README.md

permalink
/

tf

local tf = import "github.com/tf-libsonnet/core/main.libsonnet"

tf implements core utility functions for generating Terraform code.

Index

• fn isStringArray(v)
• fn mergeAll(objs)
• fn objItems(obj)
• fn objItemsAll(obj)
• fn withData(type, label, attrs, _meta={})
• fn withLocal(name, value)
• fn withLocalList(locals)
• fn withLocalMap(map)
• fn withModule(name, source, inpuuts, version='null', _meta={})
• fn withOutput(name, value, description='null', sensitive='null')
• fn withOutputList(outputs)
• fn withOutputMap(map)
• fn withProvider(name, attrs, alias='null', src='null', version='null')
• fn withResource(type, label, attrs, _meta={})
• fn withSensitiveOutputMap(map)
• fn withVariable(name, isRequired=true, type='null', description='null', default='null')
• obj meta
  • fn new(count='null', depends_on='null', for_each='null', provider='null', lifecycle='null', connection='null', provisioner='null')
  • fn newForModule(count='null', depends_on='null', for_each='null', providers='null')
  • obj meta.lifecycle
    • fn new(create_before_destroy='null', prevent_destroy='null', ignore_changes='null', replace_triggered_by='null', precondition='null', postcondition='null')
    • obj meta.lifecycle.condition
      • fn new(condition, error_message)
  • obj meta.provisioner
    • fn newFile(destination, source='null', content='null', connection='null', when='null', on_failure='null')
    • fn newLocalExec(command, working_dir='null', interpreter='null', environment='null', when='null', on_failure='null')
    • fn newRemoteExec(inline='null', script='null', scripts='null', connection='null', when='null', on_failure='null')
    • obj meta.provisioner.connection
      • fn new(host, type='null', user='null', password='null', port='null', timeout='null', script_path='null', private_key='null', certificate='null', agent='null', agent_identity='null', host_key='null', target_platform='null', bastion_host='null', bastion_host_key='null', bastion_port='null', bastion_user='null', bastion_password='null', bastion_private_key='null', bastion_certificate='null', https='null', insecure='null', use_ntlm='null', cacert='null')

Fields

fn isStringArray

isStringArray(v)

tf.isStringArray returns true if the given value is an array with all elements as string.

Args:

• v (any): The value being evaluated.

Returns:

• A boolean indicating whether the given arg is a string array.

fn mergeAll

mergeAll(objs)

tf.mergeAll takes a list of objects and merges them together into a single object, processing the objects from left to right. This is useful when you want to create many Terraform resources using list comprehension.

Args:

• objs (list[obj]): List of objects to be merged into one.

Returns:

• A single object that is the result of merging all the input objects together.

fn objItems

objItems(obj)

tf.objItems takes an object and returns a list of objects with two attributes:

• k: The object key being iterated.
• v: The object value being iterated.

This is useful when iterating the keys and values of an object. For example, if you had the object:

{
  one: 1,
  two: 2,
  three: 3,
}

You can create iterate the fields like so:

[i.k + ' = ' + i.v for i in objItems(obj)]

Args:

• obj (obj): The object whose fields and values to iterate.

Returns:

• A list of objects with attributes k and v to denote the object keys and values.

fn objItemsAll

objItemsAll(obj)

tf.objItemsAll is like objItems, but also includes hidden fields.

Args:

• obj (obj): The object whose fields and values to iterate.

Returns:

• A list of objects with attributes k and v to denote the object keys and values.

fn withData

withData(type, label, attrs, _meta={})

tf.withData injects a new Terraform data source block into the root configuration.

Additionally, this inserts a private function into the _ref attribute that generates references to attributes of the data source. For example, if you added a new data source using:

withData('azurerm_virtual_network', 'foo', {})

You can get the reference to the id field of the azurerm_virtual_network using the reference:

$._ref.azurerm_virtual_network.foo.get('id')

NOTE: When chaining and merging multiple calls to withResource, withData, and withModule, you may not be able to use super, self, or $ to get the reference to _ref. Instead, make an explicit binding to the outer object using local.

Args:

• type (string): The data source type to create (e.g., aws_instance, local_file, etc).
• label (string): The label to apply to the instance of the data source.
• attrs (obj): The attributes for the instance of the data source being created.
• _meta (obj): An optional meta-argument object that (see meta.libsonnet). Note that while technically you can set the meta-arguments on the attrs object, it is recommended to use the _meta arg to highlight the meta-arguments.

