Mirrors/SubtitleEdit
1
0
Fork 0
mirror of https://github.com/SubtitleEdit/subtitleedit.git synced 2024-11-25 04:33:04 +01:00
Code Issues Projects Releases Wiki Activity

Fix #7275

Browse Source
This commit is contained in:
niksedk 2023-08-22 16:40:39 +02:00
parent 32524d27de
commit 4fe20da148
3 changed files with 22 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
src/libse/Common/LanguageAutoDetect.cs
View File

@ -1886,16 +1886,13 @@ namespace Nikse.SubtitleEdit.Core.Common
return false;
}
public static string KoreanLetters = "가나다라마바사아자차카타파하아야어여오요우유으이대한민국활화산동물들천국의섬유독춤을춥니다";
public static string JapaneseLetters = "あいうえおかきくけこがぎぐげごさしすせそざじずぜぞたちつてとだぢづでどなにぬねのはひふへほばびぶべぼぱぴぷぺぽまみむめもやゆよらりるれろわをん";
public static string GetEncodingViaLetter(string text)
{
var dictionary = new Dictionary<string, int>();
var dictionary = new Dictionary<string, int>(6);
// Arabic
int count = 0;
foreach (var letter in "غظضذخثتشرقصفعسنملكيطحزوهدجبا")
var count = 0;
foreach (var letter in Alphabets.Arabic)
{
if (text.Contains(letter))
{
@ -1906,7 +1903,7 @@ namespace Nikse.SubtitleEdit.Core.Common
// Korean
count = 0;
foreach (var letter in KoreanLetters)
foreach (var letter in Alphabets.Korean)
{
if (text.Contains(letter))
{
@ -1917,7 +1914,7 @@ namespace Nikse.SubtitleEdit.Core.Common
// Japanese
count = 0;
foreach (var letter in JapaneseLetters)
foreach (var letter in Alphabets.Japanese)
{
if (text.Contains(letter))
{
@ -1928,7 +1925,7 @@ namespace Nikse.SubtitleEdit.Core.Common
// Thai
count = 0;
foreach (var letter in "กขฃคฅฆงจฉชซฌญฎฏฐฑฒณดตถทธนบปผฝพฟภมยรลวศษสหฬอฮ")
foreach (var letter in Alphabets.Thai)
{
if (text.Contains(letter))
{
@ -1939,7 +1936,7 @@ namespace Nikse.SubtitleEdit.Core.Common
// Sinhalese
count = 0;
foreach (var letter in "අආඇඈඉඊඋඌඍඎඏඐඑඒඓඔඕඖකඛගඝඞඟචඡජඣඤඥඦටඨඩඪණඬතථදධනඳපඵබභමඹයරලවශෂසහළෆ්ාැෑිීුූෘෙේෛොෝෞෟ෦෧෨෩෪෫෬෭෮෯")
foreach (var letter in Alphabets.Sinhalese)
{
if (text.Contains(letter))
{
