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(content, HtmlSanitizeEx.Scrubber.StripTags, object, module)
  38   end
  39
  40   def ensure_scrubbed_html(
  41         content,
  42         scrubbers
  43       ) do
  44     {:commit, filter_tags(content, scrubbers)}
  45   end
  46
  47   defp generate_scrubber_signature(scrubber) when is_atom(scrubber) do
  48     generate_scrubber_signature([scrubber])
  49   end
  50
  51   defp generate_scrubber_signature(scrubbers) do
  52     Enum.reduce(scrubbers, "", fn scrubber, signature ->
  53       "#{signature}#{to_string(scrubber)}"
  54     end)
  55   end
  56 end
  57
  58 defmodule Pleroma.HTML.Scrubber.TwitterText do
  59   @moduledoc """
  60   An HTML scrubbing policy which limits to twitter-style text.  Only
  61   paragraphs, breaks and links are allowed through the filter.
  62   """
  63
  64   @markup Application.get_env(:pleroma, :markup)
  65   @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
  66   @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
  67
  68   require HtmlSanitizeEx.Scrubber.Meta
  69   alias HtmlSanitizeEx.Scrubber.Meta
  70
  71   Meta.remove_cdata_sections_before_scrub()
  72   Meta.strip_comments()
  73
  74   # links
  75   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
  76   Meta.allow_tag_with_these_attributes("a", ["name", "title"])
  77
  78   # paragraphs and linebreaks
  79   Meta.allow_tag_with_these_attributes("br", [])
  80   Meta.allow_tag_with_these_attributes("p", [])
  81
  82   # microformats
  83   Meta.allow_tag_with_these_attributes("span", [])
  84
  85   # allow inline images for custom emoji
  86   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
  87
  88   if @allow_inline_images do
  89     # restrict img tags to http/https only, because of MediaProxy.
  90     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
  91
  92     Meta.allow_tag_with_these_attributes("img", [
  93       "width",
  94       "height",
  95       "title",
  96       "alt"
  97     ])
  98   end
  99
 100   Meta.strip_everything_not_covered()
 101 end
 102
 103 defmodule Pleroma.HTML.Scrubber.Default do
 104   @doc "The default HTML scrubbing policy: no "
 105
 106   require HtmlSanitizeEx.Scrubber.Meta
 107   alias HtmlSanitizeEx.Scrubber.Meta
 108
 109   @markup Application.get_env(:pleroma, :markup)
 110   @uri_schemes Application.get_env(:pleroma, :uri_schemes, [])
 111   @valid_schemes Keyword.get(@uri_schemes, :valid_schemes, [])
 112
 113   Meta.remove_cdata_sections_before_scrub()
 114   Meta.strip_comments()
 115
 116   Meta.allow_tag_with_uri_attributes("a", ["href", "data-user", "data-tag"], @valid_schemes)
 117   Meta.allow_tag_with_these_attributes("a", ["name", "title"])
 118
 119   Meta.allow_tag_with_these_attributes("abbr", ["title"])
 120
 121   Meta.allow_tag_with_these_attributes("b", [])
 122   Meta.allow_tag_with_these_attributes("blockquote", [])
 123   Meta.allow_tag_with_these_attributes("br", [])
 124   Meta.allow_tag_with_these_attributes("code", [])
 125   Meta.allow_tag_with_these_attributes("del", [])
 126   Meta.allow_tag_with_these_attributes("em", [])
 127   Meta.allow_tag_with_these_attributes("i", [])
 128   Meta.allow_tag_with_these_attributes("li", [])
 129   Meta.allow_tag_with_these_attributes("ol", [])
 130   Meta.allow_tag_with_these_attributes("p", [])
 131   Meta.allow_tag_with_these_attributes("pre", [])
 132   Meta.allow_tag_with_these_attributes("span", [])
 133   Meta.allow_tag_with_these_attributes("strong", [])
 134   Meta.allow_tag_with_these_attributes("u", [])
 135   Meta.allow_tag_with_these_attributes("ul", [])
 136
 137   @allow_inline_images Keyword.get(@markup, :allow_inline_images)
 138
 139   if @allow_inline_images do
 140     # restrict img tags to http/https only, because of MediaProxy.
 141     Meta.allow_tag_with_uri_attributes("img", ["src"], ["http", "https"])
 142
 143     Meta.allow_tag_with_these_attributes("img", [
 144       "width",
 145       "height",
 146       "title",
 147       "alt"
 148     ])
 149   end
 150
 151   @allow_tables Keyword.get(@markup, :allow_tables)
 152
 153   if @allow_tables do
 154     Meta.allow_tag_with_these_attributes("table", [])
 155     Meta.allow_tag_with_these_attributes("tbody", [])
 156     Meta.allow_tag_with_these_attributes("td", [])
 157     Meta.allow_tag_with_these_attributes("th", [])
 158     Meta.allow_tag_with_these_attributes("thead", [])
 159     Meta.allow_tag_with_these_attributes("tr", [])
 160   end
 161
 162   @allow_headings Keyword.get(@markup, :allow_headings)
 163
 164   if @allow_headings do
 165     Meta.allow_tag_with_these_attributes("h1", [])
 166     Meta.allow_tag_with_these_attributes("h2", [])
 167     Meta.allow_tag_with_these_attributes("h3", [])
 168     Meta.allow_tag_with_these_attributes("h4", [])
 169     Meta.allow_tag_with_these_attributes("h5", [])
 170   end
 171
 172   @allow_fonts Keyword.get(@markup, :allow_fonts)
 173
 174   if @allow_fonts do
 175     Meta.allow_tag_with_these_attributes("font", ["face"])
 176   end
 177
 178   Meta.strip_everything_not_covered()
 179 end
 180
 181 defmodule Pleroma.HTML.Transform.MediaProxy do
 182   @moduledoc "Transforms inline image URIs to use MediaProxy."
 183
 184   alias Pleroma.Web.MediaProxy
 185
 186   def before_scrub(html), do: html
 187
 188   def scrub_attribute("img", {"src", "http" <> target}) do
 189     media_url =
 190       ("http" <> target)
 191       |> MediaProxy.url()
 192
 193     {"src", media_url}
 194   end
 195
 196   def scrub_attribute(_tag, attribute), do: attribute
 197
 198   def scrub({"img", attributes, children}) do
 199     attributes =
 200       attributes
 201       |> Enum.map(fn attr -> scrub_attribute("img", attr) end)
 202       |> Enum.reject(&is_nil(&1))
 203
 204     {"img", attributes, children}
 205   end
 206
 207   def scrub({:comment, _children}), do: ""
 208
 209   def scrub({tag, attributes, children}), do: {tag, attributes, children}
 210   def scrub({_tag, children}), do: children
 211   def scrub(text), do: text
 212 end