OEmbed support for PreviewCard (#2337)

* OEmbed support for PreviewCard

* Improve ProviderDiscovery code failure treatment

* Do not crawl links if there is a content warning, since those
don't display a link card anyway

* Reset db schema

* Fresh migrate

* Fix rubocop style issues
Fix #1681 - return existing access token when applicable instead of creating new

* Fix test

* Extract http client to helper

* Improve oembed controller
This commit is contained in:
Eugen Rochko 2017-04-27 14:42:22 +02:00 committed by GitHub
parent be0a01145b
commit 88725d6ce8
22 changed files with 278 additions and 36 deletions

View file

@ -49,6 +49,7 @@ gem 'rails-settings-cached'
gem 'redis', '~>3.2', require: ['redis', 'redis/connection/hiredis'] gem 'redis', '~>3.2', require: ['redis', 'redis/connection/hiredis']
gem 'rqrcode' gem 'rqrcode'
gem 'ruby-oembed', require: 'oembed' gem 'ruby-oembed', require: 'oembed'
gem 'sanitize'
gem 'sidekiq' gem 'sidekiq'
gem 'sidekiq-unique-jobs' gem 'sidekiq-unique-jobs'
gem 'simple-navigation' gem 'simple-navigation'

View file

@ -123,6 +123,7 @@ GEM
connection_pool (2.2.1) connection_pool (2.2.1)
crack (0.4.3) crack (0.4.3)
safe_yaml (~> 1.0.0) safe_yaml (~> 1.0.0)
crass (1.0.2)
debug_inspector (0.0.2) debug_inspector (0.0.2)
devise (4.2.1) devise (4.2.1)
bcrypt (~> 3.0) bcrypt (~> 3.0)
@ -258,6 +259,8 @@ GEM
nio4r (2.0.0) nio4r (2.0.0)
nokogiri (1.7.1) nokogiri (1.7.1)
mini_portile2 (~> 2.1.0) mini_portile2 (~> 2.1.0)
nokogumbo (1.4.10)
nokogiri
oj (2.18.5) oj (2.18.5)
openssl (2.0.3) openssl (2.0.3)
orm_adapter (0.5.0) orm_adapter (0.5.0)
@ -398,6 +401,10 @@ GEM
ruby-oembed (0.12.0) ruby-oembed (0.12.0)
ruby-progressbar (1.8.1) ruby-progressbar (1.8.1)
safe_yaml (1.0.4) safe_yaml (1.0.4)
sanitize (4.4.0)
crass (~> 1.0.2)
nokogiri (>= 1.4.4)
nokogumbo (~> 1.4.1)
sass (3.4.23) sass (3.4.23)
sass-rails (5.0.6) sass-rails (5.0.6)
railties (>= 4.0.0, < 6) railties (>= 4.0.0, < 6)
@ -540,6 +547,7 @@ DEPENDENCIES
rspec-sidekiq rspec-sidekiq
rubocop rubocop
ruby-oembed ruby-oembed
sanitize
sass-rails (~> 5.0) sass-rails (~> 5.0)
sidekiq sidekiq
sidekiq-unique-jobs sidekiq-unique-jobs

View file

@ -13,7 +13,7 @@ export function fetchStatusCard(id) {
dispatch(fetchStatusCardRequest(id)); dispatch(fetchStatusCardRequest(id));
api(getState).get(`/api/v1/statuses/${id}/card`).then(response => { api(getState).get(`/api/v1/statuses/${id}/card`).then(response => {
if (!response.data.url || !response.data.title || !response.data.description) { if (!response.data.url) {
return; return;
} }