Returns:

• A mixin object that injects the new data source into the root Terraform configuration.

fn withLocal

withLocal(name, value)

tf.withLocal injects a new Terraform local definition into the root configuration.

Args:

• name (string): The name of the local to define.
• value (any): The value to bind to the local.

Returns:

• A mixin object that injects the new local into the root Terraform configuration.

fn withLocalList

withLocalList(locals)

tf.withLocalList injects the list of local configurations as Terraform local definitions in the root configuration.

Args:

• locals (list[obj]): List of local configurations, where each element describes a local. Each element should have the keys n (for name) and v (for value).

Returns:

• A mixin object that injects all the locals into the Terraform config.

fn withLocalMap

withLocalMap(map)

tf.withLocalMap injects all the key value pairs of the input map as Terraform local definitions in the root configuration.

Args:

• map (map[str, str]): Map of local keys to local values.

Returns:

• A mixin object that injects all the key value pairs as locals.

fn withModule

withModule(name, source, inpuuts, version='null', _meta={})

tf.withModule injects a new module block into the root configuration.

Additionally, this inserts a private function into the _ref attribute that generates references to attributes of the module call. For example, if you added a new module call using:

withModule('foo', 'some-source', {})

You can get the reference to the id output using the reference:

$._ref.module.foo.get('id')

NOTE: When chaining and merging multiple calls to withResource, withData, and withModule, you may not be able to use super, self, or $ to get the reference to _ref. Instead, make an explicit binding to the outer object using local.

Args:

• name (string): The name of the module block.
• source (string): The source for the module block.
• inputs (obj): The input values to pass into the module block.
• version (string): The version of the module source to pull in, if the module source references a registry. When null, the version field is omitted from the resulting module block.
• _meta (obj): An optional meta-argument object that (see meta.libsonnet). Note that while technically you can set the meta-arguments on the inputs object, it is recommended to use the _meta arg to highlight the meta-arguments.

Returns:

• A mixin object that injects the new module block into the root Terraform configuration.

fn withOutput

withOutput(name, value, description='null', sensitive='null')

tf.withOutput injects a new Terraform output block into the root configuration.

Args:

• name (string): The name of the output.
• value (string): The expression to bind to the output name.
• description (string): The description of the output. When null, the description field is omitted from the object.
• sensitive (bool): Whether the output contains sensitive information. When null, the sensitive field is omitted from the object.

Returns:

• A mixin object that injects the new output into the root Terraform configuration.

fn withOutputList

withOutputList(outputs)

withOutputList injects the list of output configurations as Terraform output blocks into the root configuration.

Args:

• outputs (list[obj]): List of output configurations, where each element describes an output block. Each element should have the keys n (for name), v (for value), d (for description), and s (for sensitive).

Returns:

• A mixin object that injects all the outputs as output blocks.

fn withOutputMap

withOutputMap(map)

tf.withOutputMap injects all the key value pairs of the input map as Terraform output blocks into the root configuration.

Args:

• map (map[str, str]): Map of output keys to output values.

Returns:

• A mixin object that injects all the key value pairs as output blocks.

fn withProvider

withProvider(name, attrs, alias='null', src='null', version='null')

tf.withProvider injects a new Terraform provider block into the root configuration

Args:

• name (string): The name of the provider to inject.
• attrs (obj): The attributes to apply to the provider block being injected.
• alias (string): The alias to bind to the provider block. When null, the alias attribute is omitted from the provider attributes.
• src (string): Where to source the provider. If specified, an entry to required_providers will be added specifying the source. If both src and version is null, the required_providers entry is omitted.
• version (string): What version of the provider to use. If specified, an entry to required_providers will be added specifying the version. If both src and version is null, the required_providers entry is omitted.

Returns:

• A mixin object that injects the new provider block into the root Terraform configuration.

fn withResource

withResource(type, label, attrs, _meta={})

tf.withResource injects a new Terraform resource block into the root configuration.

