blob: 0fd2d1181ea9fcabe1e8084e39403e77023b328c [file] [log] [blame]
# © 2016 and later: Unicode, Inc. and others.
# License & terms of use: http://www.unicode.org/copyright.html#License
#
# File: fa_fa_Latn_BGN.txt
# Generated from CLDR
#
#
########################################################################
# 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:][ءآابةتثجحخدذرزسشصضطظعغفقكلمنهویي\u064E\u064F\u0650\u0651\u0652٠١٢٣٤٥٦٧٨٩پچژگی]] ;
:: NFKD (NFC) ;
#
#
########################################################################
#
########################################################################
#
# Define All Transformation Variables
#
########################################################################
#
$alef = ’;
$ayin = ‘;
$disambig = \u0331 ;
#
#
# 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 ( \u0651 ) over a consonant character. In romanization
# the letter should be doubled. [The remainder of this rule deals with
# the definite article and is lexical.]
#
########################################################################
#
ب\u0651 bb ; # ARABIC LETTER BEH + SHADDA
پ\u0651 pp ; # ARABIC LETTER PEH + SHADDA
ت\u0651 tt ; # ARABIC LETTER TEH + SHADDA
ث\u0651 s\u0304s\u0304 ; # ARABIC LETTER THEH + SHADDA
ج\u0651 jj ; # ARABIC LETTER JEEM + SHADDA
چ\u0651 chch ; # ARABIC LETTER TCHEH + SHADDA
ح\u0651 ḥḥ ; # ARABIC LETTER HAH + SHADDA
خ\u0651 khkh ; # ARABIC LETTER KHAH + SHADDA
د\u0651 dd ; # ARABIC LETTER DAL + SHADDA
ذ\u0651 z\u0304z\u0304 ; # ARABIC LETTER THAL + SHADDA
ر\u0651 rr ; # ARABIC LETTER REH + SHADDA
ز\u0651 zz ; # ARABIC LETTER ZAIN + SHADDA
ژ\u0651 zhzh ; # ARABIC LETTER JEH + SHADDA
س\u0651 ss ; # ARABIC LETTER SEEN + SHADDA
ش\u0651 shsh ; # ARABIC LETTER SHEEN + SHADDA
ص\u0651 ṣṣ ; # ARABIC LETTER SAD + SHADDA
ض\u0651 ḍḍ ; # ARABIC LETTER DAD + SHADDA
ط\u0651 ṭṭ ; # ARABIC LETTER TAH + SHADDA
ظ\u0651 ẓẓ ; # ARABIC LETTER ZAH + SHADDA
ع\u0651 $ayin $ayin ; # ARABIC LETTER AIN + SHADDA
غ\u0651 ghgh ; # ARABIC LETTER GHAIN + SHADDA
ف\u0651 ff ; # ARABIC LETTER FEH + SHADDA
ق\u0651 qq ; # ARABIC LETTER QAF + SHADDA
ك\u0651 kk ; # ARABIC LETTER KAF + SHADDA
ل\u0651 ll ; # ARABIC LETTER LAM + SHADDA
م\u0651 mm ; # ARABIC LETTER MEEM + SHADDA
ن\u0651 nn ; # ARABIC LETTER NOON + SHADDA
ه\u0651 hh ; # ARABIC LETTER HEH + SHADDA
و\u0651 ww ; # ARABIC LETTER WAW + SHADDA
ی\u0651 yy ; # ARABIC LETTER FARSI YEH + SHADDA
#
#
########################################################################
#
# End Rule 7
#
########################################################################
#
########################################################################
#
# Start of Transformations
#
########################################################################
#
$wordBoundary ; # ARABIC LETTER HAMZA
ء $alef ; # ARABIC LETTER HAMZA
$wordBoundary ; # 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 s\u0304 ; # ARABIC LETTER THEH
ج j ; # ARABIC LETTER JEEM
چ ch ; # ARABIC LETTER TCHEH
ح ; # ARABIC LETTER HAH
خ kh ; # ARABIC LETTER KHAH
د d ; # ARABIC LETTER DAL
ذ z\u0304 ; # 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 ; # ARABIC LETTER SAD
ض ; # ARABIC LETTER DAD
ط ; # ARABIC LETTER TAH
ظ ; # 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
\u064Eا ā ; # ARABIC FATHA + ALEF
\u064Eی á ; # ARABIC FATHA + FARSI YEH
\u064Eو\u0652 ow ; # ARABIC FATHA + WAW + SUKUN
\u064E a ; # ARABIC FATHA
\u0650ي ī ; # ARABIC KASRA + YEH
\u0650 e ; # ARABIC KASRA
\u064Fو ū ; # ARABIC DAMMA + WAW
\u064F o ; # ARABIC DAMMA
\u0652 ; # ARABIC SUKUN
::NFC (NFD) ;
#
#
########################################################################