Replace extended characters with ASCII equivalents in Playdate metadata tags

config/devices.yml

@@ -40,3 +40,4 @@ devices:
     - 44100
   file_types:
     - mp3
+  transliterate_metadata: true

lib/wavesync/audio.rb

@@ -69,6 +69,46 @@ def write_cue_points(cue_points)
       FileUtils.mv(temp_path, @file_path)
     end
 
+    ID3V2_FRAME_TITLE = 'TIT2'
+    ID3V2_FRAME_ARTIST = 'TPE1'
+    ID3V2_FRAME_ALBUM = 'TALB'
+    ID3V2_FRAME_ALBUM_ARTIST = 'TPE2'
+    ID3V2_FRAME_GENRE = 'TCON'
+    ID3V2_FRAME_COMPOSER = 'TCOM'
+    ID3V2_FRAME_ENCODED_BY = 'TENC'
+    ID3V2_FRAME_COMPILATION = 'TCMP'
+
+    COMBINING_MARKS = /\p{Mn}/
+
+    TRANSLITERATE_FRAME_IDS = [
+      ID3V2_FRAME_TITLE,
+      ID3V2_FRAME_ARTIST,
+      ID3V2_FRAME_ALBUM,
+      ID3V2_FRAME_ALBUM_ARTIST,
+      ID3V2_FRAME_GENRE,
+      ID3V2_FRAME_COMPOSER,
+      ID3V2_FRAME_ENCODED_BY,
+      ID3V2_FRAME_COMPILATION
+    ].freeze
+
+    #: () -> void
+    def transliterate_tags
+      return unless @file_ext == '.mp3'
+
+      TagLib::MPEG::File.open(@file_path) do |file|
+        tag = file.id3v2_tag
+        next if tag.nil?
+
+        TRANSLITERATE_FRAME_IDS.each do |frame_id|
+          tag.frame_list(frame_id).each do |frame|
+            frame.text = transliterate(frame.to_string)
+          end
+        end
+
+        file.save
+      end
+    end
+
     #: (String | Integer | Float bpm) -> void
     def write_bpm(bpm)
       case @file_ext
@@ -222,5 +262,12 @@ def write_id3v2_bpm(tag, bpm)
       frame.text = bpm.to_s
       tag.add_frame(frame)
     end
+
+    #: (String string) -> String
+    def transliterate(string)
+      string
+        .unicode_normalize(:nfd)
+        .gsub(COMBINING_MARKS, '')
+    end
   end
 end

lib/wavesync/device.rb

@@ -11,16 +11,18 @@ class Device
     attr_reader :bpm_source #: Symbol?
     attr_reader :bar_multiple #: Integer?
     attr_reader :unsupported_characters #: Array[String]
+    attr_reader :transliterate_metadata #: bool
 
-    #: (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String]) -> void
-    def initialize(name:, sample_rates:, file_types:, bit_depths: [], bpm_source: nil, bar_multiple: nil, unsupported_characters: [])
+    #: (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String], ?transliterate_metadata: bool) -> void
+    def initialize(name:, sample_rates:, file_types:, bit_depths: [], bpm_source: nil, bar_multiple: nil, unsupported_characters: [], transliterate_metadata: false)
       @name = name
       @sample_rates = sample_rates
       @bit_depths = bit_depths
       @file_types = file_types
       @bpm_source = bpm_source
       @bar_multiple = bar_multiple
       @unsupported_characters = unsupported_characters
+      @transliterate_metadata = transliterate_metadata
     end
 
     #: () -> String
@@ -49,7 +51,8 @@ def self.load_from_yaml
           file_types: attrs['file_types'],
           bpm_source: attrs['bpm_source']&.to_sym,
           bar_multiple: attrs['bar_multiple'],
-          unsupported_characters: attrs['unsupported_characters'] || []
+          unsupported_characters: attrs['unsupported_characters'] || [],
+          transliterate_metadata: attrs['transliterate_metadata'] || false
         )
       end
     end

lib/wavesync/scanner.rb

@@ -60,6 +60,7 @@ def sync(target_library_path, device, pad: false)
                                          before_transcode: -> { @ui.conversion_progress(source_format, target_format) }) do |local_temp_path|
             inject_acid_bpm(local_temp_path, bpm, device)
             inject_cue_points(local_temp_path, audio, source_format, target_format)
+            inject_transliterated_metadata(local_temp_path, device)
           end
           path_resolver.resolve(file, audio, target_file_type: target_format.file_type)
         else
@@ -76,7 +77,8 @@ def sync(target_library_path, device, pad: false)
             end
           else
             copied = copy_file(audio, file, path_resolver)
-            path_resolver.resolve(file, audio)
+            target_path = path_resolver.resolve(file, audio)
+            inject_transliterated_metadata(target_path.to_s, device) if copied
           end
           @ui.copy(source_format)
         end
@@ -116,6 +118,13 @@ def copy_file(audio, source_file_path, path_resolver)
       end
     end
 