Additionally, this inserts a private function into the _ref attribute that generates references to attributes of the resource. For example, if you added a new resource using:

withResource('null_resource', 'foo', {})

You can get the reference to the id field of the null_resource using the reference:

$._ref.null_resource.foo.get('id')

NOTE: When chaining and merging multiple calls to withResource, withData, and withModule, you may not be able to use super, self, or $ to get the reference to _ref. Instead, make an explicit binding to the outer object using local.

Args:

• type (string): The resource type to create (e.g., aws_instance, null_resource, etc).
• label (string): The label to apply to the instance of the resource.
• attrs (obj): The attributes for the instance of the resource being created.
• _meta (obj): An optional meta-argument object that (see meta.libsonnet). Note that while technically you can set the meta-arguments on the attrs object, it is recommended to use the _meta arg to highlight the meta-arguments.

Returns:

• A mixin object that injects the new resource into the root Terraform configuration.

fn withSensitiveOutputMap

withSensitiveOutputMap(map)

tf.withSensitiveOutputMap injects all the key value pairs of the input map as Terraform output blocks with sensitive set to true into the root configuration.

Args:

• map (map[str, str]): Map of output keys to output values.

Returns:

• A mixin object that injects all the key value pairs as output blocks.

fn withVariable

withVariable(name, isRequired=true, type='null', description='null', default='null')

tf.withVariable injects a new Terraform variable block into the root configuration.

Args:

• name (string): The name of the variable.
• isRequired (bool): Whether the variable is required. When true, the default value is omitted from the object.
• type (string): The type of the variable. When null, the type field is omitted from the object.
• description (string): The description of the variable. When null, the description field is omitted from the object.
• default (any): The default value of the variable. Omitted when isRequired is true.

Returns:

• A mixin object that injects the new variable into the root Terraform configuration.

obj meta

fn meta.new

new(count='null', depends_on='null', for_each='null', provider='null', lifecycle='null', connection='null', provisioner='null')

tf.meta.new will generate an object that can be mixed into any resource or data source to set the Terraform meta arguments. This can be passed in as the _meta parameter for any call to [tf.withResource](#fn-withresource] or tf.withData.

Note that this is for resources and data sources. If you wish to generate meta-arguments for a module block, use the tf.meta.newForModule function.

Refer to the meta-arguments tab on the Terraform language reference for more information.

Args:

• count (number): Set the count meta-argument on the block. When null, the count attribute is omitted.
• depends_on (list[str] or string): Set the depends_on meta-argument on the block. When null, the depends_on attribute is omitted. This can be a string if using interpolation in the resulting Terraform.
• for_each (map[str, any], list[any] or string): Set the for_each meta-argument on the block. When null, the for_each attribute is omitted. This can be a string if using interpolation in the resulting Terraform.
• provider (string): Set the provider meta-argument on the block. When null, the provider attribute is omitted.
• lifecycle (obj): Set the lifecycle meta-argument block on the resulting block. When null, the lifecycle block is omitted. It is recommended to generate this using tf.meta.lifecycle.new.
• connection (obj): Set the connection meta-argument blocks on the resulting block. When null, there will be no provisioner blocks added. It is recommended to generate this using in tf.meta.provisioner.connection.new.
• provisioner (list[obj]): Set the list of provisioner meta-argument blocks on the resulting block. When null, there will be no provisioner blocks added. It is recommended to generate this using functions in tf.meta.provisioner.

Returns:

• A mixin that can be merged with a resource or data source object to set meta-arguments.

fn meta.newForModule

newForModule(count='null', depends_on='null', for_each='null', providers='null')

tf.meta.newForModule will generate an object that can be mixed into any module block to set the Terraform meta arguments. This can be passed in as the _meta parameter for any call to [tf.withModule](#fn-withmodule].

Note that this is for module calls. If you wish to generate meta-arguments for a resource or data source block, use the tf.meta.new function.

Refer to the meta-arguments tab on the Terraform language reference for more information.

Args:

• count (number): Set the count meta-argument on the block. When null, the count attribute is omitted.
• depends_on (list[str] or string): Set the depends_on meta-argument on the block. When null, the depends_on attribute is omitted. This can be a string if using interpolation in the resulting Terraform.
• for_each (map[str, any], list[any] or string): Set the for_each meta-argument on the block. When null, the for_each attribute is omitted. This can be a string if using interpolation in the resulting Terraform.
• providers (map[str, str]): Set the providers meta-argument on the block. When null, the providers attribute is omitted.

