Merge pull request #4457 from lllusion3469/fix_da

Fix Deviantart
2020-05-11 16:22:48 -05:00
parent 1578841a8a 45ae8bfb6f
commit deeb465b72
8 changed files with 282 additions and 70 deletions
--- a/app/logical/deviant_art_api_client.rb
+++ b/app/logical/deviant_art_api_client.rb
@@ -1,59 +1,69 @@
-# https://github.com/danbooru/danbooru/issues/4144
+# Authentication is via OAuth2 with the client credentials grant. Register a
+# new app at https://www.deviantart.com/developers/ to obtain a client_id and
+# client_secret. The app doesn't need to be published.
 #
 # API requests must send a user agent and must use gzip compression, otherwise
 # 403 errors will be returned.
+#
+# API calls operate on UUIDs. The deviation ID in the URL is not the UUID. UUIDs
+# are obtained by scraping the HTML page for the <meta property="da:appurl"> element.
+#
+# * https://www.deviantart.com/developers/
+# * https://www.deviantart.com/developers/authentication
+# * https://www.deviantart.com/developers/errors
+# * https://www.deviantart.com/developers/http/v1/20160316

-DeviantArtApiClient = Struct.new(:deviation_id) do
-  extend Memoist
+class DeviantArtApiClient
+  class Error < StandardError; end
+  BASE_URL = "https://www.deviantart.com/api/v1/oauth2/"

-  def extended_fetch
-    params = { deviationid: deviation_id, type: "art", include_session: false }
-    get("https://www.deviantart.com/_napi/da-deviation/shared_api/deviation/extended_fetch", params: params)
+  attr_reader :client_id, :client_secret
+
+  def initialize(client_id, client_secret)
+    @client_id, @client_secret = client_id, client_secret
  end

-  def extended_fetch_json
-    JSON.parse(extended_fetch.body).with_indifferent_access
+  # https://www.deviantart.com/developers/http/v1/20160316/deviation_single/bcc296bdf3b5e40636825a942a514816
+  def deviation(uuid)
+    request("deviation/#{uuid}")
  end

-  def download_url
-    url = extended_fetch_json.dig(:deviation, :extended, :download, :url)
-    response = get(url)
-    response.headers[:location]
+  # https://www.deviantart.com/developers/http/v1/20160316/deviation_download/bed6982b88949bdb08b52cd6763fcafd
+  def download(uuid, mature_content: "1")
+    request("deviation/download/#{uuid}", mature_content: mature_content)
  end

-  def get(url, retries: 1, **options)
-    response = http.cookies(cookies).get(url, **options)
+  # https://www.deviantart.com/developers/http/v1/20160316/deviation_metadata/7824fc14d6fba6acbacca1cf38c24158
+  def metadata(*uuids, mature_content: "1", ext_submission: "1", ext_camera: "1", ext_stats: "1")
+    params = {
+      deviationids: uuids.flatten,
+      mature_content: mature_content,
+      ext_submission: ext_submission,
+      ext_camera: ext_camera,
+      ext_stats: ext_stats
+    }

-    new_cookies = response.cookies.cookies.map { |cookie| { cookie.name => cookie.value } }.reduce(&:merge)
-    new_cookies = new_cookies.slice(:userinfo, :auth, :authsecure)
-    if new_cookies.present?
-      DanbooruLogger.info("DeviantArt: updating cookies", url: url, new_cookies: new_cookies, old_cookies: cookies)
-      self.cookies = new_cookies
-    end
-
-    # If the old auth cookie expired we may get a 404 with a new auth cookie
-    # set. Try again with the new cookie.
-    if response.code == 404 && retries > 0
-      DanbooruLogger.info("DeviantArt: retrying", url: url, cookies: cookies)
-      response = get(url, retries: retries - 1, **options)
-    end
-
-    response
+    request("deviation/metadata", **params)
  end

