X-Git-Url: http://git.squeep.com/?a=blobdiff_plain;f=lib%2Fpleroma%2Fweb%2Frich_media%2Fparser.ex;h=40980def8198d134496ed429a3efd494352c4c95;hb=0e23138b50f1fdd9ea78df31eec1b3caac905e2c;hp=e4595800c2f04afebbb40b381d5a28aa2eaf9b8a;hpb=73a7420c23308d19aa270ba6345a9b2b963abee6;p=akkoma
diff --git a/lib/pleroma/web/rich_media/parser.ex b/lib/pleroma/web/rich_media/parser.ex
index e4595800c..40980def8 100644
--- a/lib/pleroma/web/rich_media/parser.ex
+++ b/lib/pleroma/web/rich_media/parser.ex
@@ -1,24 +1,20 @@
# Pleroma: A lightweight social networking server
-# Copyright © 2017-2019 Pleroma Authors
+# Copyright © 2017-2020 Pleroma Authors
# SPDX-License-Identifier: AGPL-3.0-only
defmodule Pleroma.Web.RichMedia.Parser do
- @parsers [
- Pleroma.Web.RichMedia.Parsers.OGP,
- Pleroma.Web.RichMedia.Parsers.TwitterCard,
- Pleroma.Web.RichMedia.Parsers.OEmbed
- ]
-
- @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
+ Pleroma.Config.get([:rich_media, :parsers])
+ end
+
def parse(nil), do: {:error, "No URL provided"}
- if Mix.env() == :test do
+ if Pleroma.Config.get(:env) == :test do
def parse(url), do: parse_url(url)
else
def parse(url) do
@@ -26,6 +22,7 @@ defmodule Pleroma.Web.RichMedia.Parser 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)}"}
@@ -33,22 +30,80 @@ defmodule Pleroma.Web.RichMedia.Parser do
end
end
+ @doc """
+ Set the rich media cache based on the expiration time of image.
+
+ Adopt behaviour `Pleroma.Web.RichMedia.Parser.TTL`
+
+ ## Example
+
+ defmodule MyModule 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
+ # and return ttl is unix time
+ parse_ttl_from_url(image_url)
+ end
+ end
+
+ Define the module in the config
+
+ 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
+ _ ->
+ {:ok, data}
+ end
+ end
+
+ 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 ->
+ Enum.reduce_while(parsers(), %{}, fn parser, acc ->
case parser.parse(html, acc) do
{:ok, data} -> {:halt, data}
{:error, _msg} -> {:cont, acc}
@@ -56,7 +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