How can we add support clean ups in ERB file

[Bennet-Sunder]

templating languages like EJS and ERB allow you to generate HTML by writing a mixture of HTML and another language like JavaScript or Ruby.

For the below Erb code snippet

<% if flag_enabled?(flag_name) %>
    <ul>
    <% people.each do |person| %>
        <li><%= person.name %></li>
    <% end %>
    </ul>
<% end %>

The corresponding tree generated by (https://github.com/tree-sitter/tree-sitter-embedded-template) parser is

(template [0, 0] - [6, 9]
  
  (directive [0, 0] - [0, 33]
    (code [0, 2] - [0, 31]))
  
  (content [0, 33] - [2, 4])
  
  (directive [2, 4] - [2, 33]
    (code [2, 6] - [2, 31]))
  
  (content [2, 33] - [3, 12])
  
  (output_directive [3, 12] - [3, 30]
    (code [3, 15] - [3, 28]))
  
  (content [3, 30] - [4, 4])
  
  (directive [4, 4] - [4, 13]
    (code [4, 6] - [4, 11]))
  
  (content [4, 13] - [6, 0])
  
  (directive [6, 0] - [6, 9]
    (code [6, 2] - [6, 7]))
)

The code here refers to the the ruby code and content is the html content. However, the parser doesn't recognise content node as a sibling of the code node. All code and content nodes are considered siblings of the root template node.

For a cleanup perspective this isn't helpful since it parses individual ruby lines without getting the nested sibling content nodes.

From (https://tree-sitter.github.io/tree-sitter/using-parsers#multi-language-documents)

This API allows for great flexibility in how languages can be composed. Tree-sitter is not responsible for mediating the interactions between languages. Instead, you are free to do that using arbitrary application-specific logic.

Any thoughts on if it is possible to support cleanup for Erb files in the existing polyglot piranha implementation?