2019-01-28 20:59:36 +01:00
|
|
|
# Pleroma: A lightweight social networking server
|
2020-03-02 06:08:45 +01:00
|
|
|
# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
|
2019-01-28 20:59:36 +01:00
|
|
|
# SPDX-License-Identifier: AGPL-3.0-only
|
|
|
|
|
2019-01-01 21:26:40 +01:00
|
|
|
defmodule Pleroma.Web.RichMedia.Parser do
|
2020-09-02 11:38:43 +02:00
|
|
|
require Logger
|
|
|
|
|
2019-07-11 15:04:42 +02:00
|
|
|
defp parsers do
|
|
|
|
Pleroma.Config.get([:rich_media, :parsers])
|
|
|
|
end
|
|
|
|
|
2019-01-26 17:26:11 +01:00
|
|
|
def parse(nil), do: {:error, "No URL provided"}
|
|
|
|
|
2019-06-06 22:59:51 +02:00
|
|
|
if Pleroma.Config.get(:env) == :test do
|
2020-09-02 11:38:43 +02:00
|
|
|
@spec parse(String.t()) :: {:ok, map()} | {:error, any()}
|
2019-01-05 00:50:54 +01:00
|
|
|
def parse(url), do: parse_url(url)
|
|
|
|
else
|
2020-09-02 11:38:43 +02:00
|
|
|
@spec parse(String.t()) :: {:ok, map()} | {:error, any()}
|
2019-01-26 17:26:11 +01:00
|
|
|
def parse(url) do
|
2020-09-02 15:46:11 +02:00
|
|
|
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
|
|
|
|
|
2020-09-14 14:38:00 +02:00
|
|
|
{:error, :body_too_large} = e ->
|
|
|
|
e
|
|
|
|
|
2020-09-14 14:48:27 +02:00
|
|
|
{:error, {:content_type, _}} = e ->
|
2020-09-14 14:38:00 +02:00
|
|
|
e
|
|
|
|
|
|
|
|
# The TTL is not set for the errors above, since they are unlikely to change
|
|
|
|
# with time
|
2020-09-02 15:46:11 +02:00
|
|
|
{:error, _} = e ->
|
|
|
|
ttl = Pleroma.Config.get([:rich_media, :failure_backoff], 60_000)
|
|
|
|
Cachex.expire(:rich_media_cache, url, ttl)
|
|
|
|
e
|
|
|
|
end
|
2019-01-26 17:26:11 +01:00
|
|
|
end
|
2019-01-05 00:50:54 +01:00
|
|
|
end
|
2019-01-01 21:26:40 +01:00
|
|
|
|
2019-07-16 18:52:36 +02:00
|
|
|
@doc """
|
|
|
|
Set the rich media cache based on the expiration time of image.
