A cross-platform task runner for executing commands and generating files from templates

Orbit started with the need to find a cross-platform alternative of make and sed -i commands. As it does not aim to be as powerful as these two commands, Orbit offers an elegant solution for running tasks and generating files from templates, whatever the platform you're using.

• Install
• Generating a file from a template
• Defining and running tasks

Download the latest release of Orbit from the releases page. You can get Orbit for a large range of OS and architecture.

The file you downloaded is a compressed archive. You'll need to extract the Orbit binary and move it into a folder where you can execute it easily.

Linux/MacOS:

tar -xzf orbit*.tar.gz orbit
sudo mv ./orbit /usr/local/bin && chmod +x /usr/local/bin/orbit

Windows:

Right click on the file and choose Extract All.

Move the binary to a folder like C:\Orbit. Then, add it in your Path system environment variables. Click System, Advanced system settings, Environment Variables... and open Path under System variables. Edit the Variable value by adding the folder with the Orbit binary.

Alright, you're almost done 🤘! Let's check your installation by running:

orbit version

Orbit uses the Go package text/template under the hood as a template engine. It provides a interesting amount of logic for your templates.

The Go documentation and the Hugo documentation cover a lot of features that aren't mentioned here. Don't hesitate to take a look at these links to understand the Go template engine! 😃

Also, Orbit provides Sprig library and three custom functions:

• os which returns the current OS name at runtime (you may find all available names in the official documentation).
• verbose which returns true if logging is set to info level.
• debug which returns true if logging is set to debug level.

orbit generate [flags]

Specify the path of the template. This flag is required.

Specify the output file which will be generated from the template.

Good to know: if no output is specified, Orbit will print the result to Stdout.

The flag -p allows you to specify many data sources which will be applied to your template:

orbit generate [...] -p "key_1,file_1.yml"
orbit generate [...] -p "key_1,file_1.yml;key_2,file_2.toml;key_3,file_3.json;key_4,.env;key_5,some raw data"

As you can see, Orbit handles 5 types of data sources:

• YAML files (*.yaml, *.yml)
• TOML files (*.toml)
• JSON files (*.json)
• .env files
• raw data

The data will be accessible in your template through {{ .Orbit.my_key.my_data }}.

If you don't want to specify the payload each time your running orbit generate, you may also create a file named orbit-payload.yml in the folder where your running your command:

payload:

    - key: my_key
      value: my_file.yml
      
    - key: my_other_key
      value: Some raw data

By doing so, running orbit generate [...] will be equivalent to running orbit generate [...] -p "my_key,my_file.yml;my_other_key,Some raw data".

Note: you are able to override a data source from the file orbit-payload.yml if you set the same key in the -p flag.

The flag -t allows you to specify additional templates which are used in your template:

orbit generate [...] -t "template_1.txt"
orbit generate [...] -t "template_1.txt,template_2.yml"
orbit generate [...] -t "template_1.txt,template_2.yml,../../template_3.toml"

So, in order to generate a file from this template:

{{ template "additional_template.txt" }}

You should run:

orbit generate [...] -t "path/to/additional_template.txt"

If you don't want to specify the templates each time your running orbit generate, you may also use the file orbit-payload.yml in the folder where your running your command:

payload:

[...]

templates:

  - template_1.txt
  - template_2.yml

By doing so, running orbit generate [...] will be equivalent to running orbit generate [...] -t "template_1.txt,template_2.yml".

The flag --delimiters allows you to specify an alternative template delimiter pair (left/open, right/close). If unspecified, the Go defaults ({{ and }}) will be used. Note that, if this option is used, exactly two delimiters must be specified.

Examples:

orbit generate [...] --delimiters "<<,>>"
orbit generate [...] --delimiters "<<" --delimiters ">>"

The first delimiter (<< in the examples above) is used for the left/opening delimiter while the second delimiter (>> in the examples above) is used for the right/closing delimiter. This applies regardless of whether the delimiters are specified as a comma-separated pair (first example) or by repeated use of the option (second example).

