Kebab::Parseable

Mix into a struct to parse it from command-line arguments.

Annotate the struct with Kebab::Command, its fields with Kebab::Option, Kebab::Argument, or Kebab::Subcommand, then call .parse or .run.

@[Kebab::Command(summary: "Greet someone")]
struct Greet
  include Kebab::Parseable

  @[Kebab::Argument(description: "Name to greet")]
  getter name : String

  @[Kebab::Option(short: 'l', description: "Make it loud")]
  getter? loud : Bool = false
end

case result = Greet.parse(["-l", "Ada"])
in Greet         then puts result.loud? ? "HELLO #{result.name}" : "Hello #{result.name}"
in Kebab::Help   then puts result        # the user passed --help
in Kebab::Errors then STDERR.puts result # the input was invalid
end

Parsing and dispatch

.parse never raises. It returns the parsed struct, a Kebab::Help when the user asked for help, or a Kebab::Errors when the input was invalid, so Crystal's case ... in makes you handle all three.

For a command that owns its behaviour, define def run and call .run, which parses, invokes run on success, and writes help and errors for you.

.schema returns the command tree as a Kebab::Schema::Command for help, completion, and your own tooling.