[sword-svn] r266 - trunk/source/data/translit/crosswire

[chrislit at crosswire.org]

Author: chrislit
Date: 2014-02-07 20:24:14 -0700 (Fri, 07 Feb 2014)
New Revision: 266

Modified:
   trunk/source/data/translit/crosswire/Greek_Latin_TC.txt
   trunk/source/data/translit/crosswire/Hebrew_Latin_TC.txt
Log:
updated, checked, tested Hebrew/TC transliterator
updated, checked, tested Greek/TC transliterator


Modified: trunk/source/data/translit/crosswire/Greek_Latin_TC.txt
===================================================================
--- trunk/source/data/translit/crosswire/Greek_Latin_TC.txt	2014-02-07 09:27:47 UTC (rev 265)
+++ trunk/source/data/translit/crosswire/Greek_Latin_TC.txt	2014-02-08 03:24:14 UTC (rev 266)
@@ -1,6 +1,6 @@
 # ***************************************************************************
 # *
-# *  Copyright (C) 2002-2013, CrossWire Bible Society
+# *  Copyright (C) 2002-2014, CrossWire Bible Society
 # *  All Rights Reserved.
 # *
 # ***************************************************************************
@@ -10,43 +10,48 @@
 # Based on TC Transliteration:
 # http://rosetta.reltech.org/TC/TC-translit.html
 # Greek-Latin/TC
+
+:: [[:Greek:][̣̈̓̔́̀͂ͅ·—‐’̅]];
 :: NFD (NFC);
-# Uppercase
-Α<>'A';
-Β<>'B';
-Γ<>'G';
-Δ<>'D';
-Ε<>'E';
-Ζ<>'Z';
-Η<>'H';
-Θ<>'Q';
-Ι<>'I';
-Κ<>'K';
-Λ<>'L';
-Μ<>'M';
-Ν<>'N';
-Ξ<>'C';
-Ο<>'O';
-Φ<>'F';
-Ψ<>'Y';
-Π<>'P';
-Ρ<>'R';
-Σ<>'S';
-Σ<'J';
-Τ<>'T';
-Υ<>'U';
-Χ<>'X';
-Ω<>'W';
-# Archaic
-Ϝ>;
-Ϟ>;
-Ϡ>;
-# Lowercase
+
+$betaMarks = [\+\)\(\|\/\\\=\?];
+
+# Letters
+#  Uppercase
+Α<>A;
+Β<>B;
+Γ<>G;
+Δ<>D;
+Ε<>E;
+Ϝ<>V;
+Ζ<>Z;
+Η<>H;
+Θ<>Q;
+Ι<>I;
+Κ<>K;
+Λ<>L;
+Μ<>M;
+Ν<>N;
+Ξ<>C;
+Ο<>O;
+Π<>P;
+Ρ<>R;
+Σ<>S;
+Σ<J;
+Τ<>T;
+Υ<>U;
+Φ<>F;
+Χ<>X;
+Ψ<>Y;
+Ω<>W;
+
+#  Lowercase
 α<>a;
 β<>b;
 γ<>g;
 δ<>d;
 ε<>e;
+ϝ<>v;
 ζ<>z;
 η<>h;
 θ<>q;
@@ -57,41 +62,40 @@
 ν<>n;
 ξ<>c;
 ο<>o;
-φ<>f;
-ψ<>y;
 π<>p;
 ρ<>r;
-ς<>j;
+ς<j;
+ς<s}[^[:L:][:M:]];
+ς>s;
 σ<>s;
 τ<>t;
 υ<>u;
+φ<>f;
 χ<>x;
+ψ<>y;
 ω<>w;
-# Archaic
-ϝ>;
-ϟ>;
-ϡ>;
+
 # Non-letter characters
-ͅ<>'\|';
-ͺ>'\|';
-̔<>'\(';
-ʽ>'\(';
-·<>':';
-̓<>'\)';
-ʼ>'\)';
-́<>'\/';
-´>'\/';
-ˊ>'\/';
-΄>'\/';
-͂<>'=';
-̂>'=';
-'^'>'=';
-ˆ>'=';
-̀<>'\\';
-'`'>'\\';
-ˋ>'\\';
-̈<>'+';
-¨>'+';
-̣<>'?';
-# ';'<>';';
+̈<>\+;
+̓<>\);
+̔<>\(;
+ͅ<>\|;
+́<>\/;
+̀<>\\;
+͂<>\=;
+̣<>\?;
+
+·<>\:;
+—<>\-\-;
+‐<>\-;
+’<>\';
+̅<>\_;
+
+::Null;
+
+# Move accents preceding capitals/words after the letter
+$2$1<($betaMarks+)([A-Z]);
+$2$1<[^[[A-Za-z123]$betaMarks]]{($betaMarks+)([A-Za-z]);
+
 :: NFC (NFD);
+:: ([A-Za-z\+\)\(\|\/\\\=\?\:\-\'\_]);

