Improved greek transliteration

hmlendea · Jan 26, 2022 · c8db73f · c8db73f
1 parent c830a90
commit c8db73f
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 2 deletions.
diff --git a/Service/Transliterators/TransliterateDotComTransliterator.cs b/Service/Transliterators/TransliterateDotComTransliterator.cs
@@ -41,9 +41,10 @@ private string ApplyFixes(string text, string language)
                 fixedText = Regex.Replace(fixedText, "ō̂", "ô");
 
                 fixedText = Regex.Replace(fixedText, "Ch", "Kh");
+                fixedText = Regex.Replace(fixedText, "R([^\\s])", "Rh$1");
 
                 fixedText = Regex.Replace(fixedText, "ch", "kh");
-                fixedText = Regex.Replace(fixedText, "rh", "r");
+                fixedText = Regex.Replace(fixedText, "([^\\s])rh", "$1r");
 
                 fixedText = Regex.Replace(fixedText, "Mp", "V");
                 fixedText = Regex.Replace(fixedText, "Nk", "G");
@@ -59,6 +60,7 @@ private string ApplyFixes(string text, string language)
                 fixedText = Regex.Replace(fixedText, "([Rr])(a|ṓ)", "$1h$2");
                 fixedText = Regex.Replace(fixedText, "([Rr])h([a])", "$1$2");
 
+                fixedText = Regex.Replace(fixedText, "Agí(a|os)", "Hagí$1");
                 fixedText = Regex.Replace(fixedText, "polē$", "polis");
                 fixedText = Regex.Replace(fixedText, "polē ", "polis");
             }

diff --git a/Service/Transliterators/TranslitterationDotComTransliterator.cs b/Service/Transliterators/TranslitterationDotComTransliterator.cs
@@ -58,11 +58,17 @@ private string ApplyLanguageSpecificFixes(string text, string language)
                 fixedText = Regex.Replace(fixedText, "ήr([iíί])$", "éri");
 
                 fixedText = Regex.Replace(fixedText, "([a-zA-Z])H", "$1h");
+
+
+                fixedText = Regex.Replace(fixedText, "[Ά]", "Á");
+                fixedText = Regex.Replace(fixedText, "[Έ]", "É");
+                fixedText = Regex.Replace(fixedText, "[Ό]", "Ó");
+
                 fixedText = Regex.Replace(fixedText, "ᾶ", "a");
                 fixedText = Regex.Replace(fixedText, "ά", "á");
                 fixedText = Regex.Replace(fixedText, "[έ]", "é");
                 fixedText = Regex.Replace(fixedText, "[ίή]", "í");
-                fixedText = Regex.Replace(fixedText, "ώ", "ó");
+                fixedText = Regex.Replace(fixedText, "[όώ]", "ó");
                 fixedText = Regex.Replace(fixedText, "ς", "s");
                 fixedText = Regex.Replace(fixedText, "ύ", "ú");