TinyGQL

Very experimental GraphQL parser. It's mostly reusing the lexer from GraphQL-Ruby, but the parser is a hand-written recursive descent parser.

I want to target this at server side applications, so the parser eliminates some nice stuff for humans (namely line / column information, and it throws away comments).

Right now this code:

1. Doesn't know how to execute anything. It just gives you an AST
2. Isn't used anywhere (except in your heart, *but hopefully in production someday!)

You can get an AST like this:

ast = TinyGQL.parse "{ cool }"

The AST is iterable, so you can use the each method:

ast = TinyGQL.parse "{ cool }"
ast.each do |node|
  p node.class
end

Nodes have predicate methods, so if you want to find particular nodes just use a predicate:

ast = TinyGQL.parse "{ cool }"
p ast.find_all(&:field?).map(&:name) # => ["cool"]

If you need a more advanced way to iterate nodes, you can use a visitor:

class Viz
  include TinyGQL::Visitors::Visitor

  def handle_field obj
    p obj.name # => cool
    super
  end
end

ast = TinyGQL.parse "{ cool }"
ast.accept(Viz.new)

If you would like a functional way to collect data from the tree, use the Fold module:

module Fold
  extend TinyGQL::Visitors::Fold

  def self.handle_field obj, seed
    super(obj, seed + [obj.name])
  end
end

ast = TinyGQL.parse "{ neat { cool } }"
p ast.fold(Fold, []) # => ["neat", "cool"]

Nodes store their position in the source GraphQL document. If you'd like to extract the line number of the node, you'll need to keep a reference to the document and pass it to the line method on the node:

doc = <<-eod
mutation {
  likeStory(sturyID: 12345) {
    story {
      likeCount
    }
  }
}

eod

parser = TinyGQL::Parser.new doc
ast = parser.parse

ast.find_all(&:field?).each { |node|
  p node.name => node.line(doc)
}

I've licensed this code as Apache 2.0, but the lexer is from GraphQL-Ruby and is under the MIT license.