Parser: don't be so strict about validating Embeds (leave it up to the Card)

2 files changed, 38 insertions, 9 deletions

diff --git a/lib/pleroma/web/rich_media/parser.ex b/lib/pleroma/web/rich_media/parser.ex
index 66c90682c..0532e51ad 100644
--- a/lib/pleroma/web/rich_media/parser.ex
+++ b/lib/pleroma/web/rich_media/parser.ex
@@ -4,7 +4,6 @@
 
 defmodule Pleroma.Web.RichMedia.Parser do
   require Logger
-  alias Pleroma.Web.RichMedia.Parser.Card
   alias Pleroma.Web.RichMedia.Parser.Embed
 
   @cachex Pleroma.Config.get([:cachex, :provider], Cachex)
@@ -145,8 +144,7 @@ defmodule Pleroma.Web.RichMedia.Parser do
          {:ok, %Tesla.Env{body: json}} <-
            Pleroma.Web.RichMedia.Helpers.oembed_get(oembed_url),
          {:ok, data} <- Jason.decode(json),
-         embed <- %Embed{url: url, oembed: data},
-         {:ok, %Card{}} <- Card.validate(embed) do
+         embed <- %Embed{url: url, oembed: data} do
       {:ok, embed}
     else
       {:error, error} -> {:error, error}
@@ -157,8 +155,7 @@ defmodule Pleroma.Web.RichMedia.Parser do
   defp fetch_document(url) do
     with {:ok, %Tesla.Env{body: html}} <- Pleroma.Web.RichMedia.Helpers.rich_media_get(url),
          {:ok, html} <- Floki.parse_document(html),
-         %Embed{} = embed <- parse_embed(html, url),
-         {:ok, %Card{}} <- Card.validate(embed) do
+         %Embed{} = embed <- parse_embed(html, url) do
       {:ok, embed}
     else
       {:error, error} -> {:error, error}
diff --git a/test/pleroma/web/rich_media/parser_test.exs b/test/pleroma/web/rich_media/parser_test.exs
index f7d010219..8506953bb 100644
--- a/test/pleroma/web/rich_media/parser_test.exs
+++ b/test/pleroma/web/rich_media/parser_test.exs
@@ -83,8 +83,15 @@ defmodule Pleroma.Web.RichMedia.ParserTest do
     :ok
   end
 
-  test "returns error when no metadata present" do
-    assert {:error, _} = Parser.parse("http://example.com/empty")
+  test "returns empty embed when no metadata present" do
+    expected = %Embed{
+      meta: %{},
+      oembed: nil,
+      title: nil,
+      url: "http://example.com/empty"
+    }
+
+    assert Parser.parse("http://example.com/empty") == {:ok, expected}
   end
 
   test "parses ogp" do
@@ -166,8 +173,33 @@ defmodule Pleroma.Web.RichMedia.ParserTest do
     assert Parser.parse(url) == {:ok, expected}
   end
 
-  test "rejects invalid OGP data" do
-    assert {:error, _} = Parser.parse("http://example.com/malformed")
+  test "cleans corrupted meta data" do
+    expected = %Embed{
+      meta: %{
+        "Keywords" => "Konsument i zakupy",
+        "ROBOTS" => "NOARCHIVE",
+        "fb:app_id" => "515714931781741",
+        "fb:pages" => "288018984602680",
+        "google-site-verification" => "3P4BE3hLw82QWqtseIE60qQcOtrpMxMnCNkcv62pjTA",
+        "news_keywords" => "Konsument i zakupy",
+        "og:image" =>
+          "https://bi.im-g.pl/im/f7/49/17/z24418295FBW,Prace-nad-projektem-chusty-antysmogowej-rozpoczely.jpg",
+        "og:locale" => "pl_PL",
+        "og:site_name" => "wyborcza.biz",
+        "og:type" => "article",
+        "og:url" =>
+          "http://wyborcza.biz/biznes/7,147743,24417936,pomysl-na-biznes-chusta-ktora-chroni-przed-smogiem.html",
+        "twitter:card" => "summary_large_image",
+        "twitter:image" =>
+          "https://bi.im-g.pl/im/f7/49/17/z24418295FBW,Prace-nad-projektem-chusty-antysmogowej-rozpoczely.jpg",
+        "viewport" => "width=device-width, user-scalable=yes"
+      },
+      oembed: nil,
+      title: nil,
+      url: "http://example.com/malformed"
+    }
+
+    assert Parser.parse("http://example.com/malformed") == {:ok, expected}
   end
 
   test "returns error if getting page was not successful" do