common/transforms/Persian-Latin-BGN.xml

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE supplementalData SYSTEM "../../common/dtd/ldmlSupplemental.dtd">
<!--
Copyright © 1991-2013 Unicode, Inc.
CLDR data files are interpreted according to the LDML specification (http://unicode.org/reports/tr35/)
For terms of use, see http://www.unicode.org/copyright.html
-->
<supplementalData>
	<version number="$Revision$"/>
	<transforms>
		<transform source="fa" target="fa_Latn" variant="BGN" direction="forward" draft="contributed" alias="Persian-Latin/BGN fa-Latn-t-fa-m0-bgn">
			<tRule><![CDATA[
#
########################################################################
# BGN/PCGN 1956 System
#
# This system was adopted by the BGN in 1946 and by the PCGN in 1958.
# It is used for the romanization of geographic names in Iran and
# for Persian-language names in Afghanistan.
#
# Originally prepared by Michael Everson <everson@evertype.com>
########################################################################
#
# MINIMAL FILTER: Persian-Latin
#

:: [[:arabic:][:block=ARABIC:][ءآابةتثجحخدذرزسشصضطظعغفقكلمنهویيَُِّْ٠١٢٣٤٥٦٧٨٩پچژگی]] ;
:: NFKD (NFC) ;
#
#
########################################################################


#
########################################################################
#
# Define All Transformation Variables
#
########################################################################
#

$alef = ’;
$ayin = ‘;
$disambig =  ̱ ;
#
#
# Use this $wordBoundary until bug 2034 is fixed in ICU:
# http://bugs.icu-project.org/cgi-bin/icu-bugs/transliterate?id=2034;expression=boundary;user=guest
#

$wordBoundary =  [^[:L:][:M:][:N:]] ;
#
#
########################################################################

# non-letters
[:Nd:]{٫}[:Nd:] ↔ [:Nd:]{','}[:Nd:] ; # ARABIC DECIMAL SEPARATOR
[:Nd:]{٬}[:Nd:] ↔ [:Nd:]{'.'}[:Nd:] ; # ARABIC THOUSANDS SEPARATOR
٫ ↔ ',' $disambig ; # ARABIC DECIMAL SEPARATOR
٬ ↔ '.' $disambig ; # ARABIC THOUSANDS SEPARATOR
#  ٭ ↔ ; # ARABIC FIVE POINTED STAR // no need to transliterate
، ↔ ',' ; # ARABIC COMMA
؛ ↔ ';' ; # ARABIC SEMICOLON
؟ ↔ '?' ; # ARABIC QUESTION MARK
٪ ↔ '%' ; # ARABIC PERCENT SIGN
٠ ↔ 0 $disambig ; # ARABIC-INDIC DIGIT ZERO
١ ↔ 1 $disambig ; # ARABIC-INDIC DIGIT ONE
٢ ↔ 2 $disambig ; # ARABIC-INDIC DIGIT TWO
٣ ↔ 3 $disambig ; # ARABIC-INDIC DIGIT THREE
٤ ↔ 4 $disambig ; # ARABIC-INDIC DIGIT FOUR
٥ ↔ 5 $disambig ; # ARABIC-INDIC DIGIT FIVE
٦ ↔ 6 $disambig ; # ARABIC-INDIC DIGIT SIX
٧ ↔ 7 $disambig ; # ARABIC-INDIC DIGIT SEVEN
٨ ↔ 8 $disambig ; # ARABIC-INDIC DIGIT EIGHT
٩ ↔ 9 $disambig ; # ARABIC-INDIC DIGIT NINE
۰ ↔ 0 ; # EXTENDED ARABIC-INDIC DIGIT ZERO
۱ ↔ 1 ; # EXTENDED ARABIC-INDIC DIGIT ONE
۲ ↔ 2 ; # EXTENDED ARABIC-INDIC DIGIT TWO
۳ ↔ 3 ; # EXTENDED ARABIC-INDIC DIGIT THREE
۴ ↔ 4 ; # EXTENDED ARABIC-INDIC DIGIT FOUR
۵ ↔ 5 ; # EXTENDED ARABIC-INDIC DIGIT FIVE
۶ ↔ 6 ; # EXTENDED ARABIC-INDIC DIGIT SIX
۷ ↔ 7 ; # EXTENDED ARABIC-INDIC DIGIT SEVEN
۸ ↔ 8 ; # EXTENDED ARABIC-INDIC DIGIT EIGHT
۹ ↔ 9 ; # EXTENDED ARABIC-INDIC DIGIT NINE

#
########################################################################
#
# Rules moved to front to avoid masking
#
########################################################################


#
########################################################################
#
# BGN Page 89 Rule 4
#
# The character sequences كه , زه , سه , and گه may be romanized k·h, z·h,
# s·h, and g·h in order to differentiate those romanizations from the
# digraphs kh, zh, sh, and gh.
#
########################################################################
#

كه → k·h ; # ARABIC LETTER KAF + HEH
زه → z·h ; # ARABIC LETTER ZAIN + HEH
سه → s·h ; # ARABIC LETTER SEEN + HEH
گه → g·h ; # ARABIC LETTER GAF + HEH
#
#
########################################################################
#
# End Rule 4
#
########################################################################



