From bbdad8556861c60ae1f526f63de9c5857c4ad547 Mon Sep 17 00:00:00 2001 From: Ivan Tashkinov Date: Fri, 8 May 2020 23:06:47 +0300 Subject: [PATCH] Initial implementation of image preview proxy. Media proxy tests refactoring. --- config/config.exs | 5 + lib/pleroma/helpers/mogrify_helper.ex | 25 ++++ .../web/mastodon_api/views/status_view.ex | 3 +- lib/pleroma/web/media_proxy/media_proxy.ex | 53 ++++++- .../web/media_proxy/media_proxy_controller.ex | 78 ++++++++-- lib/pleroma/web/router.ex | 2 + test/web/media_proxy/media_proxy_test.exs | 133 +++++++----------- 7 files changed, 198 insertions(+), 101 deletions(-) create mode 100644 lib/pleroma/helpers/mogrify_helper.ex diff --git a/config/config.exs b/config/config.exs index e703c1632..526901f83 100644 --- a/config/config.exs +++ b/config/config.exs @@ -388,6 +388,11 @@ ], whitelist: [] +config :pleroma, :media_preview_proxy, + enabled: false, + limit_dimensions: "400x200", + max_body_length: 25 * 1_048_576 + config :pleroma, :chat, enabled: true config :phoenix, :format_encoders, json: Jason diff --git a/lib/pleroma/helpers/mogrify_helper.ex b/lib/pleroma/helpers/mogrify_helper.ex new file mode 100644 index 000000000..67edb35c3 --- /dev/null +++ b/lib/pleroma/helpers/mogrify_helper.ex @@ -0,0 +1,25 @@ +# Pleroma: A lightweight social networking server +# Copyright © 2017-2020 Pleroma Authors +# SPDX-License-Identifier: AGPL-3.0-only + +defmodule Pleroma.Helpers.MogrifyHelper do + @moduledoc """ + Handles common Mogrify operations. + """ + + @spec store_as_temporary_file(String.t(), binary()) :: {:ok, String.t()} | {:error, atom()} + @doc "Stores binary content fetched from specified URL as a temporary file." + def store_as_temporary_file(url, body) do + path = Mogrify.temporary_path_for(%{path: url}) + with :ok <- File.write(path, body), do: {:ok, path} + end + + @spec store_as_temporary_file(String.t(), String.t()) :: Mogrify.Image.t() | any() + @doc "Modifies file at specified path by resizing to specified limit dimensions." + def in_place_resize_to_limit(path, resize_dimensions) do + path + |> Mogrify.open() + |> Mogrify.resize_to_limit(resize_dimensions) + |> Mogrify.save(in_place: true) + end +end diff --git a/lib/pleroma/web/mastodon_api/views/status_view.ex b/lib/pleroma/web/mastodon_api/views/status_view.ex index 24167f66f..2a206f743 100644 --- a/lib/pleroma/web/mastodon_api/views/status_view.ex +++ b/lib/pleroma/web/mastodon_api/views/status_view.ex @@ -419,6 +419,7 @@ def render("attachment.json", %{attachment: attachment}) do [attachment_url | _] = attachment["url"] media_type = attachment_url["mediaType"] || attachment_url["mimeType"] || "image" href = attachment_url["href"] |> MediaProxy.url() + href_preview = attachment_url["href"] |> MediaProxy.preview_url() type = cond do @@ -434,7 +435,7 @@ def render("attachment.json", %{attachment: attachment}) do id: to_string(attachment["id"] || hash_id), url: href, remote_url: href, - preview_url: href, + preview_url: href_preview, text_url: href, type: type, description: attachment["name"], diff --git a/lib/pleroma/web/media_proxy/media_proxy.ex b/lib/pleroma/web/media_proxy/media_proxy.ex index b2b524524..f4791c758 100644 --- a/lib/pleroma/web/media_proxy/media_proxy.ex +++ b/lib/pleroma/web/media_proxy/media_proxy.ex @@ -20,6 +20,14 @@ def url(url) do end end + def preview_url(url) do + if disabled?