Mirrors/SubtitleEdit
1
0
Fork 0
mirror of https://github.com/SubtitleEdit/subtitleedit.git synced 2024-11-22 03:02:35 +01:00
Code Issues Projects Releases Wiki Activity

Try to fix #8629

Browse Source
This commit is contained in:
Nikolaj Olsson 2024-07-16 20:12:09 +02:00
parent 0620bc98c4
commit c9b4c29acb
2 changed files with 12 additions and 51 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/Test/FixCommonErrors/FixCommonErrorsTest.cs
View File

@ -596,6 +596,17 @@ namespace Test.FixCommonErrors
}
}
[TestMethod]
public void FixCommonOcrErrorsFrenchHardCodedRuleNoChange5()
{
using (var target = GetFixCommonErrorsLib())
{
InitializeFixCommonErrorsLine(target, "À SÉQUENCER l'ADN DE MICRO-ORGANISMES ÉTEINTS");
target.FixOcrErrorsViaReplaceList("fra");
Assert.AreEqual("À SÉQUENCER L'ADN DE MICRO-ORGANISMES ÉTEINTS", target.Subtitle.Paragraphs[0].Text);
}
}
#endregion Fix OCR errors
#region Fix missing spaces

52
src/ui/Logic/Ocr/OcrFixEngine.cs
View File

@ -663,43 +663,6 @@ namespace Nikse.SubtitleEdit.Logic.Ocr
return word;
}
private static bool IsToKeepCasing(string sentence)
{
// related to https://github.com/SubtitleEdit/subtitleedit/issues/8052
if (sentence.Length > 2 && IsCurrentWordUpperCase(sentence))
{
// do not change 'L' to lowercase in text like "L'ASSASSIN"
return char.IsUpper(sentence[2]) && (sentence[1] == '\'' || sentence[1] == '’');
}
return false;
}
private static bool IsCurrentWordUpperCase(string sentence)
{
var maxMinusOne = sentence.Length - 1;
for (var i = 2; i <= maxMinusOne; i++)
{
var ch = sentence[i];
if (!char.IsLetter(ch))
{
return true;
}
if (!char.IsUpper(ch))
{
return false;
}
if (i == maxMinusOne)
{
return true;
}
}
return false;
}
public static string FixFrenchLApostrophe(string input, string affix, string prevLine)
{
var text = input;
@ -731,22 +694,9 @@ namespace Nikse.SubtitleEdit.Logic.Ocr
var start = text.IndexOf(affix, StringComparison.Ordinal);
while (start > 0)
{
prevLine = HtmlUtil.RemoveHtmlTags(text.Substring(0, start)).TrimEnd().TrimEnd('-').TrimEnd();
isPreviousLineClose = string.IsNullOrEmpty(prevLine) || prevLine.EndsWith('.') || prevLine.EndsWith('!') || prevLine.EndsWith('?');
if (start < text.Length - 4)
{
if (start == 1 && text.StartsWith('-'))
{
isPreviousLineClose = true;
}
if (start > 1)
{
var beforeThis = HtmlUtil.RemoveHtmlTags(text.Substring(0, start));
isPreviousLineClose = beforeThis.EndsWith('.') || beforeThis.EndsWith('!') || beforeThis.EndsWith('?');
}
if (isPreviousLineClose || IsToKeepCasing(text.Substring(start + 1)))
if (text.Replace("l'", "L'") == text.ToUpperInvariant())
{
text = text.Remove(start + 1, 1).Insert(start + 1, "L");
}