tf

Less verbose and more shell-friendly Terraform.

Better this:	Than this:

The original Terraform is abusively oververbosed. The tf wrapper makes Terraform more CLI-friendly:

• removes unnecessary output
• adds own colored progress indicator
• provides an extra compact mode
• allows to use of multiple arguments when original command doesn't support it
• adds automatically quotes for indexed resources (xargs friendly)
• generates a log file when TF_OUTPUT_PATH environment variable is used

Install

Download from releases. Put anywhere in the $PATH.

asdf

Install with asdf:

asdf plugin-add tf https://github.com/dex4er/asdf-tf.git
asdf install tf latest
asdf global tf latest

mise

Install with mise-en-place:

mise plugins install tf https://github.com/dex4er/asdf-tf.git
mise use -g tf@latest

Homebrew

Install with homebrew:

brew tap dex4er/tap
brew install tf

Zsh

With Zsh noglob should be used to prevent errors when indexed parameters are arguments for tf command.

Add to ~/.zshenv:

alias tf='noglob tf'

You can use re-use Oh My Zsh plugin dedicated to the original Terraform:

compdef _terraform tf

Usage

tf init
tf plan
tf apply
tf list
tf show

etc...

You can combine commands, ie.:

tf apply $(tf list | grep aws_vpc)
tf list | grep data.aws_region.current | xargs tf refresh
tf list | grep random_password | xargs tf rm
tf list | grep aws_subnet | xargs tf show
tf list | grep module.one | while read r; do echo tf mv $r ${r/module.one./module.two.}; done | bash -x
tf list | fzf | xargs tf show | highlight -O xterm256 --syntax terraform

It is recommended to use $() rather than xargs for tf apply or tf destroy because these commands are interactive.

For not recognized commands tf passes all arguments to terraform command.

terraform command might be replaced with another using TERRAFORM_PATH environment variable, ie.:

export TERRAFORM_PATH=tofu

tf apply

The same as terraform apply with less verbose output.

Instead of Reading/Creating/Destroying... messages it will show a short progress indicator.

It will skip (known after apply) lines from the -short mode output. Also it will hide a plan for data sources (data.xxx will be read during apply blocks) from the -short and -compact mode output.

Additional options can be used: -counters shows counters with processed resources, -compact skips the content of the resources, -short removes unecessary lines (default), -full keeps original manifest, -fan hides messages about progress and shows short indicator (default), -dots hides messages about progress and shows single dot or character for each line, -verbatim keeps original messages about progress, -verbose adds counters to original messages (default if TF_IN_AUTOMATION=1), -no-outputs hides outputs (default, -no-outputs=false shows it again).

The command accepts the resource name as an argument without -target= option. If the argument misses quotes inside square brackets then they will be added.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

The variable TF_PLAN_FORMAT switches the default format for the plan (compact, short, full). The variable TF_PROGRESS_FORMAT switches the default format for the progress indicator (counters, fan, dots, verbose, verbatim).

tf destroy

The same as terraform destroy with less verbose output.

Instead of Reading/Creating/Destroying... messages it will show a short progress indicator.

It will skip (known after apply) lines from the -short mode output. Also it will hide a plan for data sources (data.xxx will be read during apply blocks) from the -short and -compact mode output.

Additional options can be used: -counters shows counters with processed resources, -compact skips the content of the resources, -shortremoves unnecessary lines (default),-fullkeeps original manifest,-fanhides messages about progress and shows short indicator (default),-dotshides messages about progress and shows single dot or character for each line,-verbatimkeeps original messages about progress,-verboseadds counters to original messages (default ifTF_IN_AUTOMATION=1), -no-outputshides outputs (default,-no-outputs=false` shows it again).

The command accepts the resource name as an argument without -target= option. If the argument misses quotes inside square brackets then they will be added.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

The variable TF_PLAN_FORMAT switches the default format for the plan (compact, short, full). The variable TF_PROGRESS_FORMAT switches the default format for the progress indicator (counters, fan, dots, verbose, verbatim).

tf import

The same as terraform import with better handling of arguments.

If the first argument misses quotes inside square brackets then they will be added.

The command appends the third argument and more to the second with a space as the separator.

The tf show for this resource is run after import unless additional option -no-show has been used.

tf init

The same as terraform init with less verbose output.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

Additional option -codesign causes that all providers will have replaced existing signature and checksums will be removed from the lock file. It will allow to run providers by Terraform on MacOS when the system blocks applications with unrecognized signature.

tf list

The same as terraform state list with less verbose output and ANSI stripped.

If the argument misses quotes inside square brackets then they will be added.

tf mv

The same as terraform state mv with less verbose output.

If the argument misses quotes inside square brackets then they will be added.

tf plan

The same as terraform plan with less verbose output.

Instead of Reading... messages it will show a short progress indicator.

It will skip (known after apply) lines from the -short mode output. Also, it will hide a plan for data sources (data.xxx will be read during apply blocks) from the -short and -compact mode output.

Additional options can be used: -counters shows counters with processed resources, -compact skips the content of the resources, -short removes unnecessary lines (default), -full keeps original manifest, -fan hides messages about progress and shows short indicator (default), -dots hides messages about progress and shows single dot or character for each line, -verbatim keeps original messages about progress, -verbose adds counters to original messages (default if TF_IN_AUTOMATION=1).

The command accepts the resource name as an argument without -target= option. If the argument misses quotes inside square brackets then they will be added.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

The variable TF_PLAN_FORMAT switches the default format for the plan (compact, short, full). The variable TF_PROGRESS_FORMAT switches the default format for the progress indicator (counters, fan, dots, verbose, verbatim).

tf refresh

The same as terraform refresh with less verbose output.

Additional options can be used: -counters shows counters with processed resources, -compact skips the content of the resources, -short removes unnecessary lines (default), -full keeps original manifest, -fan hides messages about progress and shows short indicator (default), -dots hides messages about progress and shows single dot or character for each line, -verbatim keeps original messages about progress, -verbose adds counters to original messages (default if TF_IN_AUTOMATION=1), -no-outputs hides outputs (default, -no-outputs=false shows it again).

The command accepts resource name as an argument without -target= option. If the argument misses quotes inside square brackets then they will be added.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

The variable TF_PLAN_FORMAT switches the default format for the plan (compact, short, full). The variable TF_PROGRESS_FORMAT switches the default format for the progress indicator (counters, fan, dots, verbose, verbatim).

tf rm

The same as terraform state rm with less verbose output.

The command accepts multiple arguments. If the argument misses quotes inside square brackets then they will be added.

tf show

The same as terraform show and terraform state show with less verbose output and ANSI stripped.

terraform show is used when the command is run without arguments and terraform state show when arguments are used.

The command accepts multiple arguments. If the argument misses quotes inside square brackets then they will be added.

An additional option can be used: -no-outputs hides outputs (default, -no-outputs=false shows it again).

tf taint

The same as terraform taint and it accepts multiple arguments. If the argument misses quotes inside square brackets then they will be added.

tf untaint

The same as terraform untaint and it accepts multiple arguments. If the argument misses quotes inside square brackets then they will be added.

tf upgrade

The same as terraform init -upgrade with less verbose output.

The command will log to the file named in TF_OUTPUT_PATH environment variable. The file name is resolved with strftime's % sequences.

Additional option -codesign causes that all providers will have replaced existing signature and checksums will be removed from the lock file. It will allow to run providers by Terraform on MacOS when the system blocks applications with unrecognized signature.

License

Copyright (c) 2020-2024 Piotr Roszatycki piotr.roszatycki@gmail.com

MIT