() or whitelisted?(url) do + url + else + encode_preview_url(url) + end + end + defp disabled?, do: !Config.get([:media_proxy, :enabled], false) defp local?(url), do: String.starts_with?(url, Pleroma.Web.base_url()) @@ -43,17 +51,29 @@ defp whitelisted?(url) do end) end - def encode_url(url) do + defp base64_sig64(url) do base64 = Base.url_encode64(url, @base64_opts) sig64 = base64 - |> signed_url + |> signed_url() |> Base.url_encode64(@base64_opts) + {base64, sig64} + end + + def encode_url(url) do + {base64, sig64} = base64_sig64(url) + build_url(sig64, base64, filename(url)) end + def encode_preview_url(url) do + {base64, sig64} = base64_sig64(url) + + build_preview_url(sig64, base64, filename(url)) + end + def decode_url(sig, url) do with {:ok, sig} <- Base.url_decode64(sig, @base64_opts), signature when signature == sig <- signed_url(url) do @@ -71,10 +91,10 @@ def filename(url_or_path) do if path = URI.parse(url_or_path).path, do: Path.basename(path) end - def build_url(sig_base64, url_base64, filename \\ nil) do + defp proxy_url(path, sig_base64, url_base64, filename) do [ Pleroma.Config.get([:media_proxy, :base_url], Web.base_url()), - "proxy", + path, sig_base64, url_base64, filename @@ -82,4 +102,29 @@ def build_url(sig_base64, url_base64, filename \\ nil) do |> Enum.filter(& &1) |> Path.join() end + + def build_url(sig_base64, url_base64, filename \\ nil) do + proxy_url("proxy", sig_base64, url_base64, filename) + end + + def build_preview_url(sig_base64, url_base64, filename \\ nil) do + proxy_url("proxy/preview", sig_base64, url_base64, filename) + end + + def filename_matches(%{"filename" => _} = _, path, url) do + filename = filename(url) + + if filename && not basename_matches?(path, filename) do + {:wrong_filename, filename} + else + :ok + end + end + + def filename_matches(_, _, _), do: :ok + + defp basename_matches?(path, filename) do + basename = Path.basename(path) + basename == filename or URI.decode(basename) == filename or URI.encode(basename) == filename + end end diff --git a/lib/pleroma/web/media_proxy/media_proxy_controller.ex b/lib/pleroma/web/media_proxy/media_proxy_controller.ex index 4657a4383..fe3f61c18 100644 --- a/lib/pleroma/web/media_proxy/media_proxy_controller.ex +++ b/lib/pleroma/web/media_proxy/media_proxy_controller.ex @@ -5,19 +5,21 @@ defmodule Pleroma.Web.MediaProxy.MediaProxyController do use Pleroma.Web, :controller + alias Pleroma.Config + alias Pleroma.Helpers.MogrifyHelper alias Pleroma.ReverseProxy alias Pleroma.Web.MediaProxy @default_proxy_opts [max_body_length: 25 * 1_048_576, http: [follow_redirect: true]] def remote(conn, %{"sig" => sig64, "url" => url64} = params) do - with config <- Pleroma.Config.get([:media_proxy], []), - true <- Keyword.get(config, :enabled, false), + with config <- Config.get([:media_proxy], []), + {_, true} <- {:enabled, Keyword.get(config, :enabled, false)}, {:ok, url} <- MediaProxy.decode_url(sig64, url64), - :ok <- filename_matches(params, conn.request_path, url) do + :ok <- MediaProxy.filename_matches(params, conn.request_path, url) do