Mirrors/SubtitleEdit
1
0
Fork 0
mirror of https://github.com/SubtitleEdit/subtitleedit.git synced 2025-02-01 05:21:40 +01:00
Code Issues Projects Releases Wiki Activity

Add/update OCR dictionaries

Browse Source
This commit is contained in:
Nikolaj Olsson 2020-05-03 19:40:24 +02:00
parent 9608f870ed
commit 3beb5c53f4
3 changed files with 78 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Dictionaries/eng_OCRFixReplaceList.xml
View File

@ -882,6 +882,7 @@
<Word from="HeyJennifer" to="Hey Jennifer" /> <Word from="HeyJennifer" to="Hey Jennifer" />
<Word from="hiddsn" to="hidden" /> <Word from="hiddsn" to="hidden" />
<Word from="hisjunk" to="his junk" /> <Word from="hisjunk" to="his junk" />
<Word from="hisjourney" to="his journey" />
<Word from="hisseLf" to="hisself" /> <Word from="hisseLf" to="hisself" />
<Word from="Hitlershare" to="Hitler share" /> <Word from="Hitlershare" to="Hitler share" />
<Word from="Hlneed" to="I'll need" /> <Word from="Hlneed" to="I'll need" />
@ -1168,6 +1169,7 @@
<Word from="It'Il" to="It'll" /> <Word from="It'Il" to="It'll" />
<Word from="itin" to="it in" /> <Word from="itin" to="it in" />
<Word from="itis" to="it is" /> <Word from="itis" to="it is" />
<Word from="Itis" to="It is" />
<Word from="Itjust" to="It just" /> <Word from="Itjust" to="It just" />
<Word from="itjust" to="it just" /> <Word from="itjust" to="it just" />
<Word from="It'lI" to="It'll" /> <Word from="It'lI" to="It'll" />

70
Dictionaries/ita_OCRFixReplaceList.xml Normal file
View File

@ -0,0 +1,70 @@
<?xml version="1.0" encoding="utf-8"?>
<OCRFixReplaceList>
<WholeWords>
<Word from="lN" to="IN" />
<Word from="lL" to="IL" />
<Word from="l" to="I" />
</WholeWords>
<PartialWordsAlways>
<!-- Will be replaced always -->
<WordPart from="¤" to="o" />
<WordPart from="lVI" to="M" />
<WordPart from="IVl" to="M" />
<WordPart from="lVl" to="M" />
<WordPart from="I\/I" to="M" />
<WordPart from="l\/I" to="M" />
<WordPart from="I\/l" to="M" />
<WordPart from="l\/l" to="M" />
<WordPart from="IVIa" to="Ma" />
<WordPart from="IVIe" to="Me" />
<WordPart from="IVIi" to="Mi" />
<WordPart from="IVIo" to="Mo" />
<WordPart from="IVIu" to="Mu" />
<WordPart from="IVIy" to="My" />
</PartialWordsAlways>
<PartialWords>
<!-- Will be used to check words not in dictionary.
If new word(s) and longer than 5 chars and exists in spelling dictionary, it is (or they are) accepted -->
<WordPart from="IVI" to="M" />
<WordPart from="/" to="l" />
<WordPart from="|" to="I" />
<WordPart from="vv" to="w" />
<WordPart from="m" to="rn" />
<WordPart from="l" to="i" />
<WordPart from="€" to="e" />
<WordPart from="I" to="l" />
<WordPart from="c" to="o" />
<WordPart from="i" to="t" />
<WordPart from="cc" to="oo" />
<WordPart from="ii" to="tt" />
<WordPart from="ii" to="ll" />
<!-- "f " will be two words -->
<WordPart from="f" to="f " />
<WordPart from="c" to="e" />
<WordPart from="I" to="t" />
<WordPart from="n" to="o" />
<WordPart from="s" to="e" />
<WordPart from="l-l" to="H" />
<WordPart from="l_" to="L" />
<WordPart from="°" to="o" />
<WordPart from=")/" to="y" />
<WordPart from=")'" to="y" />
<WordPart from="a’" to="d" />
<WordPart from="|" to="l" />
<WordPart from="¤" to="o" />
<WordPart from="vx/" to="w" />
</PartialWords>
<WholeLines>
</WholeLines>
<PartialLinesAlways>
</PartialLinesAlways>
<PartialLines>
</PartialLines>
<BeginLines>
</BeginLines>
<EndLines>
</EndLines>
<RegularExpressions>
<RegEx find="\b\|\b" replaceWith="I" />
</RegularExpressions>
</OCRFixReplaceList>

6
Dictionaries/names.xml
View File

@ -188,6 +188,7 @@ This file is case sensitive.
<name>Amari</name> <name>Amari</name>
<name>Amarion</name> <name>Amarion</name>
<name>Amarone</name> <name>Amarone</name>
<name>Amasova</name>
<name>Amazon</name> <name>Amazon</name>
<name>Amazons</name> <name>Amazons</name>
<name>Amber</name> <name>Amber</name>
@ -553,6 +554,7 @@ This file is case sensitive.
<name>Becca</name> <name>Becca</name>
<name>Beccas</name> <name>Beccas</name>
<name>Béchar</name> <name>Béchar</name>
<name>Bechmann</name>
<name>Beck</name> <name>Beck</name>
<name>Becker</name> <name>Becker</name>
<name>Beckett</name> <name>Beckett</name>
@ -1647,6 +1649,7 @@ This file is case sensitive.
<name>Dolly</name> <name>Dolly</name>
<name>Dolly Parton</name> <name>Dolly Parton</name>
<name>Dolores</name> <name>Dolores</name>
<name>Dom Pérignon</name>
<name>Domenic</name> <name>Domenic</name>
<name>Domina</name> <name>Domina</name>
<name>Dominic</name> <name>Dominic</name>
@ -2005,6 +2008,7 @@ This file is case sensitive.
<name>Ford</name> <name>Ford</name>
<name>Fordham</name> <name>Fordham</name>
<name>Forrest Gump</name> <name>Forrest Gump</name>
<name>Forsyth</name>
<name>Forsythe</name> <name>Forsythe</name>
<name>Fort Lauderdale</name> <name>Fort Lauderdale</name>
<name>Fortenberry</name> <name>Fortenberry</name>
@ -3501,6 +3505,7 @@ This file is case sensitive.
<name>Mark Hamill</name> <name>Mark Hamill</name>
<name>Mark Zuckerberg</name> <name>Mark Zuckerberg</name>
<name>Marko</name> <name>Marko</name>
<name>Markovitz</name>
<name>Markus</name> <name>Markus</name>
<name>Marky Mark</name> <name>Marky Mark</name>
<name>Marla</name> <name>Marla</name>
@ -3750,6 +3755,7 @@ This file is case sensitive.
<name>Michelles</name> <name>Michelles</name>
<name>Michigan</name> <name>Michigan</name>
<name>Michigans</name> <name>Michigans</name>
<name>Michoacán</name>
<name>Michonne</name> <name>Michonne</name>
<name>Mick</name> <name>Mick</name>
<name>Mickens</name> <name>Mickens</name>