View file

@ -14,14 +14,11 @@ const getHostname = url => {
class Card extends React.PureComponent { class Card extends React.PureComponent {
render () { renderLink () {
const { card } = this.props; const { card } = this.props;
if (card === null) {
return null;
}
let image = ''; let image = '';
let provider = card.get('provider_name');
if (card.get('image')) { if (card.get('image')) {
image = ( image = (
@ -31,18 +28,64 @@ class Card extends React.PureComponent {
); );
} }
if (provider.length < 1) {
provider = getHostname(card.get('url'))
}
return ( return (
<a href={card.get('url')} className='status-card' target='_blank' rel='noopener'> <a href={card.get('url')} className='status-card' target='_blank' rel='noopener'>
{image} {image}
<div className='status-card__content'> <div className='status-card__content'>
<strong className='status-card__title' title={card.get('title')}>{card.get('title')}</strong> <strong className='status-card__title' title={card.get('title')}>{card.get('title')}</strong>
<p className='status-card__description'>{card.get('description').substring(0, 50)}</p> <p className='status-card__description'>{(card.get('description') || '').substring(0, 50)}</p>
<span className='status-card__host' style={hostStyle}>{getHostname(card.get('url'))}</span> <span className='status-card__host' style={hostStyle}>{provider}</span>
</div> </div>
</a> </a>
); );
} }
renderPhoto () {
const { card } = this.props;
return (
<a href={card.get('url')} className='status-card-photo' target='_blank' rel='noopener'>
<img src={card.get('url')} alt={card.get('title')} width={card.get('width')} height={card.get('height')} />
</a>
);
}
renderVideo () {
const { card } = this.props;
const content = { __html: card.get('html') };
return (
<div
className='status-card-video'
dangerouslySetInnerHTML={content}
/>
);
}
render () {
const { card } = this.props;
if (card === null) {
return null;
}
switch(card.get('type')) {
case 'link':
return this.renderLink();
case 'photo':
return this.renderPhoto();
case 'video':
return this.renderVideo();
case 'rich':
default:
return null;
}
}
} }
Card.propTypes = { Card.propTypes = {

View file

@ -1734,6 +1734,28 @@ button.icon-button.active i.fa-retweet {
} }
} }
.status-card-video, .status-card-rich, .status-card-photo {
margin-top: 14px;
overflow: hidden;
iframe {
width: 100%;
height: auto;
}
}
.status-card-photo {
display: block;
text-decoration: none;
img {
display: block;
width: 100%;
height: auto;
margin: 0;
}
}
.status-card__title { .status-card__title {
display: block; display: block;
font-weight: 500; font-weight: 500;

View file

@ -14,8 +14,20 @@ class Api::OEmbedController < ApiController
def stream_entry_from_url(url) def stream_entry_from_url(url)
params = Rails.application.routes.recognize_path(url) params = Rails.application.routes.recognize_path(url)
raise ActiveRecord::RecordNotFound unless params[:controller] == 'stream_entries' && params[:action] == 'show' raise ActiveRecord::RecordNotFound unless recognized_stream_entry_url?(params)
stream_entry(params)
end
def recognized_stream_entry_url?(params)
%w(stream_entries statuses).include?(params[:controller]) && params[:action] == 'show'
end
def stream_entry(params)
if params[:controller] == 'stream_entries'
StreamEntry.find(params[:id]) StreamEntry.find(params[:id])
else
Status.find(params[:id]).stream_entry
end
end end
end end

View file

@ -0,0 +1,13 @@
# frozen_string_literal: true
module HttpHelper
USER_AGENT = "#{HTTP::Request::USER_AGENT} (Mastodon/#{Mastodon::VERSION}; +http://#{Rails.configuration.x.local_domain}/)"
def http_client(options = {})
timeout = { write: 10, connect: 10, read: 10 }.merge(options)
HTTP.headers(user_agent: USER_AGENT)
.timeout(:per_operation, timeout)
.follow
end
end

View file

@ -1,13 +1,13 @@
# frozen_string_literal: true # frozen_string_literal: true
require 'singleton' require 'singleton'
require_relative './sanitize_config'
class Formatter class Formatter
include Singleton include Singleton
include RoutingHelper include RoutingHelper
include ActionView::Helpers::TextHelper include ActionView::Helpers::TextHelper
include ActionView::Helpers::SanitizeHelper
def format(status) def format(status)
return reformat(status.content) unless status.local? return reformat(status.content) unless status.local?
@ -23,7 +23,7 @@ class Formatter
end end
def reformat(html) def reformat(html)
sanitize(html, tags: %w(a br p span), attributes: %w(href rel class)) sanitize(html, Sanitize::Config::MASTODON_STRICT)
end end
def plaintext(status) def plaintext(status)
@ -43,6 +43,10 @@ class Formatter
html.html_safe # rubocop:disable Rails/OutputSafety html.html_safe # rubocop:disable Rails/OutputSafety
end end
def sanitize(html, config)
Sanitize.fragment(html, config)
end
private private
def encode(html) def encode(html)