ReverseProxy.call(conn, url, Keyword.get(config, :proxy_opts, @default_proxy_opts)) else - false -> + {:enabled, false} -> send_resp(conn, 404, Plug.Conn.Status.reason_phrase(404)) {:error, :invalid_signature} -> @@ -28,20 +30,68 @@ def remote(conn, %{"sig" => sig64, "url" => url64} = params) do end end - def filename_matches(%{"filename" => _} = _, path, url) do - filename = MediaProxy.filename(url) - - if filename && does_not_match(path, filename) do - {:wrong_filename, filename} + def preview(conn, %{"sig" => sig64, "url" => url64} = params) do + with {_, true} <- {:enabled, Config.get([:media_preview_proxy, :enabled], false)}, + {:ok, url} <- MediaProxy.decode_url(sig64, url64), + :ok <- MediaProxy.filename_matches(params, conn.request_path, url) do + handle_preview(conn, url) else - :ok + {:enabled, false} -> + send_resp(conn, 404, Plug.Conn.Status.reason_phrase(404)) + + {:error, :invalid_signature} -> + send_resp(conn, 403, Plug.Conn.Status.reason_phrase(403)) + + {:wrong_filename, filename} -> + redirect(conn, external: MediaProxy.build_preview_url(sig64, url64, filename)) end end - def filename_matches(_, _, _), do: :ok + defp handle_preview(conn, url) do + with {:ok, %{status: status} = head_response} when status in 200..299 <- Tesla.head(url), + {_, true} <- {:acceptable_content_length, acceptable_body_length?(head_response)} do + content_type = Tesla.get_header(head_response, "content-type") + handle_preview(content_type, conn, url) + else + {_, %{status: status}} -> + send_resp(conn, :failed_dependency, "Can't fetch HTTP headers (HTTP #{status}).") - defp does_not_match(path, filename) do - basename = Path.basename(path) - basename != filename and URI.decode(basename) != filename and URI.encode(basename) != filename + {:acceptable_content_length, false} -> + send_resp(conn, :unprocessable_entity, "Source file size exceeds limit.") + end + end + + defp handle_preview("image/" <> _, %{params: params} = conn, url) do + with {:ok, %{status: status, body: body}} when status in 200..299 <- Tesla.get(url), + {:ok, path} <- MogrifyHelper.store_as_temporary_file(url, body), + resize_dimensions <- + Map.get( + params, + "limit_dimensions", + Config.get([:media_preview_proxy, :limit_dimensions]) + ), + %Mogrify.Image{} <- MogrifyHelper.in_place_resize_to_limit(path, resize_dimensions) do + send_file(conn, 200, path) + else + {_, %{status: _}} -> + send_resp(conn, :failed_dependency, "Can't fetch the image.") + + _ -> + send_resp(conn, :failed_dependency, "Can't handle image preview.") + end + end + + defp handle_preview(content_type, conn, _url) do + send_resp(conn, :unprocessable_entity, "Unsupported content type: #{content_type}.") + end + + defp acceptable_body_length?(head_response) do + max_body_length = Config.get([:media_preview_proxy, :max_body_length], nil) + content_length = Tesla.get_header(head_response, "content-length") + content_length = with {int, _} <- Integer.parse(content_length), do: int + + content_length == :error or + max_body_length in [nil, :infinity] or + content_length <= max_body_length end end diff --git a/lib/pleroma/web/router.ex b/lib/pleroma/web/router.ex index 7a171f9fb..6fb47029a 100644 --- a/lib/pleroma/web/router.ex +++ b/lib/pleroma/web/router.ex @@ -663,6 +663,8 @@ defmodule Pleroma.Web.Router do end scope "/proxy/", Pleroma.Web.MediaProxy do + get("/preview/:sig/:url", MediaProxyController, :preview) + get("/preview/:sig/:url/:filename", MediaProxyController, :preview) get("/:sig/:url", MediaProxyController, :remote) get("/:sig/:url/:filename", MediaProxyController, :remote) end diff --git a/test/web/media_proxy/media_proxy_test.exs b/test/web/media_proxy/media_proxy_test.exs index 69c2d5dae..cad0acd30 100644 --- a/test/web/media_proxy/media_proxy_test.exs +++ b/test/web/media_proxy/media_proxy_test.exs @@ -5,42 +5,44 @@ defmodule Pleroma.Web.MediaProxyTest do use ExUnit.Case use Pleroma.Tests.Helpers - import Pleroma.Web.MediaProxy - alias Pleroma.Web.MediaProxy.MediaProxyController - setup do: clear_config([:media_proxy, :enabled]) - setup do: clear_config(Pleroma.Upload) + alias Pleroma.Config + alias Pleroma.Web.Endpoint + alias Pleroma.Web.MediaProxy + + defp decode_result(encoded) do + [_, "proxy", sig, base64 | _] = URI.parse(encoded).path |> String.split("/") + {:ok, decoded} = MediaProxy.decode_url(sig, base64) + decoded + end describe "when enabled" do - setup do - Pleroma.Config.put([:media_proxy, :enabled], true) - :ok - end + setup do: clear_config([:media_proxy, :enabled], true) test "ignores invalid url" do - assert url(nil) == nil - assert url("") == nil + assert MediaProxy.url(nil) == nil + assert MediaProxy.url("") == nil end test "ignores relative url" do - assert url("/local") == "/local" - assert url("/") == "/" + assert MediaProxy.url("/local") == "/local" + assert MediaProxy.url("/") == "/" end test "ignores local url" do - local_url = Pleroma.Web.Endpoint.url() <> "/hello" - local_root = Pleroma.Web.Endpoint.url() - assert url(local_url) == local_url - assert url(local_root) == local_root + local_url = Endpoint.url() <> "/hello" + local_root = Endpoint.url() + assert MediaProxy.url(local_url) == local_url + assert MediaProxy.url(local_root) == local_root end test "encodes and decodes URL" do url = "https://pleroma.soykaf.com/static/logo.png" - encoded = url(url) + encoded = MediaProxy.url(url) assert String.starts_with?( encoded, - Pleroma.Config.get([:media_proxy, :base_url], Pleroma.Web.base_url()) + Config.get([:media_proxy, :base_url], Pleroma.Web.base_url()) ) assert String.ends_with?(encoded, "/logo.png") @@ -50,62 +52,59 @@ test "encodes and decodes URL" do test "encodes and decodes URL without a path" do url = "https://pleroma.soykaf.com" - encoded = url(url) + encoded = MediaProxy.url(url) assert decode_result(encoded) == url end test "encodes and decodes URL without an extension" do url = "https://pleroma.soykaf.com/path/" - encoded = url(url) + encoded = MediaProxy.url(url) assert String.ends_with?(encoded, "/path") assert decode_result(encoded) == url end test "encodes and decodes URL and ignores query params for the path" do url = "https://pleroma.soykaf.com/static/logo.png?93939393939&bunny=true" - encoded = url(url) + encoded = MediaProxy.url(url) assert String.ends_with?