Make Floki use fast_html

This commit is contained in:
rinpatch 2020-02-11 00:29:25 +03:00
parent 58299fcfb4
commit ea1631d7e6
7 changed files with 30 additions and 13 deletions

View file

@ -612,6 +612,8 @@
config :pleroma, configurable_from_database: false
config :floki, :html_parser, Floki.HTMLParser.FastHtml
# Import environment specific config. This must remain at the bottom
# of this file so it overrides the configuration defined above.
import_config "#{Mix.env()}.exs"

View file

@ -108,6 +108,7 @@ def extract_first_external_url(object, content) do
Cachex.fetch!(:scrubber_cache, key, fn _key ->
result =
content
|> Floki.parse_fragment!()
|> Floki.filter_out("a.mention,a.hashtag,a[rel~=\"tag\"]")
|> Floki.attribute("a", "href")
|> Enum.at(0)

View file

@ -17,6 +17,7 @@ defp old_user?(%User{} = u) do
# does the post contain links?
defp contains_links?(%{"content" => content} = _object) do
content
|> Floki.parse_fragment!()
|> Floki.filter_out("a.mention,a.hashtag,a[rel~=\"tag\"],a.zrl")
|> Floki.attribute("a", "href")
|> length() > 0

View file

@ -8,8 +8,10 @@ defmodule Pleroma.Web.Metadata.Providers.RelMe do
@impl Provider
def build_tags(%{user: user}) do
(Floki.attribute(user.bio, "link[rel~=me]", "href") ++
Floki.attribute(user.bio, "a[rel~=me]", "href"))
bio_tree = Floki.parse_fragment!(user.bio)
(Floki.attribute(bio_tree, "link[rel~=me]", "href") ++
Floki.attribute(bio_tree, "a[rel~=me]", "href"))
|> Enum.map(fn link ->
{:link, [rel: "me", href: link], []}
end)

View file

@ -27,9 +27,10 @@ def parse(_), do: {:error, "No URL provided"}
defp parse_url(url) do
with {:ok, %Tesla.Env{body: html, status: status}} when status in 200..299 <-
Pleroma.HTTP.get(url, [], adapter: @hackney_options),
{:ok, html_tree} <- Floki.parse_document(html),
data <-
Floki.attribute(html, "link[rel~=me]", "href") ++
Floki.attribute(html, "a[rel~=me]", "href") do
Floki.attribute(html_tree, "link[rel~=me]", "href") ++
Floki.attribute(html_tree, "a[rel~=me]", "href") do
{:ok, data}
end
rescue

View file

@ -81,18 +81,18 @@ defp parse_url(url) do
{:ok, %Tesla.Env{body: html}} = Pleroma.HTTP.get(url, [], adapter: @hackney_options)
html
|> parse_html
|> parse_html()
|> maybe_parse()
|> Map.put(:url, url)
|> clean_parsed_data()
|> check_parsed_data()
rescue
e ->
{:error, "Parsing error: #{inspect(e)}"}
{:error, "Parsing error: #{inspect(e)} #{inspect(__STACKTRACE__)}"}
end
end
defp parse_html(html), do: Floki.parse(html)
defp parse_html(html), do: Floki.parse_document!(html)
defp maybe_parse(html) do
Enum.reduce_while(parsers(), %{}, fn parser, acc ->

View file

@ -7,11 +7,14 @@ defmodule Pleroma.Web.RichMedia.Parsers.TwitterCardTest do
alias Pleroma.Web.RichMedia.Parsers.TwitterCard
test "returns error when html not contains twitter card" do
assert TwitterCard.parse("", %{}) == {:error, "No twitter card metadata found"}
assert TwitterCard.parse([{"html", [], [{"head", [], []}, {"body", [], []}]}], %{}) ==
{:error, "No twitter card metadata found"}
end
test "parses twitter card with only name attributes" do
html = File.read!("test/fixtures/nypd-facial-recognition-children-teenagers3.html")
html =
File.read!("test/fixtures/nypd-facial-recognition-children-teenagers3.html")
|> Floki.parse_document!()
assert TwitterCard.parse(html, %{}) ==
{:ok,
@ -26,7 +29,9 @@ test "parses twitter card with only name attributes" do
end
test "parses twitter card with only property attributes" do
html = File.read!("test/fixtures/nypd-facial-recognition-children-teenagers2.html")
html =
File.read!("test/fixtures/nypd-facial-recognition-children-teenagers2.html")
|> Floki.parse_document!()
assert TwitterCard.parse(html, %{}) ==
{:ok,
@ -45,7 +50,9 @@ test "parses twitter card with only property attributes" do
end
test "parses twitter card with name & property attributes" do
html = File.read!("test/fixtures/nypd-facial-recognition-children-teenagers.html")
html =
File.read!("test/fixtures/nypd-facial-recognition-children-teenagers.html")
|> Floki.parse_document!()
assert TwitterCard.parse(html, %{}) ==
{:ok,
@ -73,7 +80,8 @@ test "respect only first title tag on the page" do
"YTQ5MF9EQVIgZXhodW1hdGlvbiBvZiBNYXJnYXJldCBDb3JiaW4gZ3JhdmUgMTkyNi5qcGciXSxbInAiLCJjb252ZXJ0IiwiIl0sWyJwIiwiY29udmVydCIsIi1xdWFsaXR5IDgxIC1hdXRvLW9" <>
"yaWVudCJdLFsicCIsInRodW1iIiwiNjAweD4iXV0/DAR%20exhumation%20of%20Margaret%20Corbin%20grave%201926.jpg"
html = File.read!("test/fixtures/margaret-corbin-grave-west-point.html")
html =
File.read!("test/fixtures/margaret-corbin-grave-west-point.html") |> Floki.parse_document!()
assert TwitterCard.parse(html, %{}) ==
{:ok,
@ -87,7 +95,9 @@ test "respect only first title tag on the page" do
end
test "takes first founded title in html head if there is html markup error" do
html = File.read!("test/fixtures/nypd-facial-recognition-children-teenagers4.html")
html =
File.read!("test/fixtures/nypd-facial-recognition-children-teenagers4.html")
|> Floki.parse_document!()
assert TwitterCard.parse(html, %{}) ==
{:ok,