View file

@ -0,0 +1,36 @@
# frozen_string_literal: true
class ProviderDiscovery < OEmbed::ProviderDiscovery
include HttpHelper
class << self
def discover_provider(url, options = {})
res = http_client.get(url)
format = options[:format]
raise OEmbed::NotFound, url if res.code != 200 || res.mime_type != 'text/html'
html = Nokogiri::HTML(res.to_s)
if format.nil? || format == :json
provider_endpoint ||= html.at_xpath('//link[@type="application/json+oembed"]')&.attribute('href')&.value
format ||= :json if provider_endpoint
end
if format.nil? || format == :xml
provider_endpoint ||= html.at_xpath('//link[@type="application/xml+oembed"]')&.attribute('href')&.value
format ||= :xml if provider_endpoint
end
begin
provider_endpoint = Addressable::URI.parse(provider_endpoint)
provider_endpoint.query = nil
provider_endpoint = provider_endpoint.to_s
rescue Addressable::URI::InvalidURIError
raise OEmbed::NotFound, url
end
OEmbed::Provider.new(provider_endpoint, format || OEmbed::Formatter.default)
end
end
end

View file

@ -0,0 +1,42 @@
# frozen_string_literal: true
class Sanitize
module Config
HTTP_PROTOCOLS ||= ['http', 'https', :relative].freeze
MASTODON_STRICT ||= freeze_config(
elements: %w(p br span a),
attributes: {
'a' => %w(href),
'span' => %w(class),
},
protocols: {
'a' => { 'href' => HTTP_PROTOCOLS },
}
)
MASTODON_OEMBED ||= freeze_config merge(
RELAXED,
elements: RELAXED[:elements] + %w(audio embed iframe source video),
attributes: merge(
RELAXED[:attributes],
'audio' => %w(controls),
'embed' => %w(height src type width),
'iframe' => %w(allowfullscreen frameborder height scrolling src width),
'source' => %w(src type),
'video' => %w(controls height loop width),
'div' => [:data]
),
protocols: merge(
RELAXED[:protocols],
'embed' => { 'src' => HTTP_PROTOCOLS },
'iframe' => { 'src' => HTTP_PROTOCOLS },
'source' => { 'src' => HTTP_PROTOCOLS }
)
)
end
end

View file

@ -3,6 +3,10 @@
class PreviewCard < ApplicationRecord class PreviewCard < ApplicationRecord
IMAGE_MIME_TYPES = ['image/jpeg', 'image/png', 'image/gif'].freeze IMAGE_MIME_TYPES = ['image/jpeg', 'image/png', 'image/gif'].freeze
self.inheritance_column = false
enum type: [:link, :photo, :video, :rich]
belongs_to :status belongs_to :status
has_attached_file :image, styles: { original: '120x120#' }, convert_options: { all: '-quality 80 -strip' } has_attached_file :image, styles: { original: '120x120#' }, convert_options: { all: '-quality 80 -strip' }

View file

@ -1,6 +1,8 @@
# frozen_string_literal: true # frozen_string_literal: true
class FetchAtomService < BaseService class FetchAtomService < BaseService
include HttpHelper
def call(url) def call(url)
return if url.blank? return if url.blank?
@ -45,8 +47,4 @@ class FetchAtomService < BaseService
def fetch(url) def fetch(url)
http_client.get(url).to_s http_client.get(url).to_s
end end
def http_client
HTTP.timeout(:per_operation, write: 10, connect: 10, read: 10).follow
end
end end

