X-Git-Url: https://git.squeep.com/?a=blobdiff_plain;f=lib%2Fpleroma%2Fweb%2Frich_media%2Fparser.ex;h=c06b0a0f2668338415d592fde838bff42e49e056;hb=179fa32dd5d7b4cf1b338fc3df4840ae33525606;hp=e67ecc47d052ef34fc3f1175adadf3087d507975;hpb=4d3655c47959d7b9bfb3eb7316a8b7ad9010288e;p=akkoma diff --git a/lib/pleroma/web/rich_media/parser.ex b/lib/pleroma/web/rich_media/parser.ex index e67ecc47d..c06b0a0f2 100644 --- a/lib/pleroma/web/rich_media/parser.ex +++ b/lib/pleroma/web/rich_media/parser.ex @@ -3,15 +3,20 @@ # 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 [ + pool: :media, + recv_timeout: 2_000, + max_body: 2_000_000, + with_body: true ] + 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 @@ -19,6 +24,7 @@ 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)}"} @@ -26,19 +32,70 @@ 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 try do - {:ok, %Tesla.Env{body: html}} = Pleroma.HTTP.get(url, [], adapter: [pool: :media]) + {:ok, %Tesla.Env{body: html}} = Pleroma.HTTP.get(url, [], adapter: @hackney_options) - html |> maybe_parse() |> get_parsed_data() + html + |> parse_html + |> maybe_parse() + |> Map.put(:url, url) + |> clean_parsed_data() + |> check_parsed_data() rescue e -> {:error, "Parsing error: #{inspect(e)}"} end end + defp parse_html(html), do: Floki.parse(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} @@ -46,11 +103,24 @@ defmodule Pleroma.Web.RichMedia.Parser do end) end - defp get_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 - defp get_parsed_data(data) do + defp check_parsed_data(data) do {:error, "Found metadata was invalid or incomplete: #{inspect(data)}"} end + + defp clean_parsed_data(data) do + data + |> Enum.reject(fn {key, val} -> + with {:ok, _} <- Jason.encode(%{key => val}) do + false + else + _ -> true + end + end) + |> Map.new() + end end