danbooru::http: factor out Cloudflare Polish bypassing.

* Factor out the Cloudflare Polish bypass code to a standalone feature.

* Add `http_downloader` method to the base source strategy. This is a
  HTTP client that should be used for downloading images or making
  requests to images. This client ensures that referrer spoofing and
  Cloudflare bypassing are performed.

This fixes a bug with the upload page reporting the polished filesize
instead of the original filesize when uploading ArtStation images.

app/logical/danbooru/http.rb

@@ -5,6 +5,7 @@ require "danbooru/http/redirector"
 require "danbooru/http/retriable"
 require "danbooru/http/session"
 require "danbooru/http/spoof_referrer"
+require "danbooru/http/unpolish_cloudflare"
 
 module Danbooru
   class Http
@@ -26,7 +27,6 @@ module Danbooru
         .timeout(DEFAULT_TIMEOUT)
         .headers("Accept-Encoding" => "gzip")
         .headers("User-Agent": "#{Danbooru.config.canonical_app_name}/#{Rails.application.config.x.git_hash}")
-        .use(:spoof_referrer)
         .use(:auto_inflate)
         .use(redirector: { max_redirects: MAX_REDIRECTS })
         .use(:session)
@@ -98,20 +98,9 @@ module Danbooru
     end
 
     concerning :DownloadMethods do
-      def download_media(url, no_polish: true, **options)
+      def download_media(url, file: Tempfile.new("danbooru-download-", binmode: true))
         response = get(url)
 
-        # prevent Cloudflare Polish from modifying images.
-        if no_polish && response.headers["CF-Polished"].present?
-          url.query_values = url.query_values.to_h.merge(danbooru_no_polish: SecureRandom.uuid)
-          return download_media(url, no_polish: false)
-        end
-
-        file = download_response(response, **options)
-        [response, MediaFile.open(file)]
-      end
-
-      def download_response(response, file: Tempfile.new("danbooru-download-", binmode: true))
         raise DownloadError, "Downloading #{response.uri} failed with code #{response.status}" if response.status != 200
         raise FileTooLargeError, response if @max_size && response.content_length.to_i > @max_size
 
@@ -123,7 +112,7 @@ module Danbooru
         end
 
         file.rewind
-        file
+        [response, MediaFile.open(file)]
       end
     end
 

app/logical/danbooru/http/unpolish_cloudflare.rb

@@ -0,0 +1,20 @@
+# Bypass Cloudflare Polish (https://support.cloudflare.com/hc/en-us/articles/360000607372-Using-Cloudflare-Polish-to-compress-images)
+
+module Danbooru
+  class Http
+    class UnpolishCloudflare < HTTP::Feature
+      HTTP::Options.register_feature :unpolish_cloudflare, self
+
+      def perform(request, &block)
+        response = yield request
+
+        if response.headers["CF-Polished"].present?
+          request.uri.query_values = request.uri.query_values.to_h.merge(danbooru_no_polish: SecureRandom.uuid)
+          response = yield request
+        end
+
+        response
+      end
+    end
+  end
+end

app/logical/sources/strategies/base.rb

@@ -146,7 +146,7 @@ module Sources
 
       # Returns the size of the image resource without actually downloading the file.
       def remote_size
-        response = http.head(image_url)
+        response = http_downloader.head(image_url)
         return nil unless response.status == 200 && response.content_length.present?
 
         response.content_length.to_i
@@ -156,16 +156,23 @@ module Sources
       # Download the file at the given url, or at the main image url by default.
       def download_file!(download_url = image_url)
         raise DownloadError, "Download failed: couldn't find download url for #{url}" if download_url.blank?
-        response, file = http.download_media(download_url)
+        response, file = http_downloader.download_media(download_url)
         raise DownloadError, "Download failed: #{download_url} returned error #{response.status}" if response.status != 200
         file
       end
 
+      # A http client for API requests.
       def http
-        Danbooru::Http.headers(headers).public_only.timeout(30).max_size(Danbooru.config.max_file_size)
+        Danbooru::Http.new.public_only
       end
       memoize :http
 
+      # A http client for downloading files.
+      def http_downloader
+        http.timeout(30).max_size(Danbooru.config.max_file_size).use(:spoof_referrer).use(:unpolish_cloudflare)
+      end
+      memoize :http_downloader
+
       # The url to use for artist finding purposes. This will be stored in the
       # artist entry. Normally this will be the profile url.
       def normalize_for_artist_finder
@@ -292,8 +299,8 @@ module Sources
         to_h.to_json
       end
 
-      def http_exists?(url, headers = {})
-        http.headers(headers).head(url).status.success?
+      def http_exists?(url)
+        http_downloader.head(url).status.success?
       end
 
       # Convert commentary to dtext by stripping html tags. Sites can override

app/logical/sources/strategies/moebooru.rb

@@ -155,7 +155,7 @@ module Sources
 
           # the api_response wasn't available because it's a deleted post.
           elsif post_md5.present?
-            %w[jpg png gif].find { |ext| http_exists?("https://#{site_name}/image/#{post_md5}.#{ext}", headers) }
+            %w[jpg png gif].find { |ext| http_exists?("https://#{site_name}/image/#{post_md5}.#{ext}") }
 
           else
             nil

test/unit/danbooru_http_test.rb

@@ -129,13 +129,23 @@ class DanbooruHttpTest < ActiveSupport::TestCase
 
     context "spoof referrer feature" do
       should "spoof the referer" do
-        response = Danbooru::Http.get("https://httpbin.org/anything")
+        response = Danbooru::Http.use(:spoof_referrer).get("https://httpbin.org/anything")
 
         assert_equal(200, response.status)
         assert_equal("https://httpbin.org", response.parse.dig("headers", "Referer"))
       end
     end
 
+    context "unpolish cloudflare feature" do
+      should "return the original image for polished images" do
+        url = "https://cdnb.artstation.com/p/assets/images/images/025/273/307/4k/atey-ghailan-a-sage-keyart-s-ch-04-outlined-1.jpg?1585246642"
+        response = Danbooru::Http.use(:unpolish_cloudflare).get(url)
+
+        assert_equal(200, response.status)
+        assert_equal(720_743, response.content_length)
+      end
+    end
+
     context "#download method" do
       should "download files" do
         response, file = Danbooru::Http.download_media("https://httpbin.org/bytes/1000")