Updated Thai with new characters

dmort27 · Jun 6, 2022 · a68d73b · a68d73b
1 parent d0ada5a
commit a68d73b
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 0 deletions.
diff --git a/epitran/data/map/tha-Thai.csv b/epitran/data/map/tha-Thai.csv
@@ -74,3 +74,6 @@ Orth,Phon
 ัวะ,ua̯
 ัว,uːa̯
 ʔ,ʔ
+ๅ,ː
+ฤ,rʉ
+ำ,am
diff --git a/epitran/data/pre/tha-Thai.txt b/epitran/data/pre/tha-Thai.txt
@@ -10,3 +10,37 @@
 
 % Thanthakhat (cp. Virama)
 .์ -> 0 / _
+
+% Delete numerals
+๐ -> 0 / _ % THAI DIGIT ZERO
+๑ -> 0 / _ % THAI DIGIT ONE
+๒ -> 0 / _ % THAI DIGIT TWO
+๓ -> 0 / _ % THAI DIGIT THREE
+๔ -> 0 / _ % THAI DIGIT FOUR
+๕ -> 0 / _ % THAI DIGIT FIVE
+๖ -> 0 / _ % THAI DIGIT SIX
+๗ -> 0 / _ % THAI DIGIT SEVEN
+๘ -> 0 / _ % THAI DIGIT EIGHT
+๙ -> 0 / _ % THAI DIGIT NINE
+
+% Delete pintu
+ฺ -> 0 / _ % THAI CHARACTER PHINTHU
+
+% Delete tones
+่ -> 0 / _ % THAI CHARACTER MAI EK
+๋ -> 0 / _ % THAI CHARACTER MAI CHATTAWA
+๊ -> 0 / _ % THAI CHARACTER MAI TRI
+
+% Delete archaic and exceptional
+ํ -> 0 / _ % THAI CHARACTER NIKHAHIT
+์ -> 0 / _ % THAI CHARACTER THANTHAKHAT
+
+
+% Delete reduplication mark
+ๆ -> 0 / _ % THAI CHARACTER MAIYAMOK
+
+% Delete abbreviation marker
+ฯ -> 0 / _ % THAI CHARACTER PAIYANNOI
+
+% Delete short mark (should be handled differently)
+็ -> 0 / _ % THAI CHARACTER MAITAIKHU
-Original file line number
+Diff line change
@@ Expand Up / @@ -74,3 +74,6 @@ Orth,Phon @@
     ัวะ,ua̯
     ัว,uːa̯
     ʔ,ʔ
+    ๅ,ː
+    ฤ,rʉ
+    ำ,am