View file

@ -1,8 +1,9 @@
# frozen_string_literal: true # frozen_string_literal: true
class FetchLinkCardService < BaseService class FetchLinkCardService < BaseService
include HttpHelper
URL_PATTERN = %r{https?://\S+} URL_PATTERN = %r{https?://\S+}
USER_AGENT = "#{HTTP::Request::USER_AGENT} (Mastodon/#{Mastodon::VERSION}; +http://#{Rails.configuration.x.local_domain}/)"
def call(status) def call(status)
# Get first http/https URL that isn't local # Get first http/https URL that isn't local
@ -10,13 +11,53 @@ class FetchLinkCardService < BaseService
return if url.nil? return if url.nil?
card = PreviewCard.where(status: status).first_or_initialize(status: status, url: url)
attempt_opengraph(card, url) unless attempt_oembed(card, url)
end
private
def attempt_oembed(card, url)
response = OEmbed::Providers.get(url)
card.type = response.type
card.title = response.respond_to?(:title) ? response.title : ''
card.author_name = response.respond_to?(:author_name) ? response.author_name : ''
card.author_url = response.respond_to?(:author_url) ? response.author_url : ''
card.provider_name = response.respond_to?(:provider_name) ? response.provider_name : ''
card.provider_url = response.respond_to?(:provider_url) ? response.provider_url : ''
card.width = 0
card.height = 0
case card.type
when 'link'
card.image = URI.parse(response.thumbnail_url) if response.respond_to?(:thumbnail_url)
when 'photo'
card.url = response.url
card.width = response.width.presence || 0
card.height = response.height.presence || 0
when 'video'
card.width = response.width.presence || 0
card.height = response.height.presence || 0
card.html = Formatter.instance.sanitize(response.html, Sanitize::Config::MASTODON_OEMBED)
when 'rich'
# Most providers rely on <script> tags, which is a no-no
return false
end
card.save_with_optional_image!
rescue OEmbed::NotFound
false
end
def attempt_opengraph(card, url)
response = http_client.get(url) response = http_client.get(url)
return if response.code != 200 || response.mime_type != 'text/html' return if response.code != 200 || response.mime_type != 'text/html'
page = Nokogiri::HTML(response.to_s) page = Nokogiri::HTML(response.to_s)
card = PreviewCard.where(status: status).first_or_initialize(status: status, url: url)
card.type = :link
card.title = meta_property(page, 'og:title') || page.at_xpath('//title')&.content card.title = meta_property(page, 'og:title') || page.at_xpath('//title')&.content
card.description = meta_property(page, 'og:description') || meta_property(page, 'description') card.description = meta_property(page, 'og:description') || meta_property(page, 'description')
card.image = URI.parse(Addressable::URI.parse(meta_property(page, 'og:image')).normalize.to_s) if meta_property(page, 'og:image') card.image = URI.parse(Addressable::URI.parse(meta_property(page, 'og:image')).normalize.to_s) if meta_property(page, 'og:image')
@ -26,12 +67,6 @@ class FetchLinkCardService < BaseService
card.save_with_optional_image! card.save_with_optional_image!
end end
private
def http_client
HTTP.headers(user_agent: USER_AGENT).timeout(:per_operation, write: 10, connect: 10, read: 10).follow
end
def meta_property(html, property) def meta_property(html, property)
html.at_xpath("//meta[@property=\"#{property}\"]")&.attribute('content')&.value || html.at_xpath("//meta[@name=\"#{property}\"]")&.attribute('content')&.value html.at_xpath("//meta[@property=\"#{property}\"]")&.attribute('content')&.value || html.at_xpath("//meta[@name=\"#{property}\"]")&.attribute('content')&.value
end end

View file

@ -2,6 +2,7 @@
class FollowRemoteAccountService < BaseService class FollowRemoteAccountService < BaseService
include OStatus2::MagicKey include OStatus2::MagicKey
include HttpHelper
DFRN_NS = 'http://purl.org/macgirvin/dfrn/1.0' DFRN_NS = 'http://purl.org/macgirvin/dfrn/1.0'
@ -73,7 +74,7 @@ class FollowRemoteAccountService < BaseService
end end
def get_feed(url) def get_feed(url)
response = http_client.get(Addressable::URI.parse(url).normalize) response = http_client(write: 20, connect: 20, read: 50).get(Addressable::URI.parse(url).normalize)
[response.to_s, Nokogiri::XML(response)] [response.to_s, Nokogiri::XML(response)]
end end
@ -98,8 +99,4 @@ class FollowRemoteAccountService < BaseService
def get_profile(body, account) def get_profile(body, account)
RemoteProfileUpdateWorker.perform_async(account.id, body.force_encoding('UTF-8'), false) RemoteProfileUpdateWorker.perform_async(account.id, body.force_encoding('UTF-8'), false)
end end
def http_client
HTTP.timeout(:per_operation, write: 20, connect: 20, read: 50)
end
end end

