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

Update sr_OCRFixReplaceList.xml

Browse Source
This commit is contained in:
Kruno H 2015-06-23 11:42:06 +02:00
parent 81d7b8799b
commit f525a0b2de
1 changed files with 17 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

30
Dictionaries/sr_OCRFixReplaceList.xml
View File

@ -1,6 +1,12 @@
<OCRFixReplaceList>
<!----- credit goes to: MilanRS [http://www.prijevodi-online.org] --------->
<WholeWords />
<WholeWords
<Word from="neču" to="neću" />
<Word from="nečeš" to="nećeš" />
<Word from="neče" to="neće" />
<Word from="nečemo" to="nećemo" />
<Word from="nečete" to="nećete" />
</WholeWords>
<PartialWords>
<WordPart from="¤" to="o" />
<WordPart from="vv" to="w" />
@ -18,8 +24,6 @@
<LinePart from="Jel' si ti" to="Da li si ti"/ >
<LinePart from="jel si ti" to="da li si ti"/ >
<LinePart from="jel' si ti" to="da li si ti"/ >
<LinePart from="jel' si sigur" to="Jesi li sigur"/ >
<LinePart from="jel si sigur" to="Jesi li sigur"/ >
<LinePart from="jel ste " to="jeste li "/ >
<LinePart from="Jel ste " to="Jeste li "/ >
<LinePart from="jel' ste " to="jeste li "/ >
@ -30,8 +34,8 @@
<EndLines />
<WholeLines />
<RegularExpressions>
<RegEx find="([a-zžšðš])Ij([a-zžšðš])" replaceWith="$1lj$2" />
<RegEx find="([^a-zšðčšžA-ZŠÐ蚎])Ij(ubav|ud|ut)" replaceWith="$1lj$2" />
<RegEx find="([a-zžšđš])Ij([a-zžšđš])" replaceWith="$1lj$2" />
<RegEx find="([^a-zšđčšžA-ZŠÐ蚎])Ij(ubav|ud|ut)" replaceWith="$1lj$2" />
<!-- kad ima razmak izmedju tagova </i> <i> -->
<RegEx find="(&gt;) +(&lt;)" replaceWith="$1$2" />
@ -65,7 +69,7 @@
<RegEx find="\b0([A-Za-z])" replaceWith="O$1" />
<!-- Brisanje crte - na početku 1. reda (i kada ima dva reda) -->
<RegEx find="\A- ?([A-ZŠšŽčÐa-zššžčð0-9&#132;&quot;']|\.{3})" replaceWith="$1" />
<RegEx find="\A- ?([A-ZŠšŽčÐa-zššžčđ0-9&#132;&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 -->
@ -90,23 +94,23 @@
<RegEx find="([A-Z][a-z])eyj([a-z])" replaceWith="$1ey$2" />
<!-- Sređuje zarez sa razmakom -->
<RegEx find="([A-ZŽščŠÐa-zžčššðá-úñä-ü&quot;]),([^(\s\n&lt;“,?!)])" replaceWith="$1, $2" />
<RegEx find="([A-ZŽščŠÐa-zžčššđá-úñä-ü&quot;]),([^(\s\n&lt;“,?!)])" replaceWith="$1, $2" />
<RegEx find=" , " replaceWith=", " />
<RegEx find=" ,([a-zžšðčšA-ZŠšŽčÐ])" replaceWith=", $1" />
<RegEx find=" ,([a-zžšđčšA-ZŠšŽčÐ])" replaceWith=", $1" />
<RegEx find=" ,$" replaceWith="," />
<RegEx find="([?!])-" replaceWith="$1 -" />
<!-- Space after last of some consecutive dots (eg. "...") -->
<RegEx find="([a-zšðčšž])(\.{3}|!)([a-zšðčšž])" replaceWith="$1$2 $3" />
<RegEx find="([a-zšđčšž])(\.{3}|!)([a-zšđčšž])" replaceWith="$1$2 $3" />
<!-- 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="..." /> -->
<RegEx find="([a-zžšðčšA-ZŠšŽčÐ]) \.{3} " replaceWith="$1... " /> <!-- tekst ... tekst mijenja u tekst... tekst -->
<RegEx find="([a-zžšđčšA-ZŠšŽčÐ]) \.{3} " replaceWith="$1... " /> <!-- tekst ... tekst mijenja u tekst... tekst -->
<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-zžščðš])( \.{3}|\.{2}$)" replaceWith="$1..." />
<RegEx find="([a-zžščđš])( \.{3}|\.{2}$)" replaceWith="$1..." />
<RegEx find="([a-zžšðšA-ZŽŠÐš])\(" replaceWith="$1 (" /> <!-- Razmak ispred zagrade -->
<RegEx find="([a-zžšđšA-ZŽŠÐš])\(" replaceWith="$1 (" /> <!-- Razmak ispred zagrade -->
<RegEx find="\?([A-ZŽŠÐš])" replaceWith="? $1" /> <!-- Razmak iza upitnika -->
<RegEx find="(^|&gt;)\.{3} ([A-ZŽščŠÐ])" replaceWith="$1...$2" />
<RegEx find="^&quot;\.{3} " replaceWith="&quot;" /> <!-- Brise ... kad je na poc. reda "... -->
@ -135,7 +139,7 @@
<RegEx find="d\. o\.o\." replaceWith="d.o.o." />
<!-- Kad red počinje sa ...pa malo slovo -->
<!-- <RegEx find="^\.{3}([a-zšðčšž&quot;&lt;])" replaceWith="$1" /> -->
<!-- <RegEx find="^\.{3}([a-zšđčšž&quot;&lt;])" replaceWith="$1" /> -->
<!-- <RegEx find=" ([?!.])" replaceWith="$1" /> -->
</RegularExpressions>