src/com/ibm/text/resources/Transliterator_Latin_Greek.utf8.txt

#--------------------------------------------------------------------
# Copyright (c) 1999-2001, International Business Machines
# Corporation and others.  All Rights Reserved.
#--------------------------------------------------------------------
# Date: Tue Jan 23 12:18:45 2001
#--------------------------------------------------------------------

# Latin-Greek

# ==============================================
# Modern Greek Transliteration Rules
#
# This transliterates modern Greek characters, but using rules
# that are traditional for Ancient Greek, and
# thus more resemble Greek words that have become part
# of English. It differs from the official Greek
# transliteration, which is more phonetic (since
# most modern Greek vowels, for example, have
# degenerated simply to sound like "ee").
#
# There are only a few tricky parts.
# 1. eta and omega don't map directly to Latin vowels,
#    so we use a macron on e and o, and some
#    other combinations if they are accented.
# 2. The accented, diaeresis i and y are substituted too.
# 3. Some letters use digraphs, like "ph". While typical,
#    they need some special handling.
# 4. A gamma before a gamma or a few other letters is
#    transliterated as an "n", as in "Anglo"
# 5. An ypsilon after a vowel is a "u", as in
#    "Mouseio". Otherwise it is a "y" as in "Physikon"
# 6. The construction of the rules is made simpler by making sure
#    that most rules for lowercase letters exactly correspond to the
#    rules for uppercase letters, *except* for the case of the letters
#    in the rule itself. That way, after modifying the uppercase rules,
#    you can just copy, paste, and "set to lowercase" to get
#    the rules for lowercase letters!
# ==============================================

# ==============================================
# Variables, used to make the rules more comprehensible
# and for conditionals.
# ==============================================

### $quote='\"';

# Latin Letters

### $E_MACRON=Ē;
### $e_macron=ē;
### $O_MACRON=Ō;
### $o_macron=ō;
### $Y_UMLAUT=Ÿ;
### $y_umlaut=ÿ;

#! // with real accents.
#! + "$E_MACRON_ACUTE=Ḗ;"
#! + "$e_macron_acute=ḗ;"
#! + "$O_MACRON_ACUTE=Ṓ;"
#! + "$o_macron_acute=ṓ;"
#! + "$y_umlaut_acute=ÿ́;"
#! + "$u00ef_acute=ḯ;"
#! + "$u00fc_acute=ǘ;"
#! //

# single letter equivalents

### $E_MACRON_ACUTE=Ê;
### $e_macron_acute=ê;
### $O_MACRON_ACUTE=Ô;
### $o_macron_acute=ô;
### $y_umlaut_acute=ŷ;
### $u00ef_acute=î;
### $u00fc_acute=û;

# Greek Letters

### $ALPHA=Α;
### $BETA=Β;
### $GAMMA=Γ;
### $DELTA=Δ;
### $EPSILON=Ε;
### $ZETA=Ζ;
### $ETA=Η;
### $THETA=Θ;
### $IOTA=Ι;
### $KAPPA=Κ;
### $LAMBDA=Λ;
### $MU=Μ;
### $NU=Ν;
### $XI=Ξ;
### $OMICRON=Ο;
### $PI=Π;
### $RHO=Ρ;
### $SIGMA=Σ;
### $TAU=Τ;
### $YPSILON=Υ;
### $PHI=Φ;
### $CHI=Χ;
### $PSI=Ψ;
### $OMEGA=Ω;

### $ALPHA2=Ά;
### $EPSILON2=Έ;
### $ETA2=Ή;
### $IOTA2=Ί;
### $OMICRON2=Ό;
### $YPSILON2=Ύ;
### $OMEGA2=Ώ;
### $IOTA_DIAERESIS=Ϊ;
### $YPSILON_DIAERESIS=Ϋ;

### $alpha=α;
### $beta=β;
### $gamma=γ;
### $delta=δ;
### $epsilon=ε;
### $zeta=ζ;
### $eta=η;
### $theta=θ;
### $iota=ι;
### $kappa=κ;
### $lambda=λ;
### $mu=μ;
### $nu=ν;
### $xi=ξ;
### $omicron=ο;
### $pi=π;
### $rho=ρ;
### $sigma=σ;
### $tau=τ;
### $ypsilon=υ;
### $phi=φ;
### $chi=χ;
### $psi=ψ;
### $omega=ω;

#forms

### $alpha2=ά;
### $epsilon2=έ;
### $eta2=ή;
### $iota2=ί;
### $omicron2=ό;
### $ypsilon2=ύ;
### $omega2=ώ;
### $iota_diaeresis=ϊ;
### $ypsilon_diaeresis=ϋ;
### $iota_diaeresis2=ΐ;
### $ypsilon_diaeresis2=ΰ;
### $sigma2=ς;