|
|
|
|
|
2019-07-19 07:58:42 +02:00
|
|
|
Adopt behaviour `Pleroma.Web.RichMedia.Parser.TTL`
|
2019-07-16 18:52:36 +02:00
|
|
|
|
|
|
|
## Example
|
|
|
|
|
|
|
|
defmodule MyModule do
|
2019-07-19 07:58:42 +02:00
|
|
|
@behaviour Pleroma.Web.RichMedia.Parser.TTL
|
|
|
|
def ttl(data, url) do
|
2019-07-16 18:52:36 +02:00
|
|
|
image_url = Map.get(data, :image)
|
|
|
|
# do some parsing in the url and get the ttl of the image
|
2019-07-19 07:58:42 +02:00
|
|
|
# and return ttl is unix time
|
|
|
|
parse_ttl_from_url(image_url)
|
2019-07-16 18:52:36 +02:00
|
|
|
end
|
|
|
|
end
|
|
|
|
|
|
|
|
Define the module in the config
|
|
|
|
|
|
|
|
config :pleroma, :rich_media,
|
|
|
|
ttl_setters: [MyModule]
|
|
|
|
"""
|
2020-09-02 15:46:11 +02:00
|
|
|
@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
|
|
|
|
|
2019-07-16 18:52:36 +02:00
|
|
|
_ ->
|
2020-09-02 15:46:11 +02:00
|
|
|
{:ok, :noop}
|
2019-07-16 18:52:36 +02:00
|
|
|
end
|
|
|
|
end
|
|
|
|
|
2019-07-19 07:58:42 +02:00
|
|
|
defp get_ttl_from_image(data, url) do
|
2020-09-02 11:38:43 +02:00
|
|
|
[:rich_media, :ttl_setters]
|
|
|
|
|> Pleroma.Config.get()
|
2019-07-19 07:58:42 +02:00
|
|
|
|> Enum.reduce({:ok, nil}, fn
|
|
|
|
module, {:ok, _ttl} ->
|
|
|
|
module.ttl(data, url)
|
|
|
|
|
|
|
|
_, error ->
|
|
|
|
error
|
|
|
|
end)
|
|
|
|
end
|
2019-07-16 18:52:36 +02:00
|
|
|
|
2019-01-05 00:50:54 +01:00
|
|
|
defp parse_url(url) do
|
2020-09-02 11:38:43 +02:00
|
|
|
with {:ok, %Tesla.Env{body: html}} <- Pleroma.Web.RichMedia.Helpers.rich_media_get(url),
|
|
|
|
{:ok, html} <- Floki.parse_document(html) do
|
2019-05-30 23:03:31 +02:00
|
|
|
html
|
|
|
|
|> maybe_parse()
|
2020-06-09 19:49:24 +02:00
|
|
|
|> Map.put("url", url)
|
2019-05-30 23:03:31 +02:00
|
|
|
|> clean_parsed_data()
|
|
|
|
|> check_parsed_data()
|
2019-01-27 13:21:05 +01:00
|
|
|
end
|
2019-01-02 15:02:50 +01:00
|
|
|
end
|
|
|
|
|
|
|
|
defp maybe_parse(html) do
|
2019-07-11 15:04:42 +02:00
|
|
|
Enum.reduce_while(parsers(), %{}, fn parser, acc ->
|
2019-01-01 21:26:40 +01:00
|
|
|
case parser.parse(html, acc) do
|
2020-06-11 15:57:31 +02:00
|
|
|
data when data != %{} -> {:halt, data}
|
|
|
|
_ -> {:cont, acc}
|
2019-01-01 21:26:40 +01:00
|
|
|
end
|
|
|
|
end)
|
|
|
|
end
|
2019-01-02 15:02:50 +01:00
|
|
|
|
2020-06-09 19:49:24 +02:00
|
|
|
defp check_parsed_data(%{"title" => title} = data)
|
|
|
|
when is_binary(title) and title != "" do
|
2019-01-28 21:31:43 +01:00
|
|
|
{:ok, data}
|
2019-01-02 15:02:50 +01:00
|
|
|
end
|
|
|
|
|
2019-01-31 17:03:56 +01:00
|
|
|
defp check_parsed_data(data) do
|
2019-01-28 21:31:43 +01:00
|
|
|
{:error, "Found metadata was invalid or incomplete: #{inspect(data)}"}
|
2019-01-02 15:02:50 +01:00
|
|
|
end
|
2019-01-31 17:03:56 +01:00
|
|
|
|
|
|
|
defp clean_parsed_data(data) do
|
|
|
|
data
|
2019-02-05 21:50:57 +01:00
|
|
|
|> Enum.reject(fn {key, val} ->
|
2020-06-09 19:49:24 +02:00
|
|
|
not match?({:ok, _}, Jason.encode(%{key => val}))
|
2019-01-31 17:03:56 +01:00
|
|
|
end)
|
|
|
|
|> Map.new()
|
|
|
|
end
|
2019-01-01 21:26:40 +01:00
|
|
|
end
|