feat: improve ruby parsing

[aaronlifton]

I noticed the ruby parsing didn't work for most ruby files, so I adjusted the treesitter query and implemented module and class definitions with proper namespaced names. Also implemented detection of private class methods.
Check the test result for test_ruby2:

It parses this:

# frozen_string_literal: true

require('jwt')

top_level_var = 1

def top_level_func
  inner_var_in_func = 2
end

module A
  module B
    @module_var = :foo

    def module_method
      @module_var
    end

    class C < Base
      TEST_CONST = 1
      @class_var = :bar
      attr_accessor :a, :b

      def initialize(a, b)
        @a = a
        @b = b
        super
      end

      def bar
        inner_var_in_method = 1
        true
      end

      private

      def baz(request, params)
        auth_header = request.headers['Authorization']
        parts = auth_header.try(:split, /\s+/)
        JWT.decode(parts.last)
      end
    end
  end
end

into

var top_level_var
func top_level_func() -> void
module A{}
module A::B{
  func module_method() -> void
  var @module_var
}
class A::B::C{
  func initialize(a, b) -> void
  func bar() -> void
  private func baz(request, params) -> void
  var TEST_CONST
  var @class_var
}

I create a new test test_ruby2 since I wanted to show the results of test_ruby with the changes. I noticed test_ruby includes invalid ruby syntax (functions in functions, classes in functions), so I figured it would be best to replace test_ruby with test_ruby2 since test_ruby2 includes the same tests but only valid syntax. If a user wants to create a repo map for a file with errors, I would think they should expect the parser to pick up the invalid syntax.