# Variables for conditional mappings

# Use lowercase for all variable names, to allow cut/paste below.

$letter=[~[:Lu:][:Ll:]];
$lower=[[:Ll:]];
$softener=[eiyEIY];
$vowel=[aeiouAEIOU \
 ΑΕΗΙΟΥΩ \
 ΆΈΉΊΌΎΏ \
 ΪΫ \
 αεηιουω \
 άέήίόύώ \
 ϊϋ \
 ΐΰ \
 ];
$n_gamma=[GKXCgkxc];
$gamma_n=[ΓΚΧΞγκχξ];
$pp=[Pp];

# ==============================================
# Rules
# ==============================================
# The following are special titlecases, and should
# not be copied when duplicating the lowercase
# ==============================================

Th <> Θ}$lower;
Ph <> Φ}$lower;
Ch <> Χ}$lower;
#masked: + "Ps<Φ}$lower;"

# Because there is no uppercase forms for final sigma,
# we had to move all the sigma rules up here.

# Remember to insert ' to preserve round trip, for double letters
# don't need to do this for the digraphs with h,
# since it is not created when mapping back from greek

# use special form for s

''S <> $pp{Σ; # handle PS
S <> Σ;

# The following are a bit tricky. 's' takes two forms in greek
# final or non final.
# We use ~s to represent the abnormal form: final before letter
# or non-final before non-letter.
# We use 's to separate p and s (otherwise ps is one letter)
# so, we break out the following forms:

''s < $pp{σ}$letter;
s <          σ}$letter;
'~'s <         σ;

'~'s <         ς}$letter;
''s < $pp{ς;
s <          ς;

'~'s }$letter>ς;
'~'s             > σ;
''s }$letter>σ;
''s            > ς;
s }$letter>σ;
s              > ς;

# because there are no uppercase forms, had to move these up too.

i'\"''`'>ΐ;
y'\"''`'>ΰ;

î<>ΐ;
û<>$vowel{ΰ;
ŷ<>ΰ;

# ==============================================
# Uppercase Forms.
# To make lowercase forms, just copy and lowercase below
# ==============================================

# Typing variants, in case the keyboard doesn't have accents

'A`'>Ά;
'E`'>Έ;
'EE`'>Ή;
EE>Η;
'I`'>Ί;
'O`'>Ό;
'OO`'>Ώ;
OO>Ω;
I'\"'>Ϊ;
Y'\"'>Ϋ;

# Basic Letters

A<>Α;
Á<>Ά;
B<>Β;
N }$n_gamma<>Γ}$gamma_n;
G<>Γ;
D<>Δ;
''E <> [Ee]{Ε; # handle EE
E<>Ε;
É<>Έ;
Z<>Ζ;
Ê<>Ή;
Ē<>Η;
TH<>Θ;
I<>Ι;
Í<>Ί;
Ï<>Ϊ;
K<>Κ;
L<>Λ;
M<>Μ;
N'' <> Ν}$gamma_n;
N<>Ν;
X<>Ξ;
''O <> [Oo]{ Ο; # handle OO
O<>Ο;
Ó<>Ό;
PH<>Φ; # needs ordering before P
PS<>Ψ; # needs ordering before P
P<>Π;
R<>Ρ;
T<>Τ;
U <> $vowel{Υ;
Ú <> $vowel{Ύ;
Ü <> $vowel{Ϋ;
Y<>Υ;
Ý<>Ύ;
Ÿ<>Ϋ;
CH<>Χ;
Ô<>Ώ;
Ō<>Ω;

# Extra English Letters. Mapped for completeness

C}$softener>|S;
C>|K;
F>|PH;
H>|CH;
J>|I;
Q>|K;
V>|U;
W>|U;

# ==============================================
# Lowercase Forms. Just copy above and lowercase
# ==============================================

# typing variants, in case the keyboard doesn't have accents

'a`'>ά;
'e`'>έ;
'ee`'>ή;
ee>η;
'i`'>ί;
'o`'>ό;
'oo`'>ώ;
oo>ω;
i'\"'>ϊ;
y'\"'>ϋ;

# basic letters

a<>α;
á<>ά;
b<>β;
n }$n_gamma<>γ}$gamma_n;
g<>γ;
d<>δ;
''e <> [Ee]{ε; # handle EE
e<>ε;
é<>έ;
z<>ζ;
ê<>ή;
ē<>η;
th<>θ;
i<>ι;
í<>ί;
ï<>ϊ;
k<>κ;
l<>λ;
m<>μ;
n'' <> ν}$gamma_n;
n<>ν;
x<>ξ;
''o <> [Oo]{ ο; # handle OO
o<>ο;
ó<>ό;
ph<>φ; # needs ordering before p
ps<>ψ; # needs ordering before p
p<>π;
r<>ρ;
t<>τ;
u <> $vowel{υ;
ú <> $vowel{ύ;
ü <> $vowel{ϋ;
y<>υ;
ý<>ύ;
ÿ<>ϋ;
ch<>χ;
ô<>ώ;
ō<>ω;

# extra english letters. mapped for completeness

c}$softener>|s;
c>|k;
f>|ph;
h>|ch;
j>|i;
q>|k;
v>|u;
w>|u;

# ====================================
# Normal final rule: remove '
# ====================================

#+ "''>;"

# eof