3 files changed, 46 insertions, 37 deletions

diff --git a/lib/pleroma/web/rich_media/helpers.ex b/lib/pleroma/web/rich_media/helpers.ex
index 6210f2c5a..752ca9f81 100644
--- a/lib/pleroma/web/rich_media/helpers.ex
+++ b/lib/pleroma/web/rich_media/helpers.ex
@@ -58,7 +58,7 @@ defmodule Pleroma.Web.RichMedia.Helpers do
     with true <- Config.get([:rich_media, :enabled]),
          false <- object.data["sensitive"] || false,
          {:ok, page_url} <-
-           HTML.extract_first_external_url(object, object.data["content"]),
+           HTML.extract_first_external_url_from_object(object),
          :ok <- validate_page_url(page_url),
          {:ok, rich_media} <- Parser.parse(page_url) do
       %{page_url: page_url, rich_media: rich_media}
@@ -96,6 +96,6 @@ defmodule Pleroma.Web.RichMedia.Helpers do
         @rich_media_options
       end
 
-    Pleroma.HTTP.get(url, headers, options)
+    Pleroma.HTTP.get(url, headers, adapter: options)
   end
 end
diff --git a/lib/pleroma/web/rich_media/parser.ex b/lib/pleroma/web/rich_media/parser.ex
index ca592833f..e98c743ca 100644
--- a/lib/pleroma/web/rich_media/parser.ex
+++ b/lib/pleroma/web/rich_media/parser.ex
@@ -3,6 +3,8 @@
 # SPDX-License-Identifier: AGPL-3.0-only
 
 defmodule Pleroma.Web.RichMedia.Parser do
+  require Logger
+
   defp parsers do
     Pleroma.Config.get([:rich_media, :parsers])
   end
@@ -10,17 +12,29 @@ defmodule Pleroma.Web.RichMedia.Parser do
   def parse(nil), do: {:error, "No URL provided"}
 
   if Pleroma.Config.get(:env) == :test do
+    @spec parse(String.t()) :: {:ok, map()} | {:error, any()}
     def parse(url), do: parse_url(url)
   else
+    @spec parse(String.t()) :: {:ok, map()} | {:error, any()}
     def parse(url) do
-      try do
-        Cachex.fetch!(:rich_media_cache, url, fn _ ->
-          {:commit, parse_url(url)}
-        end)
-        |> set_ttl_based_on_image(url)
-      rescue
-        e ->
-          {:error, "Cachex error: #{inspect(e)}"}
+      with {:ok, data} <- get_cached_or_parse(url),
+           {:ok, _} <- set_ttl_based_on_image(data, url) do
+        {:ok, data}
+      else
+        error ->
+          Logger.error(fn -> "Rich media error: #{inspect(error)}" end)
+      end
+    end
+
+    defp get_cached_or_parse(url) do
+      case Cachex.fetch!(:rich_media_cache, url, fn _ -> {:commit, parse_url(url)} end) do
+        {:ok, _data} = res ->
+          res
+
+        {:error, _} = e ->
+          ttl = Pleroma.Config.get([:rich_media, :failure_backoff], 60_000)
+          Cachex.expire(:rich_media_cache, url, ttl)
+          e
       end
     end
   end
@@ -47,19 +61,26 @@ defmodule Pleroma.Web.RichMedia.Parser do
       config :pleroma, :rich_media,
         ttl_setters: [MyModule]
   """
-  def set_ttl_based_on_image({:ok, data}, url) do
-    with {:ok, nil} <- Cachex.ttl(:rich_media_cache, url),
-         ttl when is_number(ttl) <- get_ttl_from_image(data, url) do
-      Cachex.expire_at(:rich_media_cache, url, ttl * 1000)
-      {:ok, data}
-    else
+  @spec set_ttl_based_on_image(map(), String.t()) ::
+          {:ok, Integer.t() | :noop} | {:error, :no_key}
+  def set_ttl_based_on_image(data, url) do
+    case get_ttl_from_image(data, url) do
+      {:ok, ttl} when is_number(ttl) ->
+        ttl = ttl * 1000
+
+        case Cachex.expire_at(:rich_media_cache, url, ttl) do
+          {:ok, true} -> {:ok, ttl}
+          {:ok, false} -> {:error, :no_key}
+        end
+
       _ ->
-        {:ok, data}
+        {:ok, :noop}
     end
   end
 
   defp get_ttl_from_image(data, url) do
-    Pleroma.Config.get([:rich_media, :ttl_setters])
+    [:rich_media, :ttl_setters]
+    |> Pleroma.Config.get()
     |> Enum.reduce({:ok, nil}, fn
       module, {:ok, _ttl} ->
         module.ttl(data, url)
@@ -70,23 +91,16 @@ defmodule Pleroma.Web.RichMedia.Parser do
   end
 
   defp parse_url(url) do
-    try do
-      {:ok, %Tesla.Env{body: html}} = Pleroma.Web.RichMedia.Helpers.rich_media_get(url)
-
+    with {:ok, %Tesla.Env{body: html}} <- Pleroma.Web.RichMedia.Helpers.rich_media_get(url),
+         {:ok, html} <- Floki.parse_document(html) do
       html
-      |> parse_html()
       |> maybe_parse()
       |> Map.put("url", url)
       |> clean_parsed_data()
       |> check_parsed_data()
-    rescue
-      e ->
-        {:error, "Parsing error: #{inspect(e)} #{inspect(__STACKTRACE__)}"}
     end
   end
 
-  defp parse_html(html), do: Floki.parse_document!(html)
-
   defp maybe_parse(html) do
     Enum.reduce_while(parsers(), %{}, fn parser, acc ->
       case parser.parse(html, acc) do
diff --git a/lib/pleroma/web/rich_media/parsers/ttl/aws_signed_url.ex b/lib/pleroma/web/rich_media/parsers/ttl/aws_signed_url.ex
index 0dc1efdaf..c5aaea2d4 100644
--- a/lib/pleroma/web/rich_media/parsers/ttl/aws_signed_url.ex
+++ b/lib/pleroma/web/rich_media/parsers/ttl/aws_signed_url.ex
@@ -10,20 +10,15 @@ defmodule Pleroma.Web.RichMedia.Parser.TTL.AwsSignedUrl do
       |> parse_query_params()
       |> format_query_params()
       |> get_expiration_timestamp()
+    else
+      {:error, "Not aws signed url #{inspect(image)}"}
     end
   end
 
-  defp is_aws_signed_url(""), do: nil
-  defp is_aws_signed_url(nil), do: nil
-
-  defp is_aws_signed_url(image) when is_binary(image) do
+  defp is_aws_signed_url(image) when is_binary(image) and image != "" do
     %URI{host: host, query: query} = URI.parse(image)
 
-    if String.contains?(host, "amazonaws.com") and String.contains?(query, "X-Amz-Expires") do
-      image
-    else
-      nil
-    end
+    String.contains?(host, "amazonaws.com") and String.contains?(query, "X-Amz-Expires")
   end
 
   defp is_aws_signed_url(_), do: nil
@@ -46,6 +41,6 @@ defmodule Pleroma.Web.RichMedia.Parser.TTL.AwsSignedUrl do
       |> Map.get("X-Amz-Date")
       |> Timex.parse("{ISO:Basic:Z}")
 
-    Timex.to_unix(date) + String.to_integer(Map.get(params, "X-Amz-Expires"))
+    {:ok, Timex.to_unix(date) + String.to_integer(Map.get(params, "X-Amz-Expires"))}
   end
 end