common/transforms/und_FONIPA-fa.xml

<?xml version="1.0" encoding="UTF-8" ?>
<!DOCTYPE supplementalData SYSTEM "../../common/dtd/ldmlSupplemental.dtd">
<!--
Copyright © 1991-2016 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="und_FONIPA" target="fa" direction="forward" alias="fa-t-und-fonipa">
		  <tRule><![CDATA[
# Vowels
# ------
# In these rules, we produce ی و ا both for short and for long vowels.
# This would be wrong for writing Farsi or Arabic, but when transliterating
# foreign words and names, it is strongly preferred to vowel marks.
# Short schwa [ə] and a few other, schwa-like vowels get omitted entirely
# unless at the end of the word, in which case we emit ه whose Farsi
# word-final pronunciation comes close to [ə]. At the beginning of words,
# Farsi speakers prefer to see آ for [ɑ] and a few other similar-sounding
# dark vowels; note that this use of آ is quite different from Arabic.

$IVowel = [i ɪ e {e̞}];
$UVowel = [y {ɨ} {ʉ} ɯ u ʏ {ɪ̈} {ʊ̈} {ɯ̽} {ʊ} ø ɤ o {ø̞} {ɤ̞} {o̞} ɔ w {w̥} ʍ ʷ];
$AVowel = [ɛ œ ɜ æ ɶ];
$DarkAVowel = [ʌ a ɑ ɒ ɐ ɞ {ä} {ɒ̈}];  # آ instead of ا at beginning of words
$SchwaVowel = [ɘ ɵ ə {ɵ̞}];
$Click = [ʘ ɋ ǀ ʇ ǃ ʗ ǂ ʄ ǁ ʖ];
$Boundary =  [^[:L:][:M:][:N:]];

::NFD;
[ʰ ʱ ʼ  ̃  ̰  ̋  ́  ̄  ̀  ̏  ̌  ̂ ˥ ˦ ˧ ˨ ˩ ꜜ ꜛ ↗ ↘  ͡  ͜  ̯] → ;
ʲ → j;
ᵐ → m;
ⁿ → n;
ᵑ → ŋ;
::NFC;

# TODO: Diphthongs probably need more work.

# Romanian [sekujesk] → [sekujask], for emitting سیکویاسک not سیکویسک
$UVowel [j $IVowel] [e {e̞} $SchwaVowel] → uia;

# Kazakh Аягөз [ɑjɑɡy̯ʉz] → [ɑjɑɡiuz], to emit آیاگیوز not آیاگووز
yʉ → iu;

::NULL;

# Vowels
$Boundary {$SchwaVowel ː?} → ای;
$SchwaVowel ː → ی;
{[$SchwaVowel e {e̞}]} [^[:L:][:M:][:N:][\.]] → ه;
$SchwaVowel → ;

$Boundary {$IVowel ː?} → ای;
$IVowel ː? j? → ی;

$Boundary {$UVowel ː?} → او;
$UVowel ː? → و;

$Boundary {$AVowel ː?} → ا;
$AVowel ː? → ا;

$Boundary {$DarkAVowel ː?} → آ;
$DarkAVowel ː? → ا;

# Shadda for long (geminated) consonants
ː → ّ;

# Affricates
[{t͡ʃ} ʧ] → چ;

# Clicks
[ɡ g ɠ k] $Click → کچ;
[n ɲ]? $Click → نچ;

# Nasal stops
[{m̥} m ɱ] → م;
[{n̼̊} {n̼} {n̥} n {ɳ̊} ɳ {ɲ̊} {ɲ̥} ɲ] → ن;
[{ŋ̊} ŋ {ɴ̥} ɴ] k → نک;
[{ŋ̊} ŋ {ɴ̥} ɴ] [ɡ g]? → نگ;

# Non-nasal stops
[p {p̪}] → پ;
[b {b̪} ɓ] → ب;
[{d̼} d ɗ ᶑ] → د;
[{t̼} t] → ت;
[ʈ] → ط;
[ɖ] → ض;
c → چ;
ɟ → دج;
k → ک;
[ɡ g ɠ] → گ;
[q ɢ ʡ ʛ] → ق;
ʔ → ;

# Sibilant fricatives
s → س;
z → ز;
[ʃ ʂ ɕ ʄ] → ش;
[ʒ ʐ ʑ] → ژ;

# Non-sibilant fricatives
[ɸ f] → ف;
[β v] → و;
[{θ̼} θ {θ̱}] → ث;
[{ð̼} ð {ð̠}] → ذ;
ç → ش;
ʝ $IVowel? ː? → ی;
[x χ] → خ;
[ɣ ʁ] → غ;
ħ → ح;
ʕ → ع;
[h ɦ {ʔ̞}] → ه;

# Approximants, trills, flaps
ʋ → و;
ʙ → بر;
{r̝} → رژ;
[{ɹ̥} {ɹ} {ɻ̊} {ɻ} {ɾ̥} ɾ {ɽ̊} ɽ {r̼} {r̥} r] → ر;
[{ʀ̥} ʀ] → غ;
ʜ → ح;
ʢ → ع;
j $IVowel? ː? → ی;

# Laterals
ɬ → شل;
ɮ → ژل;
{[{ʎ̥} ʎ]} [^ $IVowel j ʝ] → لی;
[{l̼} {l̥} l {ɭ̊} ɭ {ʎ̥} ʎ] → ل;
[ʟ {ʟ̠}] → غ;

# Independent pass for misc cleanup.
::NULL;

# Strip off syllable markers
\. → ;  

# Sequences of three or more ووو look very confusing; we shorten them.
# Polish Darłowo [darwɔvɔ] → داروو → داروووو
ووو+ → وو;

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