(encoded, "/logo.png") assert decode_result(encoded) == url end test "validates signature" do - secret_key_base = Pleroma.Config.get([Pleroma.Web.Endpoint, :secret_key_base]) + secret_key_base = Config.get([Endpoint, :secret_key_base]) + clear_config([Endpoint, :secret_key_base], secret_key_base) - on_exit(fn -> - Pleroma.Config.put([Pleroma.Web.Endpoint, :secret_key_base], secret_key_base) - end) + encoded = MediaProxy.url("https://pleroma.social") - encoded = url("https://pleroma.social") - - Pleroma.Config.put( - [Pleroma.Web.Endpoint, :secret_key_base], + Config.put( + [Endpoint, :secret_key_base], "00000000000000000000000000000000000000000000000" ) [_, "proxy", sig, base64 | _] = URI.parse(encoded).path |> String.split("/") - assert decode_url(sig, base64) == {:error, :invalid_signature} + assert MediaProxy.decode_url(sig, base64) == {:error, :invalid_signature} end - test "filename_matches preserves the encoded or decoded path" do - assert MediaProxyController.filename_matches( + test "`filename_matches/_` preserves the encoded or decoded path" do + assert MediaProxy.filename_matches( %{"filename" => "/Hello world.jpg"}, "/Hello world.jpg", "http://pleroma.social/Hello world.jpg" ) == :ok - assert MediaProxyController.filename_matches( + assert MediaProxy.filename_matches( %{"filename" => "/Hello%20world.jpg"}, "/Hello%20world.jpg", "http://pleroma.social/Hello%20world.jpg" ) == :ok - assert MediaProxyController.filename_matches( + assert MediaProxy.filename_matches( %{"filename" => "/my%2Flong%2Furl%2F2019%2F07%2FS.jpg"}, "/my%2Flong%2Furl%2F2019%2F07%2FS.jpg", "http://pleroma.social/my%2Flong%2Furl%2F2019%2F07%2FS.jpg" ) == :ok - assert MediaProxyController.filename_matches( + assert MediaProxy.filename_matches( %{"filename" => "/my%2Flong%2Furl%2F2019%2F07%2FS.jp"}, "/my%2Flong%2Furl%2F2019%2F07%2FS.jp", "http://pleroma.social/my%2Flong%2Furl%2F2019%2F07%2FS.jpg" @@ -116,7 +115,7 @@ test "encoded url are tried to match for proxy as `conn.request_path` encodes th # conn.request_path will return encoded url request_path = "/ANALYSE-DAI-_-LE-STABLECOIN-100-D%C3%89CENTRALIS%C3%89-BQ.jpg" - assert MediaProxyController.filename_matches( + assert MediaProxy.filename_matches( true, request_path, "https://mydomain.com/uploads/2019/07/ANALYSE-DAI-_-LE-STABLECOIN-100-DÉCENTRALISÉ-BQ.jpg" @@ -124,20 +123,12 @@ test "encoded url are tried to match for proxy as `conn.request_path` encodes th end test "uses the configured base_url" do - base_url = Pleroma.Config.get([:media_proxy, :base_url]) - - if base_url do - on_exit(fn -> - Pleroma.Config.put([:media_proxy, :base_url], base_url) - end) - end - - Pleroma.Config.put([:media_proxy, :base_url], "https://cache.pleroma.social") + clear_config([:media_proxy, :base_url], "https://cache.pleroma.social") url = "https://pleroma.soykaf.com/static/logo.png" - encoded = url(url) + encoded = MediaProxy.url(url) - assert String.starts_with?(encoded, Pleroma.Config.get([:media_proxy, :base_url])) + assert String.starts_with?