+    #: (String path, Device device) -> void
+    def inject_transliterated_metadata(path, device)
+      return unless device.transliterate_metadata
+
+      Audio.new(path).transliterate_tags
+    end
+
     #: (String local_temp_path, (Integer | Float | String)? bpm, Device device) -> void
     def inject_acid_bpm(local_temp_path, bpm, device)
       return unless device.bpm_source == :acid_chunk && bpm && File.extname(local_temp_path).downcase == '.wav'

sig/generated/wavesync/audio.rbs

@@ -31,6 +31,29 @@ module Wavesync
     # : (Array[{identifier: Integer, sample_offset: Integer, label: String?}] cue_points) -> void
     def write_cue_points: (Array[{ identifier: Integer, sample_offset: Integer, label: String? }] cue_points) -> void
 
+    ID3V2_FRAME_TITLE: ::String
+
+    ID3V2_FRAME_ARTIST: ::String
+
+    ID3V2_FRAME_ALBUM: ::String
+
+    ID3V2_FRAME_ALBUM_ARTIST: ::String
+
+    ID3V2_FRAME_GENRE: ::String
+
+    ID3V2_FRAME_COMPOSER: ::String
+
+    ID3V2_FRAME_ENCODED_BY: ::String
+
+    ID3V2_FRAME_COMPILATION: ::String
+
+    COMBINING_MARKS: ::Regexp
+
+    TRANSLITERATE_FRAME_IDS: untyped
+
+    # : () -> void
+    def transliterate_tags: () -> void
+
     # : (String | Integer | Float bpm) -> void
     def write_bpm: (String | Integer | Float bpm) -> void
 
@@ -39,8 +62,8 @@ module Wavesync
 
     private
 
-    # : (Integer? target_sample_rate, Integer? target_bit_depth, ?Numeric? padding_seconds) -> Hash[Symbol, untyped]
-    def build_transcode_options: (Integer? target_sample_rate, Integer? target_bit_depth, ?Numeric? padding_seconds) -> Hash[Symbol, untyped]
+    # : (String target_file_type, Integer? target_bit_depth) -> String
+    def transcode_codec: (String target_file_type, Integer? target_bit_depth) -> String
 
     # : () -> (String | Integer)?
     def bpm_from_file: () -> (String | Integer)?
@@ -80,5 +103,8 @@ module Wavesync
 
     # : (untyped tag, String | Integer | Float bpm) -> void
     def write_id3v2_bpm: (untyped tag, String | Integer | Float bpm) -> void
+
+    # : (String string) -> String
+    def transliterate: (String string) -> String
   end
 end

sig/generated/wavesync/device.rbs

@@ -16,8 +16,10 @@ module Wavesync
 
     attr_reader unsupported_characters: Array[String]
 
-    # : (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String]) -> void
-    def initialize: (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String]) -> void
+    attr_reader transliterate_metadata: bool
+
+    # : (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String], ?transliterate_metadata: bool) -> void
+    def initialize: (name: String, sample_rates: Array[Integer], file_types: Array[String], ?bit_depths: Array[Integer], ?bpm_source: Symbol?, ?bar_multiple: Integer?, ?unsupported_characters: Array[String], ?transliterate_metadata: bool) -> void
 
     # : () -> String
     def self.config_path: () -> String

sig/generated/wavesync/scanner.rbs

@@ -16,6 +16,9 @@ module Wavesync
     # : (Audio audio, String source_file_path, PathResolver path_resolver) -> bool
     def copy_file: (Audio audio, String source_file_path, PathResolver path_resolver) -> bool
 
+    # : (String path, Device device) -> void
+    def inject_transliterated_metadata: (String path, Device device) -> void
+
     # : (String local_temp_path, (Integer | Float | String)? bpm, Device device) -> void
     def inject_acid_bpm: (String local_temp_path, (Integer | Float | String)? bpm, Device device) -> void
 

sig/generated/wavesync/set_editor.rbs

@@ -43,12 +43,6 @@ module Wavesync
     # : (Float semitones) -> String
     def format_pitch_shift: (Float semitones) -> String
 
-    # : () -> void
-    def kill_player: () -> void
-
-    # : () -> void
-    def check_player: () -> void
-
     private
 