-  def cookies
-    Cache.get("deviantart_cookies", 10.years.to_i) do
-      JSON.parse(Danbooru.config.deviantart_cookies)
-    end
+  def request(url, **params)
+    params = { access_token: access_token.token, **params }
+
+    url = URI.join(BASE_URL, url).to_s
+    response = Danbooru::Http.cache(1.minute).get(url, params: params)
+    response.parse.with_indifferent_access
  end

-  def cookies=(new_cookies)
-    Cache.put("deviantart_cookies", new_cookies, 10.years.to_i)
+  def oauth
+    OAuth2::Client.new(client_id, client_secret, site: "https://www.deviantart.com", token_url: "/oauth2/token")
  end

-  def http
-    HTTP.use(:auto_inflate).headers(Danbooru.config.http_headers.merge("Accept-Encoding" => "gzip"))
+  def access_token
+    @access_token = oauth.client_credentials.get_token if @access_token.nil? || @access_token.expired?
+    @access_token
  end

-  memoize :extended_fetch, :extended_fetch_json, :download_url
+  def access_token=(hash)
+    @access_token = OAuth2::AccessToken.from_hash(oauth, hash)
+  end
 end
--- a/app/logical/sources/strategies.rb
+++ b/app/logical/sources/strategies.rb
@@ -5,6 +5,7 @@ module Sources
        Strategies::Pixiv,
        Strategies::NicoSeiga,
        Strategies::Twitter,
+        Strategies::Stash, # must come before DeviantArt
        Strategies::DeviantArt,
        Strategies::Tumblr,
        Strategies::ArtStation,
--- a/app/logical/sources/strategies/deviant_art.rb
+++ b/app/logical/sources/strategies/deviant_art.rb
@@ -86,33 +86,26 @@ module Sources
        # work is private, deleted, or the url didn't contain a deviation id; use image url as given by user.
        if api_deviation.blank?
          url
-        elsif api_deviation[:isDownloadable]
-          api_client.download_url
+        elsif api_deviation[:is_downloadable]
+          api_download[:src]
+        elsif api_deviation[:flash].present?
+          api_deviation.dig(:flash, :src)
+        elsif api_deviation[:videos].present?
+          api_deviation[:videos].max_by { |x| x[:filesize] }[:src]
        else
-          media = api_deviation[:media]
-          token = media[:token].first
-          fullview = media[:types].find { |data| data[:t] == "fullview" && data[:c].present? }
-
-          if fullview.present?
-            op = fullview[:c].gsub('<prettyName>', media[:prettyName])
-            src = "#{media[:baseUri]}/#{op}?token=#{token}"
-          else
-            src = "#{media[:baseUri]}?token=#{token}"
+          src = api_deviation.dig(:content, :src)
+          if deviation_id && deviation_id.to_i <= 790677560 && src =~ /^https:\/\/images-wixmp-/ && src !~ /\.gif\?/
+            src = src.sub(%r!(/f/[a-f0-9-]+/[a-f0-9-]+)!, '/intermediary\1')
+            src = src.sub(%r!/v1/(fit|fill)/.*\z!i, "")
          end
-
-          if deviation_id && deviation_id.to_i <= 790677560 && src =~ /\Ahttps:\/\/images-wixmp-/i
-            src = src.gsub(%r!(/f/[a-f0-9-]+/[a-f0-9-]+)!, '/intermediary\1')
-            src = src.gsub(%r!/v1/(fit|fill)/.*\z!i, "")
-          end
-
-          src = src.gsub(%r!\Ahttps?://orig\d+\.deviantart\.net!i, "http://origin-orig.deviantart.net")
+          src = src.sub(%r!\Ahttps?://orig\d+\.deviantart\.net!i, "http://origin-orig.deviantart.net")
          src = src.gsub(%r!q_\d+,strp!, "q_100")
          src
        end
      end

