Vietnamese optimizations

* Simplified the layout to reduce the number of suggestions per key and to be more consistent with the original Nokia way

	* a migration to enforce the new layout

app/languages/definitions/Vietnamese.yml

@@ -4,11 +4,11 @@ dictionaryFile: vi-utf8.csv
 layout:
   - [SPECIAL] # 0
   - [PUNCTUATION] # 1
-  - [a, b, c, ă, â, á, ắ, ấ, à, ằ, ầ, ả, ẳ, ẩ, ã, ẵ, ẫ, ạ, ặ, ậ] # 2
-  - [d, e, f, đ, ê, é, ế, è, ề, ẻ, ể, ẽ, ễ, ẹ, ệ] # 3
-  - [g, h, i, í, ì, ỉ, ĩ, ị] # 4
-  - [j, k, l] # 5
-  - [m, n, o, ô, ơ, ó, ố, ớ, ò, ồ, ờ, ỏ, ổ, ở, õ, ỗ, ỡ, ọ, ộ, ợ] # 6
+  - [a, ă, â, b, c, ́] # 2
+  - [d, đ, e, ê, f, ̀] # 3
+  - [g, h, i, ̉] # 4
+  - [j, k, l, ̃] # 5
+  - [m, n, o, ô, ơ, ̣] # 6
   - [p, q, r, s] # 7
-  - [t, u, v, ư, ú, ứ, ù, ừ, ủ, ử, ũ, ữ, ụ, ự] # 8
-  - [w, x, y, z, ý, ỳ, ỷ, ỹ, ỵ] # 9
+  - [t, u, ư, v] # 8
+  - [w, x, y, z] # 9

app/src/main/java/io/github/sspanak/tt9/db/sqlite/Migration.java

@@ -21,6 +21,11 @@ class Migration {
 				" SET maxWordsPerSequence = " + SettingsStore.SUGGESTIONS_POSITIONS_LIMIT +
 				", fileHash = '0'",
 			832
+		),
+		new Migration(
+			// enforce the new Vietnamese layout
+			"DELETE FROM " + Tables.LANGUAGES_META + " WHERE langId = 481590",
+			952
 		)
 	};
 

app/src/main/java/io/github/sspanak/tt9/ime/modes/predictions/LocaleWordsSorter.java

@@ -13,11 +13,14 @@ import io.github.sspanak.tt9.languages.LanguageKind;
 
 
 class LocaleWordsSorter {
+	private final Language language;
 	private final Pattern sortingPattern;
 
 
 	LocaleWordsSorter(@Nullable Language language) {
-		sortingPattern = LanguageKind.isIndic(language) ? Pattern.compile("\\p{L}\\p{M}+") : null;
+		this.language = language;
+		boolean isAlphabetWithModifiers = LanguageKind.isIndic(language) || LanguageKind.isVietnamese(language);
+		sortingPattern = isAlphabetWithModifiers ? Pattern.compile("\\p{L}\\p{M}+") : null;
 	}
 
 
@@ -49,7 +52,9 @@ class LocaleWordsSorter {
 	}
 
 
-	boolean shouldSort(@Nullable Language language, @NonNull String stem, @NonNull String digitSequence) {
-		return LanguageKind.isIndic(language) && !stem.isEmpty() && stem.length() == digitSequence.length() - 1;
+	boolean shouldSort(@NonNull String stem, @NonNull String digitSequence) {
+		return
+			(LanguageKind.isIndic(language) && !stem.isEmpty() && stem.length() == digitSequence.length() - 1)
+			|| LanguageKind.isVietnamese(language);
 	}
 }

app/src/main/java/io/github/sspanak/tt9/ime/modes/predictions/WordPredictions.java

@@ -101,7 +101,7 @@ public class WordPredictions extends Predictions {
 			words.addAll(dbWords);
 		} else {
 			suggestStem();
-			dbWords = localeWordsSorter.shouldSort(language, stem, digitSequence) ? localeWordsSorter.sort(dbWords) : dbWords;
+			dbWords = localeWordsSorter.shouldSort(stem, digitSequence) ? localeWordsSorter.sort(dbWords) : dbWords;
 			dbWords = rearrangeByPairFrequency(dbWords);
 			suggestMissingWords(generatePossibleStemVariations(dbWords));
 			suggestMissingWords(dbWords.isEmpty() ? generateWordVariations(inputWord) : dbWords);

app/src/main/java/io/github/sspanak/tt9/languages/LanguageKind.java

@@ -20,4 +20,5 @@ public class LanguageKind {
 	public static boolean isIndic(Language language) { return isGujarati(language) || isHindi(language); }
 	public static boolean isKorean(Language language) { return language != null && language.getId() == KOREAN; }
 	public static boolean isUkrainian(Language language) { return language != null && language.getId() == 54645; }
+	public static boolean isVietnamese(Language language) { return language != null && language.getId() == 481590; }
 }