     # : () -> void
@@ -105,9 +99,15 @@ module Wavesync
     # : () -> void
     def jump_to_next_cue: () -> void
 
+    # : () -> void
+    def kill_player: () -> void
+
     # : () -> void
     def stop_playback: () -> void
 
+    # : () -> void
+    def check_player: () -> void
+
     # : () -> void
     def advance_and_play: () -> void
 

sig/taglib.rbs

@@ -42,6 +42,7 @@ module TagLib
     class TextIdentificationFrame
       def initialize: (::String id, untyped encoding) -> void
       def text=: (::String text) -> ::String
+      def to_string: () -> ::String
     end
 
     class Tag

test/wavesync/audio_test.rb

@@ -64,6 +64,52 @@ class AudioTest < Wavesync::TestCase
       assert_nil audio('44100_16.aiff').bpm
     end
 
+    test 'transliterate_tags replaces umlauts across all supported id3v2 frames for mp3' do
+      with_temp_copy('44100.mp3') do |path|
+        write_id3v2_frames(
+          path,
+          'TIT2' => 'Jóga',
+          'TPE1' => 'Björk',
+          'TALB' => 'Åström Remixes',
+          'TPE2' => 'Sigur Rós',
+          'TCON' => 'Électronique',
+          'TCOM' => 'Jean-Michel Jarré',
+          'TENC' => 'Encöder',
+          'TCMP' => '1'
+        )
+
+        Audio.new(path).transliterate_tags
+
+        frames = read_id3v2_frames(path)
+        assert_equal 'Joga', frames['TIT2']
+        assert_equal 'Bjork', frames['TPE1']
+        assert_equal 'Astrom Remixes', frames['TALB']
+        assert_equal 'Sigur Ros', frames['TPE2']
+        assert_equal 'Electronique', frames['TCON']
+        assert_equal 'Jean-Michel Jarre', frames['TCOM']
+        assert_equal 'Encoder', frames['TENC']
+        assert_equal '1', frames['TCMP']
+      end
+    end
+
+    test 'transliterate_tags leaves ascii-only tags unchanged' do
+      with_temp_copy('44100.mp3') do |path|
+        write_id3v2_frames(path, 'TPE1' => 'Aphex Twin', 'TIT2' => 'Windowlicker')
+
+        Audio.new(path).transliterate_tags
+
+        frames = read_id3v2_frames(path)
+        assert_equal 'Aphex Twin', frames['TPE1']
+        assert_equal 'Windowlicker', frames['TIT2']
+      end
+    end
+
+    test 'transliterate_tags is a no-op for non-mp3 files' do
+      with_temp_copy('44100_16.wav') do |path|
+        Audio.new(path).transliterate_tags
+      end
+    end
+
     test 'write_bpm round-trips for wav via acid chunk' do
       with_temp_copy('44100_16.wav') do |path|
         Audio.new(path).write_bpm(128)
@@ -155,6 +201,30 @@ def audio(name)
       Audio.new(File.join(FIXTURES_PATH, name))
     end
 
+    def write_id3v2_frames(path, frames)
+      TagLib::MPEG::File.open(path) do |file|
+        tag = file.id3v2_tag(true)
+        frames.each do |frame_id, text|
+          tag.remove_frames(frame_id)
+          frame = TagLib::ID3v2::TextIdentificationFrame.new(frame_id, TagLib::String::UTF8)
+          frame.text = text
+          tag.add_frame(frame)
+        end
+        file.save
+      end
+    end
+
+    def read_id3v2_frames(path)
+      result = {}
+      TagLib::MPEG::File.open(path) do |file|
+        tag = file.id3v2_tag
+        tag.frame_list.each do |frame|
+          result[frame.frame_id] = frame.to_string
+        end
+      end
+      result
+    end
+
     def with_temp_copy(name)
       ext = File.extname(name)
       tmp = Tempfile.new(['audio_test', ext])

test/wavesync/device_test.rb

@@ -67,6 +67,20 @@ class DeviceTest < Wavesync::TestCase
       octatrack = Device.find_by(name: 'Octatrack')
       assert_includes octatrack.unsupported_characters, '"'
     end
+    test 'Playdate has transliterate_metadata enabled' do
+      playdate = Device.find_by(name: 'Playdate')
+      assert_equal true, playdate.transliterate_metadata
+    end
+
+    test 'transliterate_metadata defaults to false' do
+      device = Device.new(
+        name: 'Test',
+        sample_rates: [44_100],
+        file_types: ['wav']
+      )
+      assert_equal false, device.transliterate_metadata
+    end
+
     test 'unsupported_characters defaults to empty array' do
       device = Device.new(
         name: 'Test',