Typed structs on steroids based on algebraic types and pattern matching seamlessly integrating with standard Ruby features.
- Documentation: http://blog.pitr.ch/algebrick
- Source: https://github.com/pitr-ch/algebrick
- Blog: http://blog.pitr.ch/tag/algebrick.html
- Well defined data structures.
- Actor messages see new Actor implementation in concurrent-ruby.
- Describing message protocols in message-based cross-process communication. Algebraic types play nice with JSON de/serialization.
- and more...
Let's define a Tree
Tree = Algebrick.type do |tree|
variants Empty = atom,
Leaf = type { fields Integer },
Node = type { fields tree, tree }
end
Now types Tree(Empty | Leaf | Node)
, Empty
, Leaf(Integer)
and Node(Tree, Tree)
are defined.
Let's add a method, don't miss the pattern matching example.
module Tree
# compute depth of a tree
def depth
match self,
(on Empty, 0),
(on Leaf, 1),
# ~ will store and pass matched parts to variables left and right
(on Node.(~any, ~any) do |left, right|
1 + [left.depth, right.depth].max
end)
end
end
Method defined in module Tree
are passed down to all values of type Tree
Empty.depth # => 0
Leaf[10].depth # => 1
Node[Leaf[4], Empty].depth # => 2
Node[Empty, Node[Leaf[1], Empty]].depth # => 3