After start my saga on Ruby AST Search I built a few tools to help me to understand the AST and to inspect it.

After a few days around I discover everything I needed existed already built in and I just need to learn about it.

Ruby Ripper

Ripper provides a symbolic expression tree for your code.

As it’s in the standard library, you can only start using it.

I love to use ruby -e to test some code inline. Then, let’s start with something to learn about the lexer:


We can start with a simply tokenize. That will split your code into atoms to compile each piece in the sequence.

$ ruby -r ripper -e ‘p Ripper.tokenize(“1 + 2”)’

["1", " ", "+", " ", "2"]


The lexer will tag each tokenized piece and identify each part of the statement:

$ ruby -r ripper -e ‘p Ripper.sexp(“1+2”)’

The output should be something like:

[:program, [[:binary, [:@int, "1", [1, 0]], :+, [:@int, "2", [1, 2]]]]]

Checking one more example:

$ ruby -r ripper -r pp -e ‘pp Ripper.lex(“def a; end”)’

[[[1, 0], :on_kw, "def"],
 [[1, 3], :on_sp, " "],
 [[1, 4], :on_ident, "a"],
 [[1, 5], :on_semicolon, ";"],
 [[1, 6], :on_sp, " "],
 [[1, 7], :on_kw, "end"]]

You can also call the Ripper.sexp to get the symbolic expression of your code:

Let’s use pp to make it more easy to ready:

$ ruby -r ripper -r pp -e ‘pp Ripper.sexp(“1+2 + ( 3 * 4)”) ‘

   [:binary, [:@int, "1", [1, 0]], :+, [:@int, "2", [1, 2]]],
   [:paren, [[:binary, [:@int, "3", [1, 8]], :*, [:@int, "4", [1, 12]]]]]]]]


This feature is cool and strange. I can’t find a proper documentation, and I debug it by myself trying to understand better the strategy. It’s a kind of regular expression mixed with the lexer. Take a look at the example:

Ripper.slice('(1 + 2.0 + 1.1)','int.*float.*float')
=> "1 + 2.0 + 1.1"

It’s useful for grep for specific sequence of node types, for example, lets check for a sequence of int float float in the operation:

Ripper.slice('2.2 / 4.2 + (1 + 2.0+ 1.1) * 2','int.*float.*float')
=> "1 + 2.0+ 1.1"

Take a look at the official example, and it’s cooler than mine.

Ripper.slice('def m(a) nil end', '[ident lparen rparen]+')  #=> "m(a)"

That’s all for today! Thanks for reading o/

Share → Twitter Facebook Linkedin

Hello there, my name is Jônatas Davi Paganini and this is my personal blog.
I'm developer advocate at Timescale and I also have a few open source projects on github.

Check my talks or connect with me via linkedin / twitter / github / instagram / facebook / strava / meetup.