firehose/blogex/lib/blogex/external_links.ex

defmodule Blogex.ExternalLinks do
  @moduledoc """
  Post-processes compiled post HTML so that external links open in a new
  tab and carry a marker class for styling.

  Anchors whose `href` starts with `http://` or `https://` get:

    * `target="_blank"` — the link opens in a new tab
    * `rel="noopener"`  — the new tab cannot access `window.opener`
    * `class="blogex-external"` — hook for CSS to draw an external-link
      marker icon (the host app's stylesheet owns the actual styling)
    * a `<span class="blogex-external-hint"> (opens in a new tab)</span>`
      at the end of the link text — announced by screen readers, hidden
      visually by the host app's stylesheet

  Internal links (relative paths), `mailto:` and `#anchor` links are left
  untouched.

  The rewrite is idempotent: applying it to already-rewritten HTML does
  not duplicate attributes or hints.

  Assumptions: the HTML comes from NimblePublisher's Earmark conversion,
  which emits double-quoted, lowercase attribute names and never adds
  `target`, `rel` or `class` to anchors. Attribute values are matched
  with case-insensitive regexes to stay robust for other producers.
  Anchors are matched together with their closing tag (the hint goes
  inside), so a stray unclosed `<a>` is left alone.
  """

  @marker_class "blogex-external"
  @hint_class "blogex-external-hint"
  @hint_text " (opens in a new tab)"
  @external_href ~r|^https?://|i
  @anchor ~r|(<a\s[^>]*>)(.*?)(</a>)|is

  @doc """
  Rewrites all external anchors in an HTML fragment.
  """
  @spec rewrite(String.t()) :: String.t()
  def rewrite(html) when is_binary(html) do
    Regex.replace(@anchor, html, fn _anchor, opening, content, closing ->
      {opening, content} = decorate_anchor(opening, content)
      opening <> content <> closing
    end)
  end

  defp decorate_anchor(opening, content) do
    if external?(href(opening)) do
      opening =
        opening
        |> ensure_attribute("target", "_blank")
        |> ensure_rel("noopener")
        |> ensure_class(@marker_class)

      {opening, ensure_hint(content)}
    else
      {opening, content}
    end
  end

  defp ensure_hint(content) do
    if String.contains?(content, @hint_class) do
      content
    else
      content <> ~s|<span class="#{@hint_class}">#{@hint_text}</span>|
    end
  end

  defp href(anchor) do
    case Regex.run(~r/\shref\s*=\s*"([^"]*)"/i, anchor) do
      [_match, href] -> href
      _ -> nil
    end
  end

  defp external?(href), do: is_binary(href) and Regex.match?(@external_href, href)

  defp ensure_attribute(anchor, name, value) do
    if Regex.match?(~r/\s#{name}\s*=/i, anchor) do
      anchor
    else
      String.replace_suffix(anchor, ">", ~s( #{name}="#{value}">))
    end
  end

  defp ensure_rel(anchor, keyword) do
    case Regex.run(~r/\srel\s*=\s*"([^"]*)"/i, anchor) do
      [_match, existing] ->
        add_space_separated(anchor, "rel", existing, keyword)

      _ ->
        ensure_attribute(anchor, "rel", keyword)
    end
  end

  defp ensure_class(anchor, class) do
    case Regex.run(~r/\sclass\s*=\s*"([^"]*)"/i, anchor) do
      [_match, existing] ->
        add_space_separated(anchor, "class", existing, class)

      _ ->
        ensure_attribute(anchor, "class", class)
    end
  end

  # Appends a keyword/class to an existing space-separated attribute value,
  # unless it is already present.
  defp add_space_separated(anchor, name, existing, addition) do
    if addition in String.split(existing) do
      anchor
    else
      String.replace(anchor, ~s(#{name}="#{existing}"), ~s(#{name}="#{existing} #{addition}"))
    end
  end
end