Python Makefile
Switch branches/tags
Nothing to show
Clone or download
Latest commit aebb993 Oct 11, 2017

README.md

python-packer

Build Status PyPI PypI

A Python interface for packer.io

Packer version

The interface has been developed vs. Packer v0.7.5.

Installation

You must have Packer installed prior to using this client though as installer class is provided to install packer for you.

 pip install python-packer

 # or, for dev:
 pip install https://github.com/nir0s/python-packer/archive/master.tar.gz

Usage Examples

Packer.build()

import packer

packerfile = 'packer/tests/resources/packerfile.json'
exc = []
only = ['my_first_image', 'my_second_image']
vars = {"variable1": "value1", "variable2": "value2"}
var_file = 'path/to/var/file'
packer_exec_path = '/usr/bin/packer'

p = packer.Packer(packerfile, exc=exc, only=only, vars=vars,
                  var_file=var_file, exec_path=packer_exec_path)
p.build(parallel=True, debug=False, force=False)

Packer.fix()

...

p = packer.Packer(packerfile, ...)
output_file = 'packer/tests/resources/packerfile_fixed.json'
print(p.fix(output_file))

The output_file parameter will write the output of the fix function to a file.

Packer.inspect()

A -machine-readable (mrf) argument is provided.

If the mrf argument is set to True, the output will be parsed and an object containing the parsed output will be exposed as a dictionary containing the components:

...

p = packer.Packer(packerfile, ...)
result = p.inspect(mrf=True)
print(result.parsed_output)
# print(result.stdout) can also be used here

# output:
"variables": [
  {
    "name": "aws_access_key",
    "value": "{{env `AWS_ACCESS_KEY_ID`}}"
  },
  {
    "name": "aws_secret_key",
    "value": "{{env `AWS_ACCESS_KEY`}}"
  }
],
"provisioners": [
  {
    "type": "shell"
  }
],
"builders": [
  {
    "type": "amazon-ebs",
    "name": "amazon"
  }
]

If the mrf argument is set to False, the output will not be parsed but rather returned as is:

...

p = packer.Packer(packerfile, ...)
result = p.inspect(mrf=True)
print(result.stdout)

# output:
Optional variables and their defaults:

  aws_access_key          = {{env `AWS_ACCESS_KEY_ID`}}
  aws_secret_key          = {{env `AWS_ACCESS_KEY`}}

Builders:

  amazon                   (amazon-ebs)

Provisioners:

  shell

...

Packer.push()

You must be logged into Atlas to use the push function:

...

p = packer.Packer(packerfile, ...)
atlas_token = 'oi21mok3mwqtk31om51o2joj213m1oo1i23n1o2'
p.push(create=True, token=atlas_token)

Packer.validate()

...

p = packer.Packer(packerfile, ...)
p.validate(syntax_only=False)

Packer.version()

...

p = packer.Packer(packerfile, ...)
print(p.version())

PackerInstaller.install()

This installs packer to packer_path using the installer_path and verifies that the installation was successful.

packer_path = '/usr/bin/'
installer_path = 'Downloads/packer_0.7.5_linux_amd64.zip'

p = packer.Installer(packer_path, installer_path)
p.install()

Shell Interaction

The sh Python module is used to execute Packer. As such, return values from all functional methods (validate, build, etc..) other than the version method will return an sh execution object. This is meant for you to be able to read stdout, stderr, exit codes and more after executing the commands. With the progression of python-packer less abstract objects will return and more concise return values will be provided.

Additionally, to verify that all errors return with as much info as possible, error handling is done gently. Most errors will raise an sh exception so that you're able to interact with them. Again, as this module progresses, these exceptions will be handled properly.

Testing

Please contribute. Currently tests are not really developed.

git clone git@github.com:nir0s/python-packer.git
cd python-packer
pip install tox
tox

Contributions..

..are always welcome.