Python binding for Terraform.
$ pip install libterraform
NOTE
- Please install version 0.3.1 or above, which solves the memory leak problem.
- This library does not support multithreading.
TerraformCommand
is used to invoke various Terraform commands.
Now, support all commands (plan
, apply
, destroy
etc.), and return a CommandResult
object. The CommandResult
object has the following properties:
retcode
indicates the command return code. A value of 0 or 2 is normal, otherwise is abnormal.value
represents command output. If json=True
is specified when executing the command, the output will be loadedjson
indicates whether to load the output as json.error
indicates command error output.To get Terraform verison:
>>> from libterraform import TerraformCommand
>>> TerraformCommand().version()
<CommandResult retcode=0 json=True>
>>> _.value
{'terraform_version': '1.8.4', 'platform': 'darwin_arm64', 'provider_selections': {}, 'terraform_outdated': True}
>>> TerraformCommand().version(json=False)
<CommandResult retcode=0 json=False>
>>> _.value
'Terraform v1.8.4\non darwin_arm64\n'
To init
and apply
according to Terraform configuration files in the specified directory:
>>> from libterraform import TerraformCommand
>>> cli = TerraformCommand('your_terraform_configuration_directory')
>>> cli.init()
<CommandResult retcode=0 json=False>
>>> cli.apply()
<CommandResult retcode=0 json=True>
Additionally, run()
can execute arbitrary commands, returning a tuple (retcode, stdout, stderr)
.
>>> TerraformCommand.run('version')
(0, 'Terraform v1.8.4\non darwin_arm64\n', '')
>>> TerraformCommand.run('invalid')
(1, '', 'Terraform has no command named "invalid".\n\nTo see all of Terraform\'s top-level commands, run:\n terraform -help\n\n')
TerraformConfig
is used to parse Terraform config files.
For now, only supply TerraformConfig.load_config_dir
method which reads the .tf and .tf.json files in the given
directory as config files and then combines these files into a single Module. This method returns (mod, diags)
which are both dict, corresponding to
the *Module
and hcl.Diagnostic structures in Terraform
respectively.
>>> from libterraform import TerraformConfig
>>> mod, _ = TerraformConfig.load_config_dir('your_terraform_configuration_directory')
>>> mod['ManagedResources'].keys()
dict_keys(['time_sleep.wait1', 'time_sleep.wait2'])
libterraform | Terraform |
---|---|
0.8.0 | 1.8.4 |
0.7.0 | 1.6.6 |
0.6.0 | 1.5.7 |
0.5.0 | 1.3.0 |
0.4.0 | 1.2.2 |
0.3.1 | 1.1.7 |
If you want to develop this library, should first prepare the following environments:
Then, initialize git submodule:
$ git submodule init
$ git submodule update
pip install
necessary tools:
$ pip install poetry pytest
Now, we can build and test:
$ poetry build -f wheel
$ pytest
Terraform is a great tool for deploying resources. If you need to call the Terraform command in the Python program for deployment, a new process needs to be created to execute the Terraform command on the system. A typical example of this is the python-terraform library. Doing so has the following problems:
This library compiles Terraform as a dynamic link library in advance, and then loads it for calling. So there is no need to install Terraform, nor to start a new process.
In addition, since the Terraform dynamic link library is loaded, this library can further call Terraform's internal capabilities, such as parsing Terraform config files.