Modified: trunk/source/data/translit/crosswire/Hebrew_Latin_TC.txt
===================================================================
--- trunk/source/data/translit/crosswire/Hebrew_Latin_TC.txt	2014-02-07 09:27:47 UTC (rev 265)
+++ trunk/source/data/translit/crosswire/Hebrew_Latin_TC.txt	2014-02-08 03:24:14 UTC (rev 266)
@@ -1,6 +1,6 @@
 # ***************************************************************************
 # *
-# *  Copyright (C) 2002-2013, CrossWire Bible Society
+# *  Copyright (C) 2002-2014, CrossWire Bible Society
 # *  All Rights Reserved.
 # *
 # ***************************************************************************
@@ -8,13 +8,21 @@
 # Generated from CLDR
 
 # Based on TC Transliteration:
-# http://rosetta.reltech.org/TC/TC-translit.html
+# http://rosetta.reltech.org/TC/TC-translit-main.html
+
 # Hebrew-Latin/TC
+
+:: [[:Hebrew:][\‍\ ]];
 :: NFD (); #Don't put into NFC when Hebrew is the destination
-$letter=[[:L:] [\)+\(\:\&\$\#]];
-ו ׄ<>ow;
-ו ּ<>U;
-א<>'\)';
+
+$letter=[[a-zAEFU][\)+\(\&\$\#\:\-\.\,\|\;\"\@\=\\\']];
+
+# Completeness mappings
+|w<v;
+|i<j;
+
+# Consonants
+א<>\);
 ב<>b;
 ג<>g;
 ד<>d;
@@ -22,7 +30,7 @@
 ו<>w;
 ז<>z;
 ח<>x;
-ט<>'+';
+ט<>\+;
 י<>y;
 כ<k}$letter;
 כ>k;
@@ -35,7 +43,7 @@
 נ>n;
 ן<>n;
 ס<>s;
-ע<>'\(';
+ע<>\(;
 פ<p}$letter;
 פ>p;
 ף<>p;
@@ -44,62 +52,49 @@
 ץ<>c;
 ק<>q;
 ר<>r;
-ש ׂ<>'\&';
-ש ׁ<>'\$';
-ש>'\#';
+שׂ<>\&;
+שׁ<>\$;
+ש<>\#;
 ת<>t;
+
+# Yiddish additions
 װ>ww;
 ױ>wy;
 ײ>yy;
-ְ<>':';
-ֱ<>'E';
-ֲ<>'A';
-ֳ<>'F';
-ִ<>i;
-ֵ<>'\"';
-ֶ<>e;
+
+# Vowels
 ַ<>a;
 ָ<>f;
+ִ<>i;
+ֶ<>e;
+ֵ<>\";
 ֹ<>o;
 ֻ<>u;
-ֿ>',';
-ׄ>;
-֯>'\\';
-׳>;
-״>;
-ּ<>'\.';
-ֽ<>'\|';
-־<>'-';
-׀>;
-׃>'\;';
-֑>;
-֒>;
-֓>;
-֔>;
-֕>;
-֖>;
-֗>;
-֘>;
-֙>;
-֚>;
-֛>;
-֜>;
-֝>;
-֞>;
-֟>;
-֠>;
-֡>;
-֣>;
-֤>;
-֥>;
-֦>;
-֧>;
-֨>;
-֩>;
-֪>;
-֫>;
-֬>;
-֭>;
-֮>;
-:: ([[$letter]-[AEFU]] lower);
+ּ<U;
+ְ<>\:;
+ֲ<>A;
+ֳ<>F;
+ֱ<>E;
+
+־<>\-;
+ּ<>\.;
+ֿ<>\,;
+ֽ<>\|;
+׃<>\;;
+
+# Special Masoretic marks
+׳<\@;
+׳<>\';
+֜>\';
+״<>\=;
+֞>\=;
+֯<>\\;
+
+# Delete remaining Hebrew points (mostly cantillation)
+' ׀'>;
+'‍ֽ'>;
+[֮֒׀ׅ֤֥֦֖֣֛֑֪֧֚֭֝֠֩֙ׄ֫֨֬֓֔֗֘֡֟֕]>;
+
+:: ([BCDGHIJKLMNOPQRSTVWXYZ] lower);
 :: NFC (NFD);
+:: ([[:Latin:][\)+\(\&\$\#\:\-\.\,\|\;\"\@\=\\\']]);