(encoded, Config.get([:media_proxy, :base_url])) end # Some sites expect ASCII encoded characters in the URL to be preserved even if @@ -148,7 +139,7 @@ test "preserve ASCII encoding" do url = "https://pleroma.com/%20/%21/%22/%23/%24/%25/%26/%27/%28/%29/%2A/%2B/%2C/%2D/%2E/%2F/%30/%31/%32/%33/%34/%35/%36/%37/%38/%39/%3A/%3B/%3C/%3D/%3E/%3F/%40/%41/%42/%43/%44/%45/%46/%47/%48/%49/%4A/%4B/%4C/%4D/%4E/%4F/%50/%51/%52/%53/%54/%55/%56/%57/%58/%59/%5A/%5B/%5C/%5D/%5E/%5F/%60/%61/%62/%63/%64/%65/%66/%67/%68/%69/%6A/%6B/%6C/%6D/%6E/%6F/%70/%71/%72/%73/%74/%75/%76/%77/%78/%79/%7A/%7B/%7C/%7D/%7E/%7F/%80/%81/%82/%83/%84/%85/%86/%87/%88/%89/%8A/%8B/%8C/%8D/%8E/%8F/%90/%91/%92/%93/%94/%95/%96/%97/%98/%99/%9A/%9B/%9C/%9D/%9E/%9F/%C2%A0/%A1/%A2/%A3/%A4/%A5/%A6/%A7/%A8/%A9/%AA/%AB/%AC/%C2%AD/%AE/%AF/%B0/%B1/%B2/%B3/%B4/%B5/%B6/%B7/%B8/%B9/%BA/%BB/%BC/%BD/%BE/%BF/%C0/%C1/%C2/%C3/%C4/%C5/%C6/%C7/%C8/%C9/%CA/%CB/%CC/%CD/%CE/%CF/%D0/%D1/%D2/%D3/%D4/%D5/%D6/%D7/%D8/%D9/%DA/%DB/%DC/%DD/%DE/%DF/%E0/%E1/%E2/%E3/%E4/%E5/%E6/%E7/%E8/%E9/%EA/%EB/%EC/%ED/%EE/%EF/%F0/%F1/%F2/%F3/%F4/%F5/%F6/%F7/%F8/%F9/%FA/%FB/%FC/%FD/%FE/%FF" - encoded = url(url) + encoded = MediaProxy.url(url) assert decode_result(encoded) == url end @@ -159,77 +150,55 @@ test "preserve non-unicode characters per RFC3986" do url = "https://pleroma.com/ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz1234567890-._~:/?#[]@!$&'()*+,;=|^`{}" - encoded = url(url) + encoded = MediaProxy.url(url) assert decode_result(encoded) == url end test "preserve unicode characters" do url = "https://ko.wikipedia.org/wiki/위키백과:대문" - encoded = url(url) + encoded = MediaProxy.url(url) assert decode_result(encoded) == url end end describe "when disabled" do - setup do - enabled = Pleroma.Config.get([:media_proxy, :enabled]) - - if enabled do - Pleroma.Config.put([:media_proxy, :enabled], false) - - on_exit(fn -> - Pleroma.Config.put([:media_proxy, :enabled], enabled) - :ok - end) - end - - :ok - end + setup do: clear_config([:media_proxy, :enabled], false) test "does not encode remote urls" do - assert url("https://google.fr") == "https://google.fr" + assert MediaProxy.url("https://google.fr") == "https://google.fr" end end - defp decode_result(encoded) do - [_, "proxy", sig, base64 | _] = URI.parse(encoded).path |> String.split("/") - {:ok, decoded} = decode_url(sig, base64) - decoded - end - describe "whitelist" do - setup do - Pleroma.Config.put([:media_proxy, :enabled], true) - :ok - end + setup do: clear_config([:media_proxy, :enabled], true) test "mediaproxy whitelist" do - Pleroma.Config.put([:media_proxy, :whitelist], ["google.com", "feld.me"]) + clear_config([:media_proxy, :whitelist], ["google.com", "feld.me"]) url = "https://feld.me/foo.png" - unencoded = url(url) + unencoded = MediaProxy.url(url) assert unencoded == url end test "does not change whitelisted urls" do - Pleroma.Config.put([:media_proxy, :whitelist], ["mycdn.akamai.com"]) - Pleroma.Config.put([:media_proxy, :base_url], "https://cache.pleroma.social") + clear_config([:media_proxy, :whitelist], ["mycdn.akamai.com"]) + clear_config([:media_proxy, :base_url], "https://cache.pleroma.social") media_url = "https://mycdn.akamai.com" url = "#{media_url}/static/logo.png" - encoded = url(url) + encoded = MediaProxy.url(url) assert String.starts_with?(encoded, media_url) end test "ensure Pleroma.Upload base_url is always whitelisted" do media_url = "https://media.pleroma.social" - Pleroma.Config.put([Pleroma.Upload, :base_url], media_url) + clear_config([Pleroma.Upload, :base_url], media_url) url = "#{media_url}/static/logo.png" - encoded = url(url) + encoded = MediaProxy.url(url) assert String.starts_with?(encoded, media_url) end