@ -1950,7 +1947,7 @@ namespace Nikse.SubtitleEdit.Core.Common
// Urdu
count = 0;
foreach (var letter in "ﺍﺎﺁﺂﺏﺑﺒﺐﭖﭘﭙﭗﺕﺗﺘﺖﭦﭨﭩﭧﺙﺛﺜﺚﺝﺟﺠﺞﭺﭼﭽﭻﺡﺣﺤﺢﺥﺧﺨﺦﺩﺪﺫﺬﺭﺮﮌﮍﺯﺰﮊﮋﺱﺳﺴﺲﺵﺷﺸﺶﺹﺻﺼﺺﺽﺿﻀﺾﻁﻃﻄﻂﻅﻇﻈﻆﻉﻋﻌﻊﻍﻏﻐﻎﻑﻓﻔﻒﻕﻗﻘﻖﻙﻛﻜﻚﻻﻼﻝﻟﻠﻞﻡﻣﻤﻢﻥﻧﻨﻦﻭﻮﮮﮯﮦﮨﮩﮧﯼﯾﯿﯽﮪﮬﮭﮫﴽﴼﺀﺋﺌﹱﹷﹹ")
foreach (var letter in Alphabets.Urdu)
{
if (text.Contains(letter))
{
@ -1967,5 +1964,15 @@ namespace Nikse.SubtitleEdit.Core.Common
{
return language == "ko" || language == "zh" || language == "ja" || language == "th";
}
public static class Alphabets
{
public const string Arabic = "غظضذخثتشرقصفعسنملكيطحزوهدجبا";
public const string Korean = "가나다라마바사아자차카타파하아야어여오요우유으이대한민국활화산동물들천국의섬유독춤을춥니다";
public const string Japanese = "あいうえおかきくけこがぎぐげごさしすせそざじずぜぞたちつてとだぢづでどなにぬねのはひふへほばびぶべぼぱぴぷぺぽまみむめもやゆよらりるれろわをん";
public const string Thai = "กขฃคฅฆงจฉชซฌญฎฏฐฑฒณดตถทธนบปผฝพฟภมยรลวศษสหฬอฮ";
public const string Urdu = "ﺍﺎﺁﺂﺏﺑﺒﺐﭖﭘﭙﭗﺕﺗﺘﺖﭦﭨﭩﭧﺙﺛﺜﺚﺝﺟﺠﺞﭺﭼﭽﭻﺡﺣﺤﺢﺥﺧﺨﺦﺩﺪﺫﺬﺭﺮﮌﮍﺯﺰﮊﮋﺱﺳﺴﺲﺵﺷﺸﺶﺹﺻﺼﺺﺽﺿﻀﺾﻁﻃﻄﻂﻅﻇﻈﻆﻉﻋﻌﻊﻍﻏﻐﻎﻑﻓﻔﻒﻕﻗﻘﻖﻙﻛﻜﻚﻻﻼﻝﻟﻠﻞﻡﻣﻤﻢﻥﻧﻨﻦﻭﻮﮮﮯﮦﮨﮩﮧﯼﯾﯿﯽﮪﮬﮭﮫﴽﴼﺀﺋﺌﹱﹷﹹ";
public const string Sinhalese = "අආඇඈඉඊඋඌඍඎඏඐඑඒඓඔඕඖකඛගඝඞඟචඡජඣඤඥඦටඨඩඪණඬතථදධනඳපඵබභමඹයරලවශෂසහළෆ්ාැෑිීුූෘෙේෛොෝෞෟ෦෧෨෩෪෫෬෭෮෯";
}
}
}

4
src/libse/Common/TextLengthCalculator/CalcCJK.cs
View File

@ -62,8 +62,8 @@ namespace Nikse.SubtitleEdit.Core.Common.TextLengthCalculator
count += 0.5m;
}
else if (ChineseFullWidthPunctuations.Contains(ch) ||
LanguageAutoDetect.JapaneseLetters.Contains(ch) ||
LanguageAutoDetect.KoreanLetters.Contains(ch) ||
LanguageAutoDetect.Alphabets.Japanese.Contains(ch) ||
LanguageAutoDetect.Alphabets.Korean.Contains(ch) ||
IsCjk(ch))
{
count++;

4
src/libse/Common/TextLengthCalculator/CalcCJKNoSpace.cs
View File

@ -61,8 +61,8 @@
count += 0.5m;
}
else if (CalcCjk.ChineseFullWidthPunctuations.Contains(ch) ||
LanguageAutoDetect.JapaneseLetters.Contains(ch) ||
LanguageAutoDetect.KoreanLetters.Contains(ch) ||
LanguageAutoDetect.Alphabets.Japanese.Contains(ch) ||
LanguageAutoDetect.Alphabets.Korean.Contains(ch) ||
CalcCjk.IsCjk(ch))
{
count++;