View file

@ -34,7 +34,7 @@ class PostStatusService < BaseService
process_mentions_service.call(status) process_mentions_service.call(status)
process_hashtags_service.call(status) process_hashtags_service.call(status)
LinkCrawlWorker.perform_async(status.id) LinkCrawlWorker.perform_async(status.id) unless status.spoiler_text.present?
DistributionWorker.perform_async(status.id) DistributionWorker.perform_async(status.id)
Pubsubhubbub::DistributionWorker.perform_async(status.stream_entry.id) Pubsubhubbub::DistributionWorker.perform_async(status.stream_entry.id)

View file

@ -1,5 +1,7 @@
object @card object @card
attributes :url, :title, :description attributes :url, :title, :description, :type,
:author_name, :author_url, :provider_name,
:provider_url, :html, :width, :height
node(:image) { |card| card.image? ? full_asset_url(card.image.url(:original)) : nil } node(:image) { |card| card.image? ? full_asset_url(card.image.url(:original)) : nil }

View file

@ -36,7 +36,7 @@ Doorkeeper.configure do
# Reuse access token for the same resource owner within an application (disabled by default) # Reuse access token for the same resource owner within an application (disabled by default)
# Rationale: https://github.com/doorkeeper-gem/doorkeeper/issues/383 # Rationale: https://github.com/doorkeeper-gem/doorkeeper/issues/383
# reuse_access_token reuse_access_token
# Issue access tokens with refresh token (disabled by default) # Issue access tokens with refresh token (disabled by default)
# use_refresh_token # use_refresh_token

View file

@ -1,4 +1,5 @@
# frozen_string_literal: true # frozen_string_literal: true
Kaminari.configure do |config| Kaminari.configure do |config|
config.default_per_page = 40 config.default_per_page = 40
config.window = 1 config.window = 1

View file

@ -0,0 +1,4 @@
# frozen_string_literal: true
require_relative '../../app/lib/provider_discovery'
OEmbed::Providers.register_fallback(ProviderDiscovery)

View file

@ -0,0 +1,12 @@
class AddOEmbedToPreviewCards < ActiveRecord::Migration[5.0]
def change
add_column :preview_cards, :type, :integer, default: 0, null: false
add_column :preview_cards, :html, :text, null: false, default: ''
add_column :preview_cards, :author_name, :string, null: false, default: ''
add_column :preview_cards, :author_url, :string, null: false, default: ''
add_column :preview_cards, :provider_name, :string, null: false, default: ''
add_column :preview_cards, :provider_url, :string, null: false, default: ''
add_column :preview_cards, :width, :integer, default: 0, null: false
add_column :preview_cards, :height, :integer, default: 0, null: false
end
end

View file

@ -10,7 +10,7 @@
# #
# It's strongly recommended that you check this file into your version control system. # It's strongly recommended that you check this file into your version control system.
ActiveRecord::Schema.define(version: 20170425131920) do ActiveRecord::Schema.define(version: 20170425202925) do
# These are extensions that must be enabled in order to support this database # These are extensions that must be enabled in order to support this database
enable_extension "plpgsql" enable_extension "plpgsql"
@ -203,6 +203,14 @@ ActiveRecord::Schema.define(version: 20170425131920) do
t.datetime "image_updated_at" t.datetime "image_updated_at"
t.datetime "created_at", null: false t.datetime "created_at", null: false
t.datetime "updated_at", null: false t.datetime "updated_at", null: false
t.integer "type", default: 0, null: false
t.text "html", default: "", null: false
t.string "author_name", default: "", null: false
t.string "author_url", default: "", null: false
t.string "provider_name", default: "", null: false
t.string "provider_url", default: "", null: false
t.integer "width", default: 0, null: false
t.integer "height", default: 0, null: false
t.index ["status_id"], name: "index_preview_cards_on_status_id", unique: true, using: :btree t.index ["status_id"], name: "index_preview_cards_on_status_id", unique: true, using: :btree
end end

View file

@ -9,6 +9,6 @@ RSpec.describe FetchLinkCardService do
status = Fabricate(:status, text: 'Check out http://example.中国') status = Fabricate(:status, text: 'Check out http://example.中国')
FetchLinkCardService.new.call(status) FetchLinkCardService.new.call(status)
expect(a_request(:get, 'http://example.xn--fiqs8s/')).to have_been_made expect(a_request(:get, 'http://example.xn--fiqs8s/')).to have_been_made.at_least_once
end end
end end