      def page_url
-        if api_deviation[:url].present?
+        if api_deviation.present?
          api_deviation[:url]
        elsif deviation_id.present?
          page_url_from_image_url
@@ -143,7 +136,7 @@ module Sources
      def artist_name
        if artist_name_from_url.present?
          artist_name_from_url
-        elsif api_deviation.dig(:author, :username).present?
+        elsif api_metadata.present?
          api_metadata.dig(:author, :username)
        else
          nil
@@ -151,12 +144,11 @@ module Sources
      end

      def artist_commentary_title
-        api_deviation[:title]
+        api_metadata[:title]
      end

      def artist_commentary_desc
-        return nil unless api_deviation.dig(:extended, :description).present?
-        api_deviation.dig(:extended, :description)
+        api_metadata[:description]
      end

      def normalized_for_artist_finder?
@@ -172,10 +164,12 @@ module Sources
      end

      def tags
-        return [] unless api_deviation.dig(:extended, :tags).present?
+        if api_metadata.blank?
+          return []
+        end

-        api_deviation.dig(:extended, :tags).map do |tag|
-          [tag[:name], tag[:url]]
+        api_metadata[:tags].map do |tag|
+          [tag[:tag_name], "https://www.deviantart.com/tag/#{tag[:tag_name]}"]
        end
      end

@@ -251,19 +245,70 @@ module Sources
        self.class.title_from_url(url) || self.class.title_from_url(referer_url)
      end

-      def api_client
-        @api_client ||= DeviantArtApiClient.new(deviation_id)
+      def page
+        return nil if page_url_from_image_url.blank?
+
+        resp = Danbooru::Http.cache(1.minute).get(page_url_from_image_url, follow: {max_hops: 1})
+
+        if resp.status.success?
+          Nokogiri::HTML(resp.body.to_s)
+        # the work was deleted
+        elsif resp.code == 404
+          nil
+        else
+          raise "failed to fetch page (got code #{resp.code})"
+        end
      end
+      memoize :page
+
+      # Scrape UUID from <meta property="da:appurl" content="DeviantArt://deviation/12F08C5D-A3A4-338C-2F1A-7E4E268C0E8B">
+      # For hidden or deleted works the UUID will be nil.
+      def uuid
+        return nil if page.nil?
+        meta = page.at_css('meta[property="da:appurl"]')
+        return nil if meta.nil?
+
+        appurl = meta["content"]
+        uuid = appurl[%r!\ADeviantArt://deviation/(.*)\z!, 1]
+        uuid
+      end
+      memoize :uuid
+
+      def api_client
+        api_client = DeviantArtApiClient.new(
+          Danbooru.config.deviantart_client_id,
+          Danbooru.config.deviantart_client_secret
+        )
+        api_client.access_token = Cache.get("da-access-token", 11.weeks) do
+          api_client.access_token.to_hash
+        end
+        api_client
+      end
+      memoize :api_client

      def api_deviation
-        api_client.extended_fetch_json[:deviation] || {}
+        return {} if uuid.nil?
+        api_client.deviation(uuid)
      end
+      memoize :api_deviation
+
+      def api_metadata
+        return {} if uuid.nil?
+        api_client.metadata(uuid)[:metadata].first
+      end
+      memoize :api_metadata
+
+      def api_download
+        return {} unless uuid.present? && api_deviation[:is_downloadable]
+        api_client.download(uuid)
+      end
+      memoize :api_download

