Fix to swap Sinhala Diga Is-pilla and Rakaaraansaya.

PiperOrigin-RevId: 578932892
google-research · Nov 2, 2023 · 4f0fc1e · 4f0fc1e
1 parent 61df30a
commit 4f0fc1e
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 0 deletions.
diff --git a/nisaba/scripts/brahmic/data/Sinh/visual_rewrite.textproto b/nisaba/scripts/brahmic/data/Sinh/visual_rewrite.textproto
@@ -93,6 +93,11 @@ item {
   to_uname: ["AL-LAKUNA", "ZERO WIDTH JOINER", "RAYANNA", "KETTI IS-PILLA"] to_raw: "්‍රි"
 }
 
+item {
+  uname: ["DIGA IS-PILLA", "ZERO WIDTH JOINER", "AL-LAKUNA", "ZERO WIDTH JOINER", "RAYANNA"] raw: "ී‍්‍ර"
+  to_uname: ["AL-LAKUNA", "ZERO WIDTH JOINER", "RAYANNA", "DIGA IS-PILLA"] to_raw: "්‍රී"
+}
+
 item {
   uname: ["DIGA KOMBUVA", "ZERO WIDTH JOINER", "AL-LAKUNA", "ZERO WIDTH JOINER", "RAYANNA"] raw: "ේ‍්‍ර"
   to_uname: ["AL-LAKUNA", "ZERO WIDTH JOINER", "RAYANNA", "DIGA KOMBUVA"] to_raw: "්‍රේ"

diff --git a/nisaba/scripts/brahmic/testdata/visual_norm.textproto b/nisaba/scripts/brahmic/testdata/visual_norm.textproto
@@ -32,6 +32,7 @@ rewrite { rule: "SINH" input: "ශ්‍රීලං‌කා‍" output: "
 
 rewrite { rule: "SINH" input: "ඊශ‍්‍රායල" output: "ඊශ්‍රායල" }
 rewrite { rule: "SINH" input: "බි‍්‍රතාන්‍යයේ" output: "බ්‍රිතාන්‍යයේ" }
+rewrite { rule: "SINH" input: "මන්තී‍්‍ර" output: "මන්ත්‍රී" }
 rewrite { rule: "SINH" input: "අපේ‍්‍රල්" output: "අප්‍රේල්" }
 
 # TODO: Currently, ‘preserve’ is only available when the listed sequence