uploads: allow searching uploads and media assets by metatag.

Allow searching the /uploads and /media_assets pages by the following metatags:

* id:
* md5:
* width:
* height:
* duration:
* mpixels:
* ratio:
* filesize:
* filetype:
* date:
* age:
* status:<processing|active|deleted|expunged|failed> (for /media_assets)
* status:<pending|processing|active|failed> (for /uploads)
* is:<filetype>, is:<status>
* exif:

Examples:

* https://betabooru.donmai.us/media_assets?search[ai_tags_match]=filetype:png
* https://betabooru.donmai.us/uploads?search[ai_tags_match]=filetype:png

Note that in /uploads search, the id:, date:, and age: metatags refer to the upload media asset, not
the upload itself.

Note also that uploads may contain multiple assets, so for example searching uploads by
`filetype:png` will return all uploads containing at least one PNG file, even if they contain other
non-PNG files.

app/logical/media_asset_query.rb

@@ -1,8 +1,7 @@
 # frozen_string_literal: true
 
-# An AITagQuery is a tag search performed on media assets using AI tags. Only
-# basic tags are allowed, no metatags.
-class AITagQuery
+# A MediaAssetQuery is a tag search performed on media assets (or upload media assets) using AI tags.
+class MediaAssetQuery
   extend Memoist
 
   attr_reader :search_string
@@ -36,7 +35,7 @@ class AITagQuery
         ai_tag = AITag.named(node.name).where(score: score_range)
         relation.where(ai_tag.where(AITag.arel_table[:media_asset_id].eq(relation.arel_table[foreign_key])).arel.exists)
       in :metatag
-        relation.none
+        metatag_matches(node.name, node.value, relation)
       in :wildcard
         relation.none
       in :not
@@ -54,4 +53,37 @@ class AITagQuery
       end
     end
   end
+
+  def metatag_matches(name, value, relation)
+    case name
+    when "id"
+      relation.attribute_matches(value, :id)
+    when "md5"
+      relation.attribute_matches(value, "media_assets.md5", :md5)
+    when "width"
+      relation.attribute_matches(value, "media_assets.image_width")
+    when "height"
+      relation.attribute_matches(value, "media_assets.image_height")
+    when "duration"
+      relation.attribute_matches(value, "media_assets.duration", :float)
+    when "mpixels"
+      relation.attribute_matches(value, "(media_assets.image_width * media_assets.image_height) / 1000000.0", :float)
+    when "ratio"
+      relation.attribute_matches(value, "ROUND(media_assets.image_width::numeric / media_assets.image_height::numeric, 2)", :ratio)
+    when "filesize"
+      relation.attribute_matches(value, "media_assets.file_size", :filesize)
+    when "filetype"
+      relation.attribute_matches(value, "media_assets.file_ext", :enum)
+    when "date"
+      relation.attribute_matches(value, :created_at, :date)
+    when "age"
+      relation.attribute_matches(value, :created_at, :age)
+    when "status"
+      relation.attribute_matches(value, :status, :enum)
+    when "is"
+      relation.is_matches(value)
+    when "exif"
+      relation.exif_matches(value)
+    end
+  end
 end

app/models/media_asset.rb

@@ -204,7 +204,30 @@ class MediaAsset < ApplicationRecord
   concerning :SearchMethods do
     class_methods do
       def ai_tags_match(tag_string, score_range: (50..))
-        AITagQuery.search(tag_string, relation: self, score_range: score_range)
+        MediaAssetQuery.search(tag_string, relation: self, score_range: score_range)
+      end
+
+      def is_matches(value)
+        case value.downcase
+        when *MediaAsset.statuses.keys
+          where(status: value)
+        when *FILE_TYPES
+          attribute_matches(value, :file_ext, :enum)
+        else
+          none
+        end
+      end
+
+      def exif_matches(string)
+        # string = File:ColorComponents=3
+        if string.include?("=")
+          key, value = string.split(/=/, 2)
+          hash = { key => value }
+          joins(:media_metadata).where_json_contains("media_metadata.metadata", hash)
+        # string = File:ColorComponents
+        else
+          joins(:media_metadata).where_json_has_key("media_metadata.metadata", string)
+        end
       end
 
       def search(params, current_user)