      def api_response
        {
-          code: api_client.extended_fetch.code,
-          headers: api_client.extended_fetch.headers.to_h,
-          body: api_client.extended_fetch_json
+          deviation: api_deviation,
+          metadata: api_metadata,
+          download: api_download
        }
      end
    end
--- a/app/logical/sources/strategies/stash.rb
+++ b/app/logical/sources/strategies/stash.rb
@@ -0,0 +1,55 @@
+# Page URLs:
+# * https://sta.sh/0wxs31o7nn2 (single image)
+# * https://sta.sh/21leo8mz87ue (folder)
+#
+# Image URLs:
+# * https://orig00.deviantart.net/0fd2/f/2018/252/9/c/a_pepe_by_noizave-dcmga0s.png
+#
+# Ref:
+# * https://github.com/danbooru/danbooru/issues/3877
+# * https://www.deviantartsupport.com/en/article/what-is-stash-3391708
+# * https://www.deviantart.com/developers/http/v1/20160316/stash_item/4662dd8b10e336486ea9a0b14da62b74
+#
+module Sources
+  module Strategies
+    class Stash < DeviantArt
+      STASH = %r{\Ahttps?://sta\.sh/(?<post_id>[0-9a-zA-Z]+)}i
+
+      def domains
+        ["deviantart.net", "sta.sh"]
+      end
+
+      def match?
+        parsed_urls.map(&:domain).any?("sta.sh")
+      end
+
+      def site_name
+        "Sta.sh"
+      end
+
+      def canonical_url
+        page_url
+      end
+
+      def page_url
+        page_url_from_image_url
+      end
+
+      def page_url_from_image_url
+        "https://sta.sh/#{stash_id}"
+      end
+
+      def self.stash_id_from_url(url)
+        if url =~ STASH
+          $~[:post_id].downcase
+        else
+          nil
+        end
+      end
+
+      def stash_id
+        [url, referer_url].map { |x| self.class.stash_id_from_url(x) }.compact.first
+      end
+    end
+  end
+end
--- a/config/danbooru_default_config.rb
+++ b/config/danbooru_default_config.rb
@@ -333,6 +333,15 @@ module Danbooru
      nil
    end

+    # Register at https://www.deviantart.com/developers/
+    def deviantart_client_id
+      nil
+    end
+    
+    def deviantart_client_secret
+      nil
+    end    
+
    # http://tinysubversions.com/notes/mastodon-bot/
    def pawoo_client_id
      nil
--- a/test/unit/artist_test.rb
+++ b/test/unit/artist_test.rb
@@ -205,6 +205,7 @@ class ArtistTest < ActiveSupport::TestCase

    context "when finding deviantart artists" do
      setup do
+        skip "DeviantArt API keys not set" unless Danbooru.config.deviantart_client_id.present?
        FactoryBot.create(:artist, :name => "artgerm", :url_string => "http://artgerm.deviantart.com/")
        FactoryBot.create(:artist, :name => "trixia",  :url_string => "http://trixdraws.deviantart.com/")
      end
--- a/test/unit/sources/deviant_art_test.rb
+++ b/test/unit/sources/deviant_art_test.rb
@@ -2,6 +2,11 @@ require 'test_helper'

 module Sources
  class DeviantArtTest < ActiveSupport::TestCase
+    def setup
+      super
+      skip "DeviantArt API keys not set" unless Danbooru.config.deviantart_client_id.present?
+    end
+
    context "A page url" do
      setup do
        @site = Sources::Strategies.find("https://www.deviantart.com/aeror404/art/Holiday-Elincia-424551484")
@@ -239,6 +244,36 @@ module Sources
      end
    end

