Merge branch 'develop' into gun

[akkoma] / lib / pleroma / web / rich_media / parser.ex

diff --git a/lib/pleroma/web/rich_media/parser.ex b/lib/pleroma/web/rich_media/parser.ex
index ba8dc6f2a4b0812a9d3a715ce36deefdd921b215..40980def8198d134496ed429a3efd494352c4c95 100644 (file)
--- a/lib/pleroma/web/rich_media/parser.ex
+++ b/lib/pleroma/web/rich_media/parser.ex
@@ -1,13 +1,11 @@
 # Pleroma: A lightweight social networking server
-# Copyright © 2017-2019 Pleroma Authors <https://pleroma.social/>
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
 # SPDX-License-Identifier: AGPL-3.0-only
 
 defmodule Pleroma.Web.RichMedia.Parser do
-  @hackney_options [
+  @options [
     pool: :media,
-    recv_timeout: 2_000,
-    max_body: 2_000_000,
-    with_body: true
+    max_body: 2_000_000
   ]
 
   defp parsers do
@@ -35,17 +33,17 @@ defmodule Pleroma.Web.RichMedia.Parser do
   @doc """
   Set the rich media cache based on the expiration time of image.
 
-  Define a module that has `run` function
+  Adopt behaviour `Pleroma.Web.RichMedia.Parser.TTL`
 
   ## Example
 
       defmodule MyModule do
-        def run(data, url) do
+        @behaviour Pleroma.Web.RichMedia.Parser.TTL
+        def ttl(data, url) do
           image_url = Map.get(data, :image)
           # do some parsing in the url and get the ttl of the image
-          # ttl is unix time
-          ttl = parse_ttl_from_url(image_url)  
-          Cachex.expire_at(:rich_media_cache, url, ttl * 1000)
+          # and return ttl is unix time
+          parse_ttl_from_url(image_url)
         end
       end
 
@@ -55,37 +53,55 @@ defmodule Pleroma.Web.RichMedia.Parser do
         ttl_setters: [MyModule]
   """
   def set_ttl_based_on_image({:ok, data}, url) do
-    case Cachex.ttl(:rich_media_cache, url) do
-      {:ok, nil} ->
-        modules = Pleroma.Config.get([:rich_media, :ttl_setters])
-
-        if Enum.count(modules) > 0 do
-          Enum.each(modules, & &1.run(data, url))
-        end
-
-        {:ok, data}
-
+    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
       _ ->
         {:ok, data}
     end
   end
 
-  def set_ttl_based_on_image(data, _url), do: data
+  defp get_ttl_from_image(data, url) do
+    Pleroma.Config.get([:rich_media, :ttl_setters])
+    |> Enum.reduce({:ok, nil}, fn
+      module, {:ok, _ttl} ->
+        module.ttl(data, url)
+
+      _, error ->
+        error
+    end)
+  end
 
   defp parse_url(url) do
+    opts =
+      if Application.get_env(:tesla, :adapter) == Tesla.Adapter.Hackney do
+        Keyword.merge(@options,
+          recv_timeout: 2_000,
+          with_body: true
+        )
+      else
+        @options
+      end
+
     try do
-      {:ok, %Tesla.Env{body: html}} = Pleroma.HTTP.get(url, [], adapter: @hackney_options)
+      {:ok, %Tesla.Env{body: html}} = Pleroma.HTTP.get(url, [], adapter: opts)
 
       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_document!(html)
+
   defp maybe_parse(html) do
     Enum.reduce_while(parsers(), %{}, fn parser, acc ->
       case parser.parse(html, acc) do
@@ -95,7 +111,8 @@ defmodule Pleroma.Web.RichMedia.Parser do
     end)
   end
 
-  defp check_parsed_data(%{title: title} = data) when is_binary(title) and byte_size(title) > 0 do
+  defp check_parsed_data(%{title: title} = data)
+       when is_binary(title) and byte_size(title) > 0 do
     {:ok, data}
   end