+ defp get_scrubbers(scrubber) when is_atom(scrubber), do: [scrubber]
+ defp get_scrubbers(scrubbers) when is_list(scrubbers), do: scrubbers
+ defp get_scrubbers(_), do: [Pleroma.HTML.Scrubber.Default]
+
+ def get_scrubbers do
+ Pleroma.Config.get([:markup, :scrub_policy])
+ |> get_scrubbers
+ end
+
+ def filter_tags(html, nil) do
+ filter_tags(html, get_scrubbers())
+ end
+
+ def filter_tags(html, scrubbers) when is_list(scrubbers) do
+ Enum.reduce(scrubbers, html, fn scrubber, html ->
+ filter_tags(html, scrubber)
+ end)
+ end
+
+ def filter_tags(html, scrubber), do: Scrubber.scrub(html, scrubber)
+ def filter_tags(html), do: filter_tags(html, nil)
+ def strip_tags(html), do: Scrubber.scrub(html, Scrubber.StripTags)
+
+ def get_cached_scrubbed_html_for_activity(
+ content,
+ scrubbers,
+ activity,
+ key \\ "",
+ callback \\ fn x -> x end
+ ) do
+ key = "#{key}#{generate_scrubber_signature(scrubbers)}|#{activity.id}"
+
+ Cachex.fetch!(:scrubber_cache, key, fn _key ->
+ object = Pleroma.Object.normalize(activity)
+ ensure_scrubbed_html(content, scrubbers, object.data["fake"] || false, callback)
+ end)
+ end
+
+ def get_cached_stripped_html_for_activity(content, activity, key) do
+ get_cached_scrubbed_html_for_activity(
+ content,
+ HtmlSanitizeEx.Scrubber.StripTags,
+ activity,
+ key,
+ &HtmlEntities.decode/1
+ )
+ end
+
+ def ensure_scrubbed_html(
+ content,
+ scrubbers,
+ fake,
+ callback
+ ) do
+ content =
+ content
+ |> filter_tags(scrubbers)
+ |> callback.()
+
+ if fake do
+ {:ignore, content}
+ else
+ {:commit, content}
+ end
+ end
+
+ defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
+ generate_scrubber_signature([scrubber])
+ end
+
+ defp generate_scrubber_signature(scrubbers) do
+ Enum.reduce(scrubbers, "", fn scrubber, signature ->
+ "#{signature}#{to_string(scrubber)}"
+ end)
+ end
+
+ def extract_first_external_url(_, nil), do: {:error, "No content"}
+
+ def extract_first_external_url(object, content) do
+ key = "URL|#{object.id}"
+
+ Cachex.fetch!(:scrubber_cache, key, fn _key ->
+ result =
+ content
+ |> Floki.filter_out("a.mention")
+ |> Floki.attribute("a", "href")
+ |> Enum.at(0)
+
+ {:commit, {:ok, result}}
+ end)
+ end
+end
+
+defmodule Pleroma.HTML.Scrubber.TwitterText do
+ @moduledoc """
+ An HTML scrubbing policy which limits to twitter-style text. Only
+ paragraphs, breaks and links are allowed through the filter.
+ """
+
+ @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
+
+ require HtmlSanitizeEx.Scrubber.Meta
+ alias HtmlSanitizeEx.Scrubber.Meta
+
+ Meta.remove_cdata_sections_before_scrub()
+ Meta.strip_comments()
+
+ # links
+ Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
+
+ Meta.allow_tag_with_this_attribute_values("a", "class", [
+ "hashtag",
+ "u-url",
+ "mention",
+ "u-url mention",
+ "mention u-url"
+ ])
+
+ Meta.allow_tag_with_this_attribute_values("a", "rel", [
+ "tag",
+ "nofollow",
+ "noopener",
+ "noreferrer"
+ ])
+
+ Meta.allow_tag_with_these_attributes("a", ["name", "title"])
+
+ # paragraphs and linebreaks
+ Meta.allow_tag_with_these_attributes("br", [])
+ Meta.allow_tag_with_these_attributes("p", [])
+
+ # microformats
+ Meta.allow_tag_with_this_attribute_values("span", "class", ["h-card"])
+ Meta.allow_tag_with_these_attributes("span", [])