app/models/post.rb

@@ -1284,17 +1284,7 @@ class Post < ApplicationRecord
       end
 
       def exif_matches(string)
-        # string = exif:File:ColorComponents=3
-        if string.include?("=")
-          key, value = string.split(/=/, 2)
-          hash = { key => value }
-          metadata = MediaMetadata.joins(:media_asset).where_json_contains(:metadata, hash)
-        # string = exif:File:ColorComponents
-        else
-          metadata = MediaMetadata.joins(:media_asset).where_json_has_key(:metadata, string)
-        end
-
-        where(md5: metadata.select(:md5))
+        where(md5: MediaAsset.exif_matches(string).select(:md5))
       end
 
       def ai_tags_include(value, default_confidence: ">=50")

app/models/upload.rb

@@ -126,7 +126,7 @@ class Upload < ApplicationRecord
   end
 
   def self.ai_tags_match(tag_string, score_range: (50..))
-    upload_media_assets = AITagQuery.search(tag_string, relation: UploadMediaAsset.all, foreign_key: :media_asset_id, score_range: score_range)
+    upload_media_assets = MediaAssetQuery.search(tag_string, relation: UploadMediaAsset.joins(:media_asset), foreign_key: :media_asset_id, score_range: score_range)
     where(upload_media_assets.where("upload_media_assets.upload_id = uploads.id").arel.exists)
   end
 

app/models/upload_media_asset.rb

@@ -41,6 +41,21 @@ class UploadMediaAsset < ApplicationRecord
     expired.update_all(status: :failed, error: "Stuck processing for more than 4 hours")
   end
 
+  def self.is_matches(value)
+    case value.downcase
+    when *UploadMediaAsset.statuses.keys
+      where(status: value)
+    when *MediaAsset::FILE_TYPES
+      attribute_matches(value, :file_ext, :enum)
+    else
+      none
+    end
+  end
+
+  def self.exif_matches(string)
+    merge(MediaAsset.exif_matches(string))
+  end
+
   def self.search(params, current_user)
     q = search_attributes(params, [:id, :created_at, :updated_at, :status, :source_url, :page_url, :error, :upload, :media_asset, :post], current_user: current_user)
 

test/factories/media_asset.rb

@@ -1,6 +1,6 @@
 FactoryBot.define do
   factory(:media_asset) do
-    md5 { SecureRandom.hex(32) }
+    md5 { SecureRandom.hex(16) }
     file_ext { "jpg" }
     file_size { 1_000_000 }
     image_width { 1000 }

test/unit/media_asset_test.rb

@@ -0,0 +1,74 @@
+require 'test_helper'
+
+class MediaAssetTest < ActiveSupport::TestCase
+  def assert_tag_match(assets, query)
+    assert_equal(assets.map(&:id), MediaAsset.ai_tags_match(query).order(id: :desc).pluck("id"))
+  end
+
+  context "MediaAsset" do
+    context "searching" do
+      setup do
+        @asset1 = create(:media_asset, image_width: 720, image_height: 1280, file_size: 1.megabyte, file_ext: "jpg", created_at: Time.zone.now, media_metadata: build(:media_metadata, metadata: { "File:FileType" => "JPEG" }))
+        @asset2 = create(:media_asset, image_width: 1920, image_height: 1080, file_size: 2.megabytes, file_ext: "png", duration: 3.0, created_at: Time.parse("2022-01-01"), media_metadata: build(:media_metadata, metadata: { "File:FileType" => "PNG" }))
+      end
+
+      should "return assets for the id: metatag" do
+        assert_tag_match([@asset1], "id:#{@asset1.id}")
+      end
+
+      should "return assets for the md5: metatag" do
+        assert_tag_match([@asset1], "md5:#{@asset1.md5}")
+      end
+
+      should "return assets for the width: metatag" do
+        assert_tag_match([@asset1], "width:#{@asset1.image_width}")
+      end
+
+      should "return assets for the height: metatag" do
+        assert_tag_match([@asset1], "height:#{@asset1.image_height}")
+      end
+
+      should "return assets for the duration: metatag" do
+        assert_tag_match([@asset2], "duration:3")
+      end
+
+      should "return assets for the mpixels: metatag" do
+        assert_tag_match([@asset1], "mpixels:#{(@asset1.image_width * @asset1.image_height) / 1_000_000.0}")
+      end
+
+      should "return assets for the ratio: metatag" do
+        assert_tag_match([@asset1], "ratio:#{@asset1.image_width.to_f / @asset1.image_height}")
+      end
+
+      should "return assets for the filesize: metatag" do
+        assert_tag_match([@asset1], "filesize:1mb")
+      end
+
+      should "return assets for the filetype: metatag" do
+        assert_tag_match([@asset1], "filetype:jpg")
+      end
+
+      should "return assets for the date: tag" do
+        assert_tag_match([@asset2], "date:2022-01-01")
+      end
+
+      should "return assets for the age: tag" do
+        assert_tag_match([@asset1], "age:<1minute")
+      end
+
+      should "return assets for the status: tag" do
+        assert_tag_match([@asset2, @asset1], "status:active")
+      end
+
+      should "return assets for the is: tag" do
+        assert_tag_match([@asset1], "is:jpg")
+        assert_tag_match([@asset2, @asset1], "is:active")
+      end
+
+      should "return assets for the exif: tag" do
+        assert_tag_match([@asset2, @asset1], "exif:File:FileType")
+        assert_tag_match([@asset1], "exif:File:FileType=JPEG")
+      end
+    end
+  end
+end