Returns:

• A mixin that can be merged with a resource or data source object to set meta-arguments.

obj meta.lifecycle

fn meta.lifecycle.new

new(create_before_destroy='null', prevent_destroy='null', ignore_changes='null', replace_triggered_by='null', precondition='null', postcondition='null')

tf.meta.lifecycle.new will generate a new lifecycle block.

Note that unlike the other functions, this includes type checking due to the Terraform requirement that the lifecycle block only supports literal values only. As such, it is easier to do a type check on the args since there is no possibility to use complex Terraform expressions (which will reduce to a string type in jsonnet).

Args:

• create_before_destroy (bool): Set create_before_destroy on the block. When null, the create_before_destroy attribute is omitted.
• prevent_destroy (bool): Set prevent_destroy on the block. When null, the prevent_destroy attribute is omitted.
• ignore_changes (list[str]): Set ignore_changes on the block. When null, the ignore_changes attribute is omitted.
• replace_triggered_by (list[str]): Set replace_triggered_by on the block. When null, the replace_triggered_by attribute is omitted.
• precondition (list[object]): Set precondition subblocks on the block. When null, no precondition subblocks will be rendered. It is recommended to construct this using tf.meta.lifecycle.condition.new.
• postcondition (list[object]): Set postcondition subblocks on the block. When null, no postcondition subblocks will be rendered. It is recommended to construct this using tf.meta.lifecycle.condition.new.

Returns:

• A mixin that can be merged with a meta-argument block to set the lifecycle block.

obj meta.lifecycle.condition

fn meta.lifecycle.condition.new

new(condition, error_message)

tf.meta.lifecycle.condition.new will generate a new condition block that can be used as part of precondition or postcondition in the lifecycle block.

Args:

• condition (string): Set the condition attribute on the block. This should be a Terraform expression escaped with ${}.
• error_message (string): Set the error_message attribute on the block.

Returns:

• An object that can be used as a precondition or postcondition subblock for a lifecycle block.

obj meta.provisioner

fn meta.provisioner.newFile

newFile(destination, source='null', content='null', connection='null', when='null', on_failure='null')

tf.meta.provisioner.newFile will generate a new file provisioner block that can be used as part of the provisioner meta-argument block.

Args:

• source (str): The source file or directory to copy. Exactly one of source or content must be provided.
• content (str): The direct content to copy to the destination. Exactly one of source or content must be provied.
• destination (str): The destination path to write on the remote system.
• connection (obj): Set the connection meta-argument blocks on the resulting block. When null, there will be no provisioner blocks added. It is recommended to generate this using in tf.meta.provisioner.connection.new.
• when (str): Specifies when Terraform should run the command. Can only be set to "destroy". When null, the attribute is omitted from the resulting block.
• on_failure (str): Modify how Terraform handles errors in the underlying command. Must be one of "continue" or "fail". When null, the attribute is omitted from the resulting block.

Returns:

• An object that can be used as a provisioner meta-argument block in the provisioner block list of tf.meta.new.

fn meta.provisioner.newLocalExec

newLocalExec(command, working_dir='null', interpreter='null', environment='null', when='null', on_failure='null')

tf.meta.provisioner.newLocalExec will generate a new local-exec provisioner block that can be used as part of the provisioner meta-argument block.

Args:

• command (str): The command to execute on the operator machine as part of the resource lifecycle.
• working_dir (str): The working directory where command will be executed. When null, the attribute is omitted from the resulting block.
• interpreter (list[str]): The list of interpreter arguments used to execute the command. When null, the attribute is omitted from the resulting block.
• environment (map[str, str]): Map of key-value pairs representing the environment variables that should be set. When null, the attribute is omitted from the resulting block.
• when (str): Specifies when Terraform should run the command. Can only be set to "destroy". When null, the attribute is omitted from the resulting block.
• on_failure (str): Modify how Terraform handles errors in the underlying command. Must be one of "continue" or "fail". When null, the attribute is omitted from the resulting block.

Returns:

• An object that can be used as a provisioner meta-argument block in the provisioner block list of tf.meta.new.

fn meta.provisioner.newRemoteExec

newRemoteExec(inline='null', script='null', scripts='null', connection='null', when='null', on_failure='null')

tf.meta.provisioner.newRemoteExec will generate a new remote-exec provisioner block that can be used as part of the provisioner meta-argument block.

