Make filter_out faster

[preciz]

mapper is only called for [:text, :comment] cases and takes a list so it's safe to refactor the Stream into a recursion I believe.

There is a 10x speedup in this benchmark:

  Mix.install([
    {:benchee, "~> 1.0"},
    {:floki, path: "."}
  ])

  read_file = fn name ->
    __ENV__.file
    |> Path.dirname()
    |> Path.join(name)
    |> File.read!()
    |> Floki.parse_document!()
  end

  doc = read_file.("big.html")
  doc_with_comments = [{"div", [], [{:comment, "comment 1"}, "text", {:comment, "comment 2"}]} | doc]


  IO.puts "Document parsed. Running benchmark..."

  Benchee.run(
    %{
      "filter_out :comment" => fn -> Floki.filter_out(doc_with_comments, :comment) end,
      "filter_out :text" => fn -> Floki.filter_out(doc_with_comments, :text) end
    },
    time: 5,
    memory_time: 2
  )

[Copilot]

Pull request overview

This PR optimizes the filter_out function by replacing Stream-based operations with direct tail recursion, achieving a reported 10x performance improvement. The optimization specifically targets the filtering of :text and :comment nodes from HTML trees.

Key changes:

• Replaces Stream.filter and Stream.map with a tail-recursive do_mapper helper function
• Uses an accumulator pattern with a final reverse to maintain list order
• Eliminates Stream enumeration overhead while preserving identical behavior

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[philss]

🚀 🚀 🚀 thanks!