icuSources/data/translit/Hebr_Latn.txt

   1 # © 2016 and later: Unicode, Inc. and others.
   2 # License & terms of use: http://www.unicode.org/copyright.html#License
   3 #
   4 # File: Hebr_Latn.txt
   5 # Generated from CLDR
   6 #
   7
   8 # Transliteration table for Hebrew
   9 # Based on the UNGEGN table at:
  10 #   http://www.eki.ee/wgrs/rom1_he.pdf
  11 #
  12 # Exceptions:
  13 # - Accents are added to disambiguate letters
  14 # - Combinations of dagesh, shin/sin dot that produce different
  15 #   letters are not yet encoded.
  16 #
  17 # To test, open:
  18 #   http://www.ibm.com/software/globalization/icu/demo/transform
  19 # Click Edit, paste in this file, Save As hebrew-latin/XXX
  20 # (where XXX is a username)
  21 # Now go back to the main window, and try it out.
  22 # Use hebrew-latin/XXX for Output 1, and (Inverse) for Output 2
  23 # Paste in hebrew text in Input, and hit Transliterate.
  24 #
  25 # For more information, see:
  26 #   http://icu.sourceforge.net/userguide/Transform.html
  27 :: [[:Hebrew:] [:^ccc=0:] [\u05B0-\u05B9\u05BB-\u05BC\u05C1-\u05C2ℵ-ℸ\u0304\u05BF] - [\u05BD]] ;
  28 :: nfkd (nfc) ;
  29 $letterAfter = [:M:]* [:L:] ;
  30 # move longer items here to avoid masking
  31 ח ↔ h\u0331 ;
  32 צ ↔ z\u0331 } $letterAfter;
  33 ץ ↔ z\u0331 ;
  34 ש ↔ s\u0327 ;
  35 ת ↔ t\u0327 ;
  36 א ↔ ʼ ;
  37 ב ↔ b ;
  38 ג ↔ g ;
  39 ד ↔ d ;
  40 ה ↔ h ;
  41 ו ↔ w ;
  42 ז ↔ z ;
  43 ט ↔ t ;
  44 י ↔ y ;
  45 כ ↔ k } $letterAfter;
  46 ך ↔ k ;
  47 ל ↔ l ;
  48 מ ↔ m } $letterAfter;
  49 ם ↔ m ;
  50 נ ↔ n  } $letterAfter;
  51 ן ↔ n ;
  52 ס ↔ s ;
  53 ע ↔ ʻ ;
  54 פ ↔ p } $letterAfter;
  55 ף ↔ p ;
  56 ק ↔ q ;
  57 ר ↔ r ;
  58 װ → |  וו; # HEBREW LIGATURE YIDDISH DOUBLE VAV
  59 ױ → | וי; # HEBREW LIGATURE YIDDISH VAV YOD
  60 ײ → | יי ; # HEBREW LIGATURE YIDDISH DOUBLE YOD
  61 \u05BC ↔ \u0307 ; # dagesh just goes to overdot for now
  62 \u05C1 ↔ \u030C ; # shin dot -→ sh
  63 \u05C2 ↔ \u0302 ; # sin dot -→ s
  64 # points
  65 $above = [^[:ccc=0:][:ccc=230:]]*;
  66 ‎\u05B2‎ → à ;
  67 ‎\u05B2‎ $1← a ($above)  \u0300;
  68 ‎\u05B8‎ → á ;
  69 ‎\u05B8‎  $1 ← a ($above)  \u0301;
  70 ‎\u05B1‎ → è ;
  71 ‎\u05B1‎  $1 ← e ($above)  \u0300;
  72 ‎\u05B5‎ → é ;
  73 ‎\u05B5‎  $1 ← e ($above)   \u0301;
  74 ‎\u05B0‎ → e \u0306 ;
  75 ‎\u05B0‎  $1 ← e ($above)   \u0306;
  76 ‎\u05B9‎ → ò ;
  77 ‎\u05B9‎  $1 ← o ($above)   \u0300;
  78 \u05B4 ↔ i ;
  79 \u05BB ↔ u ;
  80 \u05B7 ↔ a ;
  81 \u05B6 ↔ e ;
  82 \u05B3 ↔ o ;
  83 \u05BF ↔   \u0304 ;
  84 # fallbacks
  85 ק ← c ;
  86 פ ← f } $letterAfter;
  87 ף ← f ;
  88 ז ← j ;
  89 ו ← v ;
  90 כס ← x ;
  91 :: (lower);
  92 :: nfc (nfd) ;
  93 :: ([[:Latin:] [:^ccc=0:] [ʻ-ʼ\u0300-\u0302\u0307\u030C\u0327\u0331\u0340-\u0341    \u0304 ]]);
  94