# 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
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
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