Please use my latest project as a replacement Comonicon

CLI.jl is a library for automatically generating command line interfaces from absolutely Julia object. Is is highly inspired by python-fire. But thanks to Julia's powerful metaprogramming, we are able to do this in a much easier way. CLI.jl make use of Julia's metaprogramming and multiple dispatch feature.

warning It only supports *nix systems at the moment. I have not tested it for Windows. Julia v0.6+ is required. Still under development.

• modular implementation
• compile to binary with PackageCompiler.jl

Pkg.clone(git@github.com:Roger-luo/CLI.jl.git)

You can call @command on julia functions.

#demo.jl

using CLI

@command """
this is a demo CLI
"""


"""
add up
"""
@command function add(x::Int, y::Int; extra::Int=1)
    x + y + extra
end

main()

> demo.jl --help
[SYNOPSIS]
	demo2.jl [-h,--help] [-v,--version] <command>
[DESCRIPTION]
	this is a demo CLI

[COMMANDS]
	add
	  add up

You can create subcommands using Julia's modules. And then you can compile it to a binary application by PackageCompiler.jl

# demo.jl
module demo
using CLI

@command """
this is a demo CLI
"""

module math
using CLI

@command """
math calculations
"""

"""
add up
"""
@command function add(x::Int, y::Int; extra::Int=1)
    x + y + extra
end

"""
multiply two integers
"""
@command mul(x::Int, y::Int) = x * y

end # math

"""
plus two integers
"""
@command function plus(x::Int, y::Int)
    x + y    
end

Base.@ccallable function julia_main(ARGS::Vector{String})::Cint
    main()
end

end

then open a Julia REPL, type

> julia using PackageCompiler; build_executable("demo.jl")

you will get an binary demo, simply copy it to where you want and use it. This will reduce your CLI start time significantly. Enjoy!

CLI.jl offers several types to store your CLI application. You can directly use them to build a CLI application.

using CLI

add(x::Int, y::Int) = x + y
mul(x::Int, y::Int) = x * y

leaf1 = LeafCommand(:add, add, Signature([Int, Int], []))
leaf2 = LeafCommand(:mul, mul, Signature([Int, Int], []))
node = NodeCommand("math", [leaf1, leaf2])

maincmd = MainCommand("cal", [], [node])

stream = ARGStream(ARGS)
resolve(stream, maincmd; debug=false)

CLI.jl accepts most CLI design syntax from GNU CLI Standard, IEEE and other *nix external programs, e.g git, ls, etc.

• support @command for arbitrary Julia objects
• further performance improvements (CLI start time)

Roger Luo