Make PreviewCard records reuseable between statuses (#4642)

* Make PreviewCard records reuseable between statuses

**Warning!** Migration truncates preview_cards tablec

* Allow a wider thumbnail for link preview, display it in horizontal layout (#4648)

* Delete preview cards files before truncating

* Rename old table instead of truncating it

* Add mastodon:maintenance:remove_deprecated_preview_cards

* Ignore deprecated_preview_cards in schema definition

* Fix null behaviour

1 files changed, 60 insertions, 40 deletions

diff --git a/app/services/fetch_link_card_service.rb b/app/services/fetch_link_card_service.rb
index 20c85e0ea..c38e9e7df 100644
--- a/app/services/fetch_link_card_service.rb
+++ b/app/services/fetch_link_card_service.rb
@@ -4,29 +4,45 @@ class FetchLinkCardService < BaseService
   URL_PATTERN = %r{https?://\S+}
 
   def call(status)
-    # Get first http/https URL that isn't local
-    url = parse_urls(status)
+    @status = status
+    @url    = parse_urls
 
-    return if url.nil?
+    return if @url.nil? || @status.preview_cards.any?
 
-    url  = url.to_s
-    card = PreviewCard.where(status: status).first_or_initialize(status: status, url: url)
-    res  = Request.new(:head, url).perform
+    @url = @url.to_s
 
-    return if res.code != 200 || res.mime_type != 'text/html'
+    RedisLock.acquire(lock_options) do |lock|
+      if lock.acquired?
+        @card = PreviewCard.find_by(url: @url)
+        process_url if @card.nil?
+      end
+    end
 
-    attempt_opengraph(card, url) unless attempt_oembed(card, url)
+    attach_card unless @card.nil?
   rescue HTTP::ConnectionError, OpenSSL::SSL::SSLError
     nil
   end
 
   private
 
-  def parse_urls(status)
-    if status.local?
-      urls = status.text.match(URL_PATTERN).to_a.map { |uri| Addressable::URI.parse(uri).normalize }
+  def process_url
+    @card = PreviewCard.new(url: @url)
+    res   = Request.new(:head, @url).perform
+
+    return if res.code != 200 || res.mime_type != 'text/html'
+
+    attempt_oembed || attempt_opengraph
+  end
+
+  def attach_card
+    @status.preview_cards << @card
+  end
+
+  def parse_urls
+    if @status.local?
+      urls = @status.text.match(URL_PATTERN).to_a.map { |uri| Addressable::URI.parse(uri).normalize }
     else
-      html  = Nokogiri::HTML(status.text)
+      html  = Nokogiri::HTML(@status.text)
       links = html.css('a')
       urls  = links.map { |a| Addressable::URI.parse(a['href']).normalize unless skip_link?(a) }.compact
     end
@@ -44,41 +60,41 @@ class FetchLinkCardService < BaseService
     a['rel']&.include?('tag') || a['class']&.include?('u-url')
   end
 
-  def attempt_oembed(card, url)
-    response = OEmbed::Providers.get(url)
+  def attempt_oembed
+    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
+    @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
+    case @card.type
     when 'link'
-      card.image = URI.parse(response.thumbnail_url) if response.respond_to?(:thumbnail_url)
+      @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
+      @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)
+      @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!
+    @card.save_with_optional_image!
   rescue OEmbed::NotFound
     false
   end
 
-  def attempt_opengraph(card, url)
-    response = Request.new(:get, url).perform
+  def attempt_opengraph
+    response = Request.new(:get, @url).perform
 
     return if response.code != 200 || response.mime_type != 'text/html'
 
@@ -88,19 +104,23 @@ class FetchLinkCardService < BaseService
     detector.strip_tags = true
 
     guess = detector.detect(html, response.charset)
-    page = Nokogiri::HTML(html, nil, guess&.fetch(:encoding))
+    page  = Nokogiri::HTML(html, nil, guess&.fetch(:encoding))
 
-    card.type             = :link
-    card.title            = meta_property(page, 'og:title') || page.at_xpath('//title')&.content
-    card.description      = meta_property(page, 'og:description') || meta_property(page, 'description')
-    card.image_remote_url = meta_property(page, 'og:image') if meta_property(page, 'og:image')
+    @card.type             = :link
+    @card.title            = meta_property(page, 'og:title') || page.at_xpath('//title')&.content || ''
+    @card.description      = meta_property(page, 'og:description') || meta_property(page, 'description') || ''
+    @card.image_remote_url = meta_property(page, 'og:image') if meta_property(page, 'og:image')
 
-    return if card.title.blank?
+    return if @card.title.blank?
 
-    card.save_with_optional_image!
+    @card.save_with_optional_image!
   end
 
   def meta_property(html, property)
     html.at_xpath("//meta[@property=\"#{property}\"]")&.attribute('content')&.value || html.at_xpath("//meta[@name=\"#{property}\"]")&.attribute('content')&.value
   end
+
+  def lock_options
+    { redis: Redis.current, key: "fetch:#{@url}" }
+  end
 end