'Goodbye!'").stdout

Of course, we can run it in local session:

>>> hello(local_session)
Hello world!
Goodbye!
<BLANKLINE>

What’s nice is that we can reuse the same function in another session. Let’s create a remote SSH session using Fabric…

>>> remote_session = xal.FabricSession()
>>> remote_session.client.connect(host='localhost')
True

… then just run the same function with this remote session:

>>> hello(remote_session)
Hello world!
Goodbye!
<BLANKLINE>

Motivations

xal ideas are:

• Python users (including sysadmins and devops) have a consistent and unified API to write scripts that perform operations on system.
• such scripts are portable, i.e. they can be executed in various environments. Whatever the operating system, whatever the protocol to connect to and communicate with the system…
• Python community can share libraries that are compatible with tools such as Fabric, zc.buildout, Salt, Ansible…
• it is easier to switch from one tool to another: reconfigure the session, don’t change the scripts. Develop scripts locally, test them remotely via Fabric, distribute them using Salt… or vice-versa.

Project’s status

Today: xal is a proof-of-concept. It focuses on sample implementation of basic features such as managing files and directories, or executing sh commands. The idea is that, as a Python user, you can give it a try and, if you like it, use it for simple tasks.

Tomorrow, depending on feedback from community, xal may improve or be deprecated. As xal’s author, I would like the following things to happen:

• increased stability and performances for current features.
• more execution contexts (i.e. sessions): Salt, Fabric as sudoer, …
• more resources: users, system packages, …
• better API, preferrably built as PEPs. Just as xal’s proof of concept tries to mimic pathlib, there could be a PEP related to every resource. Sh commands (a.k.a. replacement for subprocess) are an epic example.

As xal’s author, I can’t do it alone. If you’d like to help:

• provide feedback. Do you like xal? What do you dislike in xal? Your feedback matters!
• join the project on Github.

Ressources

• Documentation: https://xal.readthedocs.org
• PyPI: https://pypi.python.org/pypi/xal
• Code repository: https://github.com/benoitbryon/xal
• Bugtracker: https://github.com/benoitbryon/xal/issues
• Continuous integration: https://travis-ci.org/benoitbryon/xal