Args:

• inline (list[str]): The list of commands to execute on the remote machine as part of the resource lifecycle. Exactly one of inline, script, or scripts must be provied.
• script (str): The path to a local script that will be copied to the remote machine and then executed. Exactly one of inline, script, or scripts must be provied.
• scripts (list[str]): The list of paths to local scripts that will be copied to the remote machine and then executed. Exactly one of inline, script, or scripts must be provied.
• connection (obj): Set the connection meta-argument blocks on the resulting block. When null, there will be no provisioner blocks added. It is recommended to generate this using in tf.meta.provisioner.connection.new.
• when (str): Specifies when Terraform should run the command. Can only be set to "destroy". When null, the attribute is omitted from the resulting block.
• on_failure (str): Modify how Terraform handles errors in the underlying command. Must be one of "continue" or "fail". When null, the attribute is omitted from the resulting block.

Returns:

• An object that can be used as a provisioner meta-argument block in the provisioner block list of tf.meta.new.

obj meta.provisioner.connection

fn meta.provisioner.connection.new

new(host, type='null', user='null', password='null', port='null', timeout='null', script_path='null', private_key='null', certificate='null', agent='null', agent_identity='null', host_key='null', target_platform='null', bastion_host='null', bastion_host_key='null', bastion_port='null', bastion_user='null', bastion_password='null', bastion_private_key='null', bastion_certificate='null', https='null', insecure='null', use_ntlm='null', cacert='null')

tf.meta.provisioner.connection.new will generate a new connection block that can be used as part of the provisioner meta-argument block.

Args:

• host (str): The address of the remote resource to connect to.
• type (str): The connection type to use when connecting to the instance. Must be one of "ssh" or "winrm". When null, the attribute is omitted from the resulting block and defaults to what Terraform has set internally ("ssh").
• user (str): The user to use for the connection. When null, the attribute is omitted from the resulting block and defaults to what Terraform has set internally.
• password (str): The password to use for the connection. When null, the attribute is omitted from the resulting block.
• port (number): The port to use for the connection. When null, the attribute is omitted from the resulting block and defaults to what Terraform has set internally.
• timeout (str): The timeout to wait for the connection. When null, the attribute is omitted from the resulting block and defaults to what Terraform has set internally ("5m").
• script_path (str): The path used to copy scripts meant for remote execution. When null, the attribute is omitted from the resulting block.
• private_key (str): The contents of an SSH key to use for the connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• certificate (str): The contents of a signed CA certificate to be used in conjunction with the private_key arg. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• agent (bool): Whether to use the ssh-agent for authenticating. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• agent_identity (str): The preferred identity from the ssh agent to use for authentication. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• host_key (str): The public key from the remote host or the signing CA. This is used to verify the connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• target_platform (str): The target platform to connect to. Must be one of "unix" or "windows". When null, the attribute is omitted from the resulting block and defaults to what Terraform has set internally ("unix"). Can only be set when type is "ssh" or null.
• bastion_host (str): The address of a bastion host to hop the connection through. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_host_key (str): The public key from the bastion host or the signing CA. This is used to verify the connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_port (number): The port to use for the bastion connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_user (str): The user to use for the bastion connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_password (str): The password to use for the bastion connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_private_key (str): The contents of an SSH key file to use for the bastion connection. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• bastion_certificate (str): The contents of a signed CA certificate to be used in conjunction with the bastion_private_key arg. When null, the attribute is omitted from the resulting block. Can only be set when type is "ssh" or null.
• https (bool): Whether to connect using HTTPS as opposed to HTTP. When null, the attribute is omitted from the resulting block. Can only be set when type is "winrm".
• insecure (bool): Whether to skip validation of the HTTPS certificate chain. When null, the attribute is omitted from the resulting block. Can only be set when type is "winrm".
• use_ntlm (bool): Whether to use NTLM authentication. When null, the attribute is omitted from the resulting block. Can only be set when type is "winrm".
• cacert (str): The CA certificate to validate against. When null, the attribute is omitted from the resulting block. Can only be set when type is "winrm".

Returns:

• An object that can be used as a connection sub block on any remote-exec or file provisioner meta-argument block.