test/unit/upload_test.rb

@@ -0,0 +1,80 @@
+require 'test_helper'
+
+class UploadTest < ActiveSupport::TestCase
+  def assert_tag_match(uploads, query)
+    assert_equal(uploads.map(&:id), Upload.ai_tags_match(query).order(id: :desc).pluck("id"))
+  end
+
+  context "Upload" do
+    context "searching" do
+      setup do
+        @asset1 = create(:media_asset, image_width: 720, image_height: 1280, file_size: 1.megabyte, file_ext: "jpg", media_metadata: build(:media_metadata, metadata: { "File:FileType" => "JPEG" }))
+        @asset2 = create(:media_asset, image_width: 1920, image_height: 1080, file_size: 2.megabytes, file_ext: "png", duration: 3.0, media_metadata: build(:media_metadata, metadata: { "File:FileType" => "PNG" }))
+
+        @uma1 = build(:upload_media_asset, media_asset: @asset1, status: "active", created_at: Time.zone.now)
+        @uma2 = build(:upload_media_asset, media_asset: @asset2, status: "active", created_at: Time.parse("2022-01-01"))
+
+        @upload1 = create(:upload, created_at: Time.zone.now, upload_media_assets: [@uma1])
+        @upload2 = create(:upload, created_at: Time.parse("2022-01-01"), upload_media_assets: [@uma2])
+      end
+
+      should "return assets for the id: metatag" do
+        assert_tag_match([@upload1], "id:#{@upload1.upload_media_assets.sole.id}")
+      end
+
+      should "return assets for the md5: metatag" do
+        assert_tag_match([@upload1], "md5:#{@asset1.md5}")
+      end
+
+      should "return assets for the width: metatag" do
+        assert_tag_match([@upload1], "width:#{@asset1.image_width}")
+      end
+
+      should "return assets for the height: metatag" do
+        assert_tag_match([@upload1], "height:#{@asset1.image_height}")
+      end
+
+      should "return assets for the duration: metatag" do
+        assert_tag_match([@upload2], "duration:3")
+      end
+
+      should "return assets for the mpixels: metatag" do
+        assert_tag_match([@upload1], "mpixels:#{(@asset1.image_width * @asset1.image_height) / 1_000_000.0}")
+      end
+
+      should "return assets for the ratio: metatag" do
+        assert_tag_match([@upload1], "ratio:#{@asset1.image_width.to_f / @asset1.image_height}")
+      end
+
+      should "return assets for the filesize: metatag" do
+        assert_tag_match([@upload1], "filesize:1mb")
+      end
+
+      should "return assets for the filetype: metatag" do
+        assert_tag_match([@upload1], "filetype:jpg")
+      end
+
+      should "return assets for the date: tag" do
+        assert_tag_match([@upload2], "date:2022-01-01")
+      end
+
+      should "return assets for the age: tag" do
+        assert_tag_match([@upload1], "age:<1minute")
+      end
+
+      should "return assets for the status: tag" do
+        assert_tag_match([@upload2, @upload1], "status:active")
+      end
+
+      should "return assets for the is: tag" do
+        assert_tag_match([@upload1], "is:jpg")
+        assert_tag_match([@upload2, @upload1], "is:active")
+      end
+
+      should "return assets for the exif: tag" do
+        assert_tag_match([@upload2, @upload1], "exif:File:FileType")
+        assert_tag_match([@upload1], "exif:File:FileType=JPEG")
+      end
+    end
+  end
+end