#
########################################################################
#
# BGN Page 91 Rule 7
#
# Doubles consonant sounds are represented in Arabic script by
# placing a shaddah ( ّ ) over a consonant character. In romanization
# the letter should be doubled. [The remainder of this rule deals with
# the definite article and is lexical.]
#
########################################################################
#

بّ → bb ; # ARABIC LETTER BEH + SHADDA
پّ → pp ; # ARABIC LETTER PEH + SHADDA
تّ → tt ; # ARABIC LETTER TEH + SHADDA
ثّ → s̄s̄ ; # ARABIC LETTER THEH + SHADDA
جّ → jj ; # ARABIC LETTER JEEM + SHADDA
چّ → chch ; # ARABIC LETTER TCHEH + SHADDA
حّ → ḥḥ ; # ARABIC LETTER HAH + SHADDA
خّ → khkh ; # ARABIC LETTER KHAH + SHADDA
دّ → dd ; # ARABIC LETTER DAL + SHADDA
ذّ → z̄z̄ ; # ARABIC LETTER THAL + SHADDA
رّ → rr ; # ARABIC LETTER REH + SHADDA
زّ → zz ; # ARABIC LETTER ZAIN + SHADDA
ژّ → zhzh ; # ARABIC LETTER JEH + SHADDA
سّ → ss ; # ARABIC LETTER SEEN + SHADDA
شّ → shsh ; # ARABIC LETTER SHEEN + SHADDA
صّ → ṣṣ ; # ARABIC LETTER SAD + SHADDA
ضّ → ḍḍ ; # ARABIC LETTER DAD + SHADDA
طّ → ṭṭ ; # ARABIC LETTER TAH + SHADDA
ظّ → ẓẓ ; # ARABIC LETTER ZAH + SHADDA
عّ → $ayin $ayin ; # ARABIC LETTER AIN + SHADDA
غّ → ghgh ; # ARABIC LETTER GHAIN + SHADDA
فّ → ff ; # ARABIC LETTER FEH + SHADDA
قّ → qq ; # ARABIC LETTER QAF + SHADDA
كّ → kk ; # ARABIC LETTER KAF + SHADDA
لّ → ll ; # ARABIC LETTER LAM + SHADDA
مّ → mm ; # ARABIC LETTER MEEM + SHADDA
نّ → nn ; # ARABIC LETTER NOON + SHADDA
هّ → hh ; # ARABIC LETTER HEH + SHADDA
وّ → ww ; # ARABIC LETTER WAW + SHADDA
یّ → yy ; # ARABIC LETTER FARSI YEH + SHADDA
#
#
########################################################################
#
# End Rule 7
#
########################################################################



#
########################################################################
#
# Start of Transformations
#
########################################################################
#

$wordBoundary{ء →  ; # ARABIC LETTER HAMZA
ء → $alef ; # ARABIC LETTER HAMZA
$wordBoundary{ا →  ; # ARABIC LETTER ALEF
آ → $alef ā ; # ARABIC FATHA ALEF WITH MADDA ABOVE
ب → b ; # ARABIC LETTER BEH
پ → p ; # ARABIC LETTER PEH
ت → t ; # ARABIC LETTER TEH
ة → h ; # ARABIC LETTER TEH MARBUTA
ث → s̄ ; # ARABIC LETTER THEH
ج → j ; # ARABIC LETTER JEEM
چ → ch ; # ARABIC LETTER TCHEH
ح → ḥ ; # ARABIC LETTER HAH
خ → kh ; # ARABIC LETTER KHAH
د → d ; # ARABIC LETTER DAL
ذ → z̄ ; # ARABIC LETTER THAL
ر → r ; # ARABIC LETTER REH
ز → z ; # ARABIC LETTER ZAIN
ژ → zh ; # ARABIC LETTER JEH
س → s ; # ARABIC LETTER SEEN
ش → sh ; # ARABIC LETTER SHEEN
ص → ṣ ; # ARABIC LETTER SAD
ض → ẕ ; # ARABIC LETTER DAD
ط → ṭ ; # ARABIC LETTER TAH
ظ → ẓ ; # ARABIC LETTER ZAH
ع → $ayin ; # ARABIC LETTER AIN
غ → gh ; # ARABIC LETTER GHAIN
ف → f ; # ARABIC LETTER FEH
ق → q ; # ARABIC LETTER QAF
ک ↔ k ; # ARABIC LETTER KEHEH
ك ↔ k $disambig ; # ARABIC LETTER KAF
گ → g ; # ARABIC LETTER GAF
ل → l ; # ARABIC LETTER LAM
م → m ; # ARABIC LETTER MEEM
ن → n ; # ARABIC LETTER NOON
ه → h ; # ARABIC LETTER HEH
و → v ; # ARABIC LETTER WAW
ی → y ; # ARABIC LETTER FARSI YEH

َا → ā ; # ARABIC FATHA + ALEF
َی → á ; # ARABIC FATHA + FARSI YEH
َوْ → ow ; # ARABIC FATHA + WAW + SUKUN
َ → a ; # ARABIC FATHA

ِي → ī ; # ARABIC KASRA + YEH
ِ → e ; # ARABIC KASRA

ُو → ū ; # ARABIC DAMMA + WAW
ُ → o ; # ARABIC DAMMA

ْ →  ; # ARABIC SUKUN
::NFC (NFD) ;

#
#
########################################################################

			]]></tRule>
		</transform>
	</transforms>
</supplementalData>