+    context "The source for a non-downloadable animated gif with id<=790677560" do
+      should "return working image url" do
+        @site = Sources::Strategies.find("https://www.deviantart.com/heartgear/art/Silent-Night-579982816")
+
+        # md5: 62caac1863aa264a56d548b4b7607097
+        assert_match(%r!\Ahttps://images-wixmp-ed30a86b8c4ca887773594c2\.wixmp\.com/f/ea95be00-c5aa-4063-bd55-f5a9183912f7/d9lb1ls-7d625444-0003-4123-bf00-274737ca7fdd.gif\?token=!, @site.image_url)
+        assert_downloaded(350_156, @site.image_url)
+      end
+    end
+
+    context "The source for a non-downloadable flash file" do
+      should "return working image url" do
+        @site = Sources::Strategies.find("https://www.deviantart.com/heartgear/art/SL-40v3-522007633")
+
+        # md5: 6adf1a3d532f898f44cf9948cbc7db7d
+        assert_match(%r!\Ahttps://api-da\.wixmp\.com/_api/download/file\?downloadToken=!, @site.image_url)
+        assert_downloaded(3_496_110, @site.image_url)
+      end
+    end
+
+    context "The source for a non-downloadable video file" do
+      should "return working image url" do
+        @site = Sources::Strategies.find("https://www.deviantart.com/gs-mantis/art/Chen-Goes-Fishing-505847233")
+
+        # md5: 344ac2b9fd5a87982af4b648aa2b2b0d
+        assert_equal("https://wixmp-ed30a86b8c4ca887773594c2.wixmp.com/v/mp4/fe046bc7-4d68-4699-96c1-19aa464edff6/d8d6281-91959e92-214f-4b2d-a138-ace09f4b6d09.1080p.8e57939eba634743a9fa41185e398d00.mp4", @site.image_url)
+        assert_downloaded(9_739_947, @site.image_url)
+      end
+    end
+
    context "The source for an DeviantArt artwork page" do
      setup do
        @site = Sources::Strategies.find("http://noizave.deviantart.com/art/test-post-please-ignore-685436408")
--- a/test/unit/sources/stash_test.rb
+++ b/test/unit/sources/stash_test.rb
@@ -0,0 +1,56 @@
+require 'test_helper'
+
+module Sources
+  class StashTest < ActiveSupport::TestCase
+    def setup
+      super
+      skip "DeviantArt API keys not set" unless Danbooru.config.deviantart_client_id.present?
+    end
+
+    context "A https://sta.sh/:id page url" do
+      should "work" do
+        @site = Sources::Strategies.find("https://sta.sh/0wxs31o7nn2")
+
+        assert_equal("noizave", @site.artist_name)
+        assert_equal("https://www.deviantart.com/noizave", @site.profile_url)
+
+        assert_equal("A pepe", @site.artist_commentary_title)
+        assert_equal("This is a test.", @site.artist_commentary_desc)
+
+        assert_equal("https://sta.sh/0wxs31o7nn2", @site.page_url)
+        assert_equal("https://sta.sh/0wxs31o7nn2", @site.canonical_url)
+        assert_match(%r!\Ahttps://api-da\.wixmp\.com/_api/download/file\?downloadToken=!, @site.image_url)
+      end
+    end
+
+    context "A https://orig00.deviantart.net/* image url" do
+      context "with a https://sta.sh/:id referer" do
+        should "work" do
+          @site = Sources::Strategies.find("https://orig00.deviantart.net/0fd2/f/2018/252/9/c/a_pepe_by_noizave-dcmga0s.png", "https://sta.sh/0wxs31o7nn2")
+
+          assert_equal("noizave", @site.artist_name)
+          assert_equal("https://www.deviantart.com/noizave", @site.profile_url)
+
+          assert_equal("A pepe", @site.artist_commentary_title)
+          assert_equal("This is a test.", @site.artist_commentary_desc)
+
+          assert_equal("https://sta.sh/0wxs31o7nn2", @site.page_url)
+          assert_equal("https://sta.sh/0wxs31o7nn2", @site.canonical_url)
+          assert_match(%r!\Ahttps://api-da\.wixmp\.com/_api/download/file\?downloadToken=!, @site.image_url)
+        end
+      end
+
+      context "without a referer" do
+        should "use the base deviantart strategy" do
+          @site = Sources::Strategies.find("https://orig00.deviantart.net/0fd2/f/2018/252/9/c/a_pepe_by_noizave-dcmga0s.png")
+
+          # if all we have is the image url, then we can't tell that this is really a sta.sh image.
+          assert_equal("Deviant Art", @site.site_name)
+
+          # this is the wrong page, but there's no way to know the correct sta.sh page without the referer.
+          assert_equal("https://www.deviantart.com/noizave/art/A-Pepe-763305148", @site.page_url)
+        end
+      end
+    end
+  end
+end