Mirrors/SubtitleEdit
1
0
Fork 0
mirror of https://github.com/SubtitleEdit/subtitleedit.git synced 2024-10-27 22:42:38 +01:00
Code Issues Projects Releases Wiki Activity

Minor fix in Ocr-fix-engine - thx Majid :)

Browse Source 
git-svn-id: https://subtitleedit.googlecode.com/svn/trunk@891 99eadd0c-20b8-1223-b5c4-2a2b2df33de2
This commit is contained in:
niksedk 2012-01-04 20:22:10 +00:00
parent 96a80936da
commit 8fb55a2e76
1 changed files with 143 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

147
src/Logic/OCR/OcrFixEngine.cs
View File

@ -295,6 +295,8 @@ namespace Nikse.SubtitleEdit.Logic.OCR
var sb = new StringBuilder();
var word = new StringBuilder();
text = ReplaceWordsBeforeLineFixes(text);
text = FixCommenOcrLineErrors(text, lastLine);
string lastWord = null;
@ -306,13 +308,9 @@ namespace Nikse.SubtitleEdit.Logic.OCR
{
string fixedWord;
if (lastWord != null && lastWord.ToUpper().Contains("COLOR="))
{
fixedWord = word.ToString();
}
else
{
fixedWord = FixCommonWordErrors(word.ToString(), lastWord);
}
sb.Append(fixedWord);
lastWord = fixedWord;
word = new StringBuilder();
@ -347,6 +345,41 @@ namespace Nikse.SubtitleEdit.Logic.OCR
return text;
}
private string ReplaceWordsBeforeLineFixes(string text)
{
string lastWord = null;
var sb = new StringBuilder();
var word = new StringBuilder();
for (int i = 0; i < text.Length; i++)
{
if (" ¡¿,.!?:;()[]{}+-£\"#&%\r\n".Contains(text[i].ToString())) // removed $
{
if (word.Length > 0)
{
string fixedWord;
if (lastWord != null && lastWord.ToUpper().Contains("COLOR="))
fixedWord = word.ToString();
else
fixedWord = FixCommonWordErrorsQuick(word.ToString(), lastWord);
sb.Append(fixedWord);
lastWord = fixedWord;
word = new StringBuilder();
}
sb.Append(text[i]);
}
else
{
word.Append(text[i]);
}
}
if (word.Length > 0) // last word
{
string fixedWord = FixCommonWordErrorsQuick(word.ToString(), lastWord);
sb.Append(fixedWord);
}
return sb.ToString();
}
private string RemoveSpaceBetweenNumbers(string text)
{
Match match = regexSpaceBetweenNumbers.Match(text);
@ -528,6 +561,111 @@ namespace Nikse.SubtitleEdit.Logic.OCR
}
}
return pre + word + post;
}
private string FixCommonWordErrorsQuick(string word, string lastWord)
{
//always replace list
foreach (string letter in _partialWordReplaceListAlways.Keys)
word = word.Replace(letter, _partialWordReplaceListAlways[letter]);
string pre = string.Empty;
string post = string.Empty;
if (word.StartsWith("<i>"))
{
pre += "<i>";
word = word.Remove(0, 3);
}
while (word.StartsWith(Environment.NewLine) && word.Length > 2)
{
pre += Environment.NewLine;
word = word.Substring(2);
}
while (word.StartsWith("-") && word.Length > 1)
{
pre += "-";
word = word.Substring(1);
}
while (word.StartsWith(".") && word.Length > 1)
{
pre += ".";
word = word.Substring(1);
}
while (word.StartsWith("\"") && word.Length > 1)
{
pre += "\"";
word = word.Substring(1);
}
if (word.StartsWith("(") && word.Length > 1)
{
pre += "(";
word = word.Substring(1);
}
if (word.StartsWith("<i>"))
{
pre += "<i>";
word = word.Remove(0, 3);
}
while (word.EndsWith(Environment.NewLine) && word.Length > 2)
{
post += Environment.NewLine;
word = word.Substring(0, word.Length - 2);
}
while (word.EndsWith("\"") && word.Length > 1)
{
post = post + "\"";
word = word.Substring(0, word.Length - 1);
}
while (word.EndsWith(".") && word.Length > 1)
{
post = post + ".";
word = word.Substring(0, word.Length - 1);
}
while (word.EndsWith(",") && word.Length > 1)
{
post = post + ",";
word = word.Substring(0, word.Length - 1);
}
while (word.EndsWith("?") && word.Length > 1)
{
post = post + "?";
word = word.Substring(0, word.Length - 1);
}
while (word.EndsWith("!") && word.Length > 1)
{
post = post + "!";
word = word.Substring(0, word.Length - 1);
}
while (word.EndsWith(")") && word.Length > 1)
{
post = post + ")";
word = word.Substring(0, word.Length - 1);
}
if (word.EndsWith("</i>"))
{
post = post + "</i>";
word = word.Remove(word.Length - 4, 4);
}
if (word.Length == 0)
return pre + word + post;
foreach (string from in _wordReplaceList.Keys)
{
if (from.Contains(word))
{
if (word == from)
return pre + _wordReplaceList[from] + post;
if (word + post == from)
return pre + _wordReplaceList[from];
if (pre + word + post == from)
return _wordReplaceList[from];
}
}
return pre + word + post;
}
@ -596,6 +734,7 @@ namespace Nikse.SubtitleEdit.Logic.OCR
return word;
}
public static string FixIor1InsideLowerCaseWord(string word)
{
if (startEndEndsWithNumber.IsMatch(word))