Sets logging to info level.

Sets logging to debug level.

Let's create our simple template template.yml:

companies:

{{- range $company := .Orbit.Values.companies }}
  - name: {{ $company.name }}
    launchers:
  {{- range $launcher := $company.launchers }}
    - {{ $launcher }}
  {{ end }}
{{- end }}

And the data provided a YAML file named data-source.yml:

companies:

  - name: SpaceX
    launchers:
      - Falcon 9
      - Falcon Heavy
      
  - name: Blue Origin
    launchers:
      - New Shepard
      - New Glenn

agencies:

  - name: ESA
    launchers:
      - Ariane 5
      - Vega

The command for generating a file from this template is quite simple:

orbit generate -f template.yml -p "Values,data-source.yml" -o companies.yml

This command will create the companies.yml file with this content:

companies:

  - name: SpaceX
    launchers:
      - Falcon 9
      - Falcon Heavy
      
  - name: Blue Origin
    launchers:
      - New Shepard
      - New Glenn

orbit run [tasks] [flags]

Like the make command with its Makefile, Orbit requires a configuration file (YAML, by default orbit.yml) where you define your tasks:

tasks:

  - use: my_first_task
    short: My first task short description
    run:
      - command [args]
      - command [args]
      - ...
      
  - use: my_second_task
    private: true
    run:
      - command [args]
      - command [args]
      - ...

• the use attribute is the name of your task.
• the short attribute is optional and is displayed when running orbit run.
• the private attribute is optional and hides the considered task when running orbit run.
• the run attribute is the stack of commands to run.
• a command is a binary which is available in your $PATH.

Once you've created your orbit.yml file, you're able to run your tasks with:

orbit run my_first_task
orbit run my_second_task
orbit run my_first_task my_second_task

Notice that you may run nested tasks 🤘!

Also a cool feature of Orbit is its ability to read its configuration through a template.

For example, if you need to execute a platform specific script, you may write:

tasks:

  - use: script
    run:
    {{ if ne "windows" os }}
      - my_script.sh
    {{ else }}
      - .\my_script.bat
    {{ end }}

Orbit will automatically detect the shell you're using (with the SHELL environment variable on POSIX system and COMSPEC on Windows).

Running the task script from the previous example will in fact executes cmd.exe /c .\my_script.bat on Windows or /bin/sh -c my_script.sh (or /bin/zsh -c my_script.sh etc.) on others OS.

Of course, if you want to specify the binary which is calling your commands, there is a shell attribute available:

tasks:

  - use: script
    shell: /bin/bash -c
    run:
      - command [args]
      - ...

Last but not least, a task is able to call others tasks within the same context thanks to the run function:

tasks:

  - use: task
    run:
      - {{ run "subtask_1" "subtask_2" }}

  - use: subtask_1
    run:
      - command [args]
      - ...
     
  - use: subtask_2
    run:
      - command [args]
      - ...

The flag -p allows you to specify many data sources which will be applied to your configuration file.

It works the same as the -p flag from the generate command.

Of course, you may also create a file named orbit-payload.yml in the same folder where you're executing Orbit.

The flag -t allows you to specify additional templates which are used in your configuration file.

It works the same as the -t flag from the generate command.

Of course, you may also create a file named orbit-payload.yml in the same folder where you're executing Orbit.

Sets logging to info level.

Sets logging to debug level.

Let's create our simple configuration file orbit.yml:

tasks:

  - use: prepare
    run:
     - orbit generate -f configuration.template.yml -o configuration.yml -p "Data,config.json"
     - echo "configuration.yml has been succesfully created!"

You are now able to run the task prepare with:

orbit run prepare

This task will:

• create a file named configuration.yml
• print configuration.yml has been succesfully created! to Stdout

Voilà! 😃

Would you like to update this documentation ? Feel free to open an issue.