git.squeep.com Git - akkoma/blob - lib/pleroma/html.ex

   1 # Pleroma: A lightweight social networking server
   2 # Copyright © 2017-2019 Pleroma Authors <https://pleroma.social/>
   3 # SPDX-License-Identifier: AGPL-3.0-only
   4
   5 defmodule Pleroma.HTML do
   6   alias HtmlSanitizeEx.Scrubber
   7
   8   defp get_scrubbers(scrubber) when is_atom(scrubber), do: [scrubber]
   9   defp get_scrubbers(scrubbers) when is_list(scrubbers), do: scrubbers
  10   defp get_scrubbers(_), do: [Pleroma.HTML.Scrubber.Default]
  11
  12   def get_scrubbers() do
  13     Pleroma.Config.get([:markup, :scrub_policy])
  14     |> get_scrubbers
  15   end
  16
  17   def filter_tags(html, nil) do
  18     filter_tags(html, get_scrubbers())
  19   end
  20
  21   def filter_tags(html, scrubbers) when is_list(scrubbers) do
  22     Enum.reduce(scrubbers, html, fn scrubber, html ->
  23       filter_tags(html, scrubber)
  24     end)
  25   end
  26
  27   def filter_tags(html, scrubber), do: Scrubber.scrub(html, scrubber)
  28   def filter_tags(html), do: filter_tags(html, nil)
  29   def strip_tags(html), do: Scrubber.scrub(html, Scrubber.StripTags)
  30
  31   def get_cached_scrubbed_html_for_object(content, scrubbers, object, module) do
  32     key = "#{module}#{generate_scrubber_signature(scrubbers)}|#{object.id}"
  33     Cachex.fetch!(:scrubber_cache, key, fn _key -> ensure_scrubbed_html(content, scrubbers) end)
  34   end
  35
  36   def get_cached_stripped_html_for_object(content, object, module) do
  37     get_cached_scrubbed_html_for_object(
  38       content,
  39       HtmlSanitizeEx.Scrubber.StripTags,
  40       object,
  41       module
  42     )
  43   end
  44
  45   def ensure_scrubbed_html(
  46         content,
  47         scrubbers
  48       ) do
  49     {:commit, filter_tags(content, scrubbers)}
  50   end
  51
  52   defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
  53     generate_scrubber_signature([scrubber])
  54   end
  55
  56   defp generate_scrubber_signature(scrubbers) do
  57     Enum.reduce(scrubbers, "", fn scrubber, signature ->
  58       "#{signature}#{to_string(scrubber)}"
  59     end)
  60   end
  61
  62   def extract_first_external_url(_, nil), do: {:error, "No content"}
  63
  64   def extract_first_external_url(object, content) do
  65     key = "URL|#{object.id}"
  66
  67     Cachex.fetch!(:scrubber_cache, key, fn _key ->
  68       result =
  69         content
  70         |> Floki.filter_out("a.mention")
  71         |> Floki.attribute("a", "href")
  72         |> Enum.at(0)
  73
  74       {:commit, {:ok, result}}
  75     end)
  76   end
  77 end
  78
  79 defmodule Pleroma.HTML.Scrubber.TwitterText do
  80   @moduledoc """
  81   An HTML scrubbing policy which limits to twitter-style text.  Only
  82   paragraphs, breaks and links are allowed through the filter.
  83   """
  84
  85   @markup Application.get_env(:pleroma, :markup)
  86   @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
  87
  88   require HtmlSanitizeEx.Scrubber.Meta
  89   alias HtmlSanitizeEx.Scrubber.Meta
  90
  91   Meta.remove_cdata_sections_before_scrub()
  92   Meta.strip_comments()
  93
  94   # links
  95   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
  96   Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
  97
  98   # paragraphs and linebreaks
  99   Meta.allow_tag_with_these_attributes("br", [])
 100   Meta.allow_tag_with_these_attributes("p", [])
 101
 102   # microformats
 103   Meta.allow_tag_with_these_attributes("span", ["class"])
 104
 105   # allow inline images for custom emoji
 106   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 107
 108   if @allow_inline_images do
 109     # restrict img tags to http/https only, because of MediaProxy.
