Mirrors/SubtitleEdit
1
0
Fork 0
mirror of https://github.com/SubtitleEdit/subtitleedit.git synced 2024-10-27 14:32:35 +01:00
Code Issues Projects Releases Wiki Activity

Updated Serbian OCRFixReplaceList

Browse Source
This commit is contained in:
Waldi Ravens 2015-06-24 14:05:26 +02:00
parent 8e012f3d71
commit 82a047cae7
1 changed files with 13 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
Dictionaries/srp_OCRFixReplaceList.xml
View File

@ -37,7 +37,7 @@
<WholeLines />
<RegularExpressions>
<RegEx find="([a-zčđšž])Ij([a-zčđšž])" replaceWith="$1lj$2" />
<RegEx find="([^A-ZÐČŠŽa-zčđšž])Ij(ubav|ud|ut)" replaceWith="$1lj$2" />
<RegEx find="([^A-ZČĐŠŽa-zčđšž])Ij(ubav|ud|ut)" replaceWith="$1lj$2" />
<!-- kad ima razmak izmedju tagova </i> <i> -->
<RegEx find="(&gt;) +(&lt;)" replaceWith="$1$2" />
<!-- ',"' to '",' -->
@ -65,19 +65,19 @@
<RegEx find="([0-9])\.O" replaceWith="$1.0" />
<RegEx find="\b0([A-Za-z])" replaceWith="O$1" />
<!-- Brisanje crte - na početku 1. reda (i kada ima dva reda) -->
<RegEx find="\A- ?([A-ZÐČŠŽa-zčđšž0-9„'&quot;]|\.{3})" replaceWith="$1" />
<RegEx find="\A- ?([A-ZČĐŠŽa-zčđšž0-9„'&quot;]|\.{3})" replaceWith="$1" />
<RegEx find="\A(&lt;[i|b|u]&gt;)- ?" replaceWith="$1" />
<RegEx find=" - " replaceWith=" -" />
<!-- Brisanje razmaka iza crte - na početku 2. reda -->
<RegEx find="([\n](&lt;[i|b|u]&gt;)?)- ([0-9A-ZÐČ芚Ž„'&quot;&lt;])" replaceWith="$1-$3" />
<RegEx find="([\n](&lt;[i|b|u]&gt;)?)- ([0-9A-ZČĐŠŽčš„'&quot;&lt;])" replaceWith="$1-$3" />
<!-- Korigovanje crte - kad je u sredini prvog reda -->
<RegEx find="([.!?&quot;&gt;]) - ([A-ZÐČ芚Ž'&quot;&lt;])" replaceWith="$1 -$2" />
<RegEx find="([.!?&quot;&gt;]) - ([A-ZČĐŠŽčš'&quot;&lt;])" replaceWith="$1 -$2" />
<!-- Zatvoren tag pa razmak poslije crtice -->
<RegEx find="(&gt;) - ([A-ZÐČ芚Ž„'&quot;])" replaceWith="$1 -$2" />
<RegEx find="(&gt;) - ([A-ZČĐŠŽčš„'&quot;])" replaceWith="$1 -$2" />
<!-- Zatvoren tag pa crtica razmak -->
<RegEx find="(&gt;)- ([A-ZÐČ芚Ž„'&quot;])" replaceWith="$1-$2" />
<RegEx find="(&gt;)- ([A-ZČĐŠŽčš„'&quot;])" replaceWith="$1-$2" />
<!-- Zagrada pa crtica razmak -->
<RegEx find="\(- ([A-ZÐČ芚Ž„'&quot;])" replaceWith="(-$1" />
<RegEx find="\(- ([A-ZČĐŠŽčš„'&quot;])" replaceWith="(-$1" />
<!-- Smart space after dot -->
<!-- osim kad je zadnje t (rijec kolt) -->
<RegEx find="([a-su-zá-úñä-ü])\.([^\s\n().:?!*^“”'&quot;&lt;])" replaceWith="$1. $2" />
@ -87,9 +87,9 @@
<!-- Joey(j)a -->
<RegEx find="([A-Z][a-z])eyj([a-z])" replaceWith="$1ey$2" />
<!-- Sređuje zarez sa razmakom -->
<RegEx find="([A-ZÐČŠŽa-zčđšžá-úñä-ü&quot;]),([^\s\n(),?!“&lt;])" replaceWith="$1, $2" />
<RegEx find="([A-ZČĐŠŽa-zčđšžá-úñä-ü&quot;]),([^\s\n(),?!“&lt;])" replaceWith="$1, $2" />
<RegEx find=" , " replaceWith=", " />
<RegEx find=" ,([A-ZÐČŠŽa-zčđšž])" replaceWith=", $1" />
<RegEx find=" ,([A-ZČĐŠŽa-zčđšž])" replaceWith=", $1" />
<RegEx find=" ,$" replaceWith="," />
<RegEx find="([?!])-" replaceWith="$1 -" />
<!-- Space after last of some consecutive dots (eg. "...") -->
@ -97,17 +97,17 @@
<!-- Delete space after "..." that is at the beginning of the line. You may delete this line if you don't like it -->
<!-- <RegEx find="^\.{3} +" replaceWith="..." /> -->
<!-- "tekst ... tekst" mijenja u "tekst... tekst" -->
<RegEx find="([A-ZÐČŠŽa-zčđšž]) \.{3} " replaceWith="$1... " />
<RegEx find="([A-ZČĐŠŽa-zčđšž]) \.{3} " replaceWith="$1... " />
<RegEx find="(\S)\. &quot;" replaceWith="$1.&quot;" />
<RegEx find="&quot; \." replaceWith="&quot;." />
<RegEx find="(\S\.{3}) &quot;(\s|$)" replaceWith="$1&quot;$2" />
<RegEx find=" \.{3}$" replaceWith="..." />
<RegEx find="([a-zčđšž])( \.{3}|\.{2}$)" replaceWith="$1..." />
<!-- Razmak ispred zagrade -->
<RegEx find="([A-ZÐČŠŽa-zčđšž])\(" replaceWith="$1 (" />
<RegEx find="([A-ZČĐŠŽa-zčđšž])\(" replaceWith="$1 (" />
<!-- Razmak iza upitnika -->
<RegEx find="\?([A-ZÐČ芚Ž])" replaceWith="? $1" />
<RegEx find="(^|&gt;)\.{3} ([A-ZÐČ芚Ž])" replaceWith="$1...$2" />
<RegEx find="\?([A-ZČĐŠŽčš])" replaceWith="? $1" />
<RegEx find="(^|&gt;)\.{3} ([A-ZČĐŠŽčš])" replaceWith="$1...$2" />
<!-- Brise ... kad je na poc. reda "... -->
<RegEx find="^&quot;\.{3} " replaceWith="&quot;" />
<RegEx find="([0-9])\$" replaceWith="$1 $" />