Merge branch 'fix/ogp-title' into 'develop'

Rich Media Parser: Attempt to use <title> from HTML as a fallback

Closes #979

See merge request pleroma/pleroma!1277
This commit is contained in:
lain 2019-06-13 13:26:39 +00:00
commit 6f5a0cc14e
3 changed files with 59 additions and 8 deletions

View File

@ -1,15 +1,19 @@
defmodule Pleroma.Web.RichMedia.Parsers.MetaTagsParser do
def parse(html, data, prefix, error_message, key_name, value_name \\ "content") do
with elements = [_ | _] <- get_elements(html, key_name, prefix),
meta_data =
Enum.reduce(elements, data, fn el, acc ->
attributes = normalize_attributes(el, prefix, key_name, value_name)
meta_data =
html
|> get_elements(key_name, prefix)
|> Enum.reduce(data, fn el, acc ->
attributes = normalize_attributes(el, prefix, key_name, value_name)
Map.merge(acc, attributes)
end) do
{:ok, meta_data}
Map.merge(acc, attributes)
end)
|> maybe_put_title(html)
if Enum.empty?(meta_data) do
{:error, error_message}
else
_e -> {:error, error_message}
{:ok, meta_data}
end
end
@ -27,4 +31,17 @@ defp normalize_attributes(html_node, prefix, key_name, value_name) do
%{String.to_atom(data[key_name]) => data[value_name]}
end
defp maybe_put_title(%{title: _} = meta, _), do: meta
defp maybe_put_title(meta, html) do
case get_page_title(html) do
"" -> meta
title -> Map.put_new(meta, :title, title)
end
end
defp get_page_title(html) do
Floki.find(html, "title") |> Floki.text()
end
end

View File

@ -0,0 +1,12 @@
<html prefix="og: http://ogp.me/ns#">
<head>
<title>The Rock (1996)</title>
<meta property="og:type" content="video.movie" />
<meta property="og:url" content="http://www.imdb.com/title/tt0117500/" />
<meta property="og:image" content="http://ia.media-imdb.com/images/rock.jpg" />
<meta property="og:description"
content="Directed by Michael Bay. With Sean Connery, Nicolas Cage, Ed Harris, John Spencer.">
</head>
</html>

View File

@ -9,6 +9,15 @@ defmodule Pleroma.Web.RichMedia.ParserTest do
} ->
%Tesla.Env{status: 200, body: File.read!("test/fixtures/rich_media/ogp.html")}
%{
method: :get,
url: "http://example.com/ogp-missing-title"
} ->
%Tesla.Env{
status: 200,
body: File.read!("test/fixtures/rich_media/ogp-missing-title.html")
}
%{
method: :get,
url: "http://example.com/twitter-card"
@ -51,6 +60,19 @@ test "parses ogp" do
}}
end
test "falls back to <title> when ogp:title is missing" do
assert Pleroma.Web.RichMedia.Parser.parse("http://example.com/ogp-missing-title") ==
{:ok,
%{
image: "http://ia.media-imdb.com/images/rock.jpg",
title: "The Rock (1996)",
description:
"Directed by Michael Bay. With Sean Connery, Nicolas Cage, Ed Harris, John Spencer.",
type: "video.movie",
url: "http://www.imdb.com/title/tt0117500/"
}}
end
test "parses twitter card" do
assert Pleroma.Web.RichMedia.Parser.parse("http://example.com/twitter-card") ==
{:ok,