 110     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 111
 112     Meta.allow_tag_with_these_attributes("img", [
 113       "width",
 114       "height",
 115       "title",
 116       "alt"
 117     ])
 118   end
 119
 120   Meta.strip_everything_not_covered()
 121 end
 122
 123 defmodule Pleroma.HTML.Scrubber.Default do
 124   @doc "The default HTML scrubbing policy: no "
 125
 126   require HtmlSanitizeEx.Scrubber.Meta
 127   alias HtmlSanitizeEx.Scrubber.Meta
 128
 129   @markup Application.get_env(:pleroma, :markup)
 130   @valid_schemes Pleroma.Config.get([:uri_schemes, :valid_schemes], [])
 131
 132   Meta.remove_cdata_sections_before_scrub()
 133   Meta.strip_comments()
 134
 135   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
 136   Meta.allow_tag_with_these_attributes("a", ["name", "title", "class"])
 137
 138   Meta.allow_tag_with_these_attributes("abbr", ["title"])
 139
 140   Meta.allow_tag_with_these_attributes("b", [])
 141   Meta.allow_tag_with_these_attributes("blockquote", [])
 142   Meta.allow_tag_with_these_attributes("br", [])
 143   Meta.allow_tag_with_these_attributes("code", [])
 144   Meta.allow_tag_with_these_attributes("del", [])
 145   Meta.allow_tag_with_these_attributes("em", [])
 146   Meta.allow_tag_with_these_attributes("i", [])
 147   Meta.allow_tag_with_these_attributes("li", [])
 148   Meta.allow_tag_with_these_attributes("ol", [])
 149   Meta.allow_tag_with_these_attributes("p", [])
 150   Meta.allow_tag_with_these_attributes("pre", [])
 151   Meta.allow_tag_with_these_attributes("span", ["class"])
 152   Meta.allow_tag_with_these_attributes("strong", [])
 153   Meta.allow_tag_with_these_attributes("u", [])
 154   Meta.allow_tag_with_these_attributes("ul", [])
 155
 156   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 157
 158   if @allow_inline_images do
 159     # restrict img tags to http/https only, because of MediaProxy.
 160     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 161
 162     Meta.allow_tag_with_these_attributes("img", [
 163       "width",
 164       "height",
 165       "title",
 166       "alt"
 167     ])
 168   end
 169
 170   @allow_tables Keyword.get(@markup, :allow_tables)
 171
 172   if @allow_tables do
 173     Meta.allow_tag_with_these_attributes("table", [])
 174     Meta.allow_tag_with_these_attributes("tbody", [])
 175     Meta.allow_tag_with_these_attributes("td", [])
 176     Meta.allow_tag_with_these_attributes("th", [])
 177     Meta.allow_tag_with_these_attributes("thead", [])
 178     Meta.allow_tag_with_these_attributes("tr", [])
 179   end
 180
 181   @allow_headings Keyword.get(@markup, :allow_headings)
 182
 183   if @allow_headings do
 184     Meta.allow_tag_with_these_attributes("h1", [])
 185     Meta.allow_tag_with_these_attributes("h2", [])
 186     Meta.allow_tag_with_these_attributes("h3", [])
 187     Meta.allow_tag_with_these_attributes("h4", [])
 188     Meta.allow_tag_with_these_attributes("h5", [])
 189   end
 190
 191   @allow_fonts Keyword.get(@markup, :allow_fonts)
 192
 193   if @allow_fonts do
 194     Meta.allow_tag_with_these_attributes("font", ["face"])
 195   end
 196
 197   Meta.strip_everything_not_covered()
 198 end
 199
 200 defmodule Pleroma.HTML.Transform.MediaProxy do
 201   @moduledoc "Transforms inline image URIs to use MediaProxy."
 202
 203   alias Pleroma.Web.MediaProxy
 204
 205   def before_scrub(html), do: html
 206
 207   def scrub_attribute("img", {"src", "http" <> target}) do
 208     media_url =
 209       ("http" <> target)
 210       |> MediaProxy.url()
 211
 212     {"src", media_url}
 213   end
 214
 215   def scrub_attribute(_tag, attribute), do: attribute
 216
 217   def scrub({"img", attributes, children}) do
 218     attributes =
 219       attributes
 220       |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
 221       |> Enum.reject(&is_nil(&1))
 222
 223     {"img", attributes, children}
 224   end
 225
 226   def scrub({:comment, _children}), do: ""
 227
 228   def scrub({tag, attributes, children}), do: {tag, attributes, children}
 229   def scrub({_tag, children}), do: children
 230   def scrub(text), do: text
 231 end