Fixed DataPreprocess bug, Added en-US lengcode in Dataprocess

EX3exp · Sep 26, 2024 · c4683c6 · c4683c6
1 parent 5c5a1f1
commit c4683c6
Show file tree

Hide file tree

Showing 6 changed files with 24 additions and 14 deletions.
diff --git a/Mirivoice/Mirivoice.Plugins.Builtin/IPAConverters/EnglishUSIPAConverter.cs b/Mirivoice/Mirivoice.Plugins.Builtin/IPAConverters/EnglishUSIPAConverter.cs
@@ -77,7 +77,7 @@ public override string ConvertToIPA(string phoneme, bool isFirstPhoneme)
                 }
             }
             string res = string.Join("\t", IPA);
-            //Log.Debug($"Converted {phoneme} to {res}");
+            Log.Debug($"Converted {phoneme} to {res}");
             return res;
         }
     }

diff --git a/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/BasePhonemizer.cs b/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/BasePhonemizer.cs
@@ -103,7 +103,7 @@ public async Task<string> ConvertToIPA(string sentence, DispatcherPriority dispa
 
             await Dispatcher.UIThread.InvokeAsync(async () =>
             {
-                IPAPhonemes.Clear();
+                IPAPhonemes = new string[variatedWords.Length].ToList();
                 bool divideWord = false;
                 var wordTasks = variatedWords
                     .Select(async (word, index) => await Task.Run(()=>
@@ -114,17 +114,17 @@ await Dispatcher.UIThread.InvokeAsync(async () =>
                             Log.Error($"[ConvertToIPA: Variated Sentence({words.Length})] - [Sentence length({variatedWords.Length})] mismatch");
                             if (UseWordDivider && phoneme.Trim() == string.Empty)
                             {
-                                IPAPhonemes.Add(" ");
+                                IPAPhonemes[index] = " ";
                             }
                             else
                             {
                                 if (index == 0)
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), true));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), true);
                                 }
                                 else
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), false));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), false);
                                 }
 
                             }
@@ -133,17 +133,17 @@ await Dispatcher.UIThread.InvokeAsync(async () =>
                         {
                             if (UseWordDivider && phoneme.Trim() == string.Empty)
                             {
-                                IPAPhonemes.Add(" ");
+                                IPAPhonemes[index] = " ";
                             }
                             else
                             {
                                 if (index == 0)
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), true));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), true);
                                 }
                                 else
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), false));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), false);
                                 }
                             }
                         }
@@ -292,7 +292,7 @@ public async Task GenerateIPAAsync(LineBoxView l, DispatcherPriority dispatcherP
 
             await Dispatcher.UIThread.InvokeAsync(async () =>
             {
-                IPAPhonemes.Clear();
+                IPAPhonemes = new string[l.MResultsCollection.Count].ToList();
 
                 bool divideWord = false;
                 var phonemeTasks = l.MResultsCollection
@@ -303,17 +303,17 @@ await Dispatcher.UIThread.InvokeAsync(async () =>
 
                             if (UseWordDivider && (phoneme == null || phoneme.Trim() == string.Empty))
                             {
-                                IPAPhonemes.Add(" ");
+                                IPAPhonemes[index] = " ";
                             }
                             else
                             {
                                 if (index == 0)
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), true));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), true);
                                 }
                                 else
                                 {
-                                    IPAPhonemes.Add(IPAConverter.ConvertToIPA(phoneme.Trim(), false));
+                                    IPAPhonemes[index] = IPAConverter.ConvertToIPA(phoneme.Trim(), false);
                                 }
                             }
 

diff --git a/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/EnglishUSPhonemizer.cs b/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/EnglishUSPhonemizer.cs
@@ -10,12 +10,13 @@ namespace Mirivoice.Mirivoice.Plugins.Builtin.Phonemizers
     public class EnglishUSPhonemizer : BasePhonemizer
     {
         public override BaseIPAConverter IPAConverter { get; set; } = new EnglishUSIPAConverter();
-        public override bool UseWordDivider { get; set; } = true;
+        public override bool UseWordDivider { get; set; } = false;
 
         protected override string[] SplitToWords(string sentence)
         {
             List<string> words = new List<string>();
             StringBuilder sb = new StringBuilder();
+            int index = 0;
             foreach (string word in sentence.Split())
             {
                 char[] charArr = word.ToCharArray();
@@ -44,8 +45,14 @@ protected override string[] SplitToWords(string sentence)
                     }
 
                 }
+
                 words.Add(sb.ToString());
                 sb.Clear();
+
+                ++index;
+                if (index != sentence.Split().Length ) { 
+                    words.Add(" ");
+                }
             }
             return words.ToArray();
         }

diff --git a/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/Utils/EnglishUSPhonemizerUtil.cs b/Mirivoice/Mirivoice.Plugins.Builtin/Phonemizers/Utils/EnglishUSPhonemizerUtil.cs
@@ -23,7 +23,7 @@ public static string WordToArpabet(string word)
         {
             // use the CMU Pronouncing Dictionary to convert words to ARPAbet
             string arpabetRes = cmuDict.CMUDict.TryGetValue(word.ToLower(), out string arpabet) ? arpabet.ToLower() : word;
-            //Log.Debug($"WordToArpabet: {word} -> {arpabetRes}");
+            Log.Debug($"WordToArpabet: {word} -> {arpabetRes}");
             return arpabetRes;
 
         }

diff --git a/Mirivoice/Views/DataPreprocessWindow.axaml b/Mirivoice/Views/DataPreprocessWindow.axaml
@@ -14,6 +14,7 @@
 				<TextBlock Text="{DynamicResource menu.tools.voicerMakingSupport.preprocessor.langCode}" VerticalAlignment="Center" FontSize="15" Margin="10 5 5 5"/>
 				<ComboBox x:Name="langSelectCombobox" Margin="15 10 5 5" SelectedIndex="0" VerticalAlignment="Center" SelectionChanged="LangChanged">
 					<ComboBoxItem Content="ko" />
+					<ComboBoxItem Content="en-US" />
 				</ComboBox>
 			</StackPanel>
 

diff --git a/Mirivoice/Views/DataPreprocessWindow.axaml.cs b/Mirivoice/Views/DataPreprocessWindow.axaml.cs
@@ -33,6 +33,8 @@ BasePhonemizer GetPhonemizerWithIndex()
         {
             case 0:
                 return LineBoxViewModel.GetPhonemizer("ko");
+            case 1:
+                return LineBoxViewModel.GetPhonemizer("en-US");
             default:
                 return new DefaultPhonemizer();
         }