GitHub/gf-rgl
1
0
Fork 1
mirror of https://github.com/GrammaticalFramework/gf-rgl.git synced 2026-05-28 09:28:54 -06:00
Code Issues Packages Projects Releases Wiki Activity

(Ara) More fixes to hamza rules

Browse Source
This commit is contained in:
Inari Listenmaa
2018-11-08 15:17:09 +01:00
parent ec0d36bd1d
commit e6d57f3b03
1 changed files with 54 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

105
src/arabic/OrthoAra.gf
View File

@@ -2,67 +2,70 @@ resource OrthoAra = open Prelude, Predef in {
flags coding=utf8 ;
oper
oper
vow : pattern Str = #("َ" | "ِ" | "ُ" | "ً" | "ٍ" | "ٌ") ;
vow : pattern Str = #("َ" | "ِ" | "ُ" | "ً" | "ٍ" | "ٌ") ;
weak : pattern Str = #("و"|"ي") ;
weak : pattern Str = #("و"|"ي") ;
-- "Sun letters": assimilate with def. article
sun : pattern Str = #("ت"|"ث"|"د"|"ذ"|"ر"|"ز"|"س"|"ش"|"ص"|"ض"|"ط"|"ظ"|"ل"|"ن") ;
-- "Sun letters": assimilate with def. article
sun : pattern Str = #("ت"|"ث"|"د"|"ذ"|"ر"|"ز"|"س"|"ش"|"ص"|"ض"|"ط"|"ظ"|"ل"|"ن") ;
-- Shadda: https://www.unicode.org/L2/L2017/17253-arabic-ordering.pdf
fixShd : Str -> Str -> Str = \word,suffix ->
case <word,suffix> of {
-- <x + "ّ", v@#vow + y> => x + v + "ّ" + y ;
<x + v@#vow, "ّ" + y> => x + "ّ" + v + y ;
_ => word + suffix
} ;
-- IL: using this to reuse patterns for weak verbs, might be strange/wrong
rmSukun : Str -> Str = \s -> case s of {
x + "ْ" + y => x + y ;
_ => s
-- Shadda: https://www.unicode.org/L2/L2017/17253-arabic-ordering.pdf
fixShd : Str -> Str -> Str = \word,suffix ->
case <word,suffix> of {
-- <x + "ّ", v@#vow + y> => x + v + "ّ" + y ;
<x + v@#vow, "ّ" + y> => x + "ّ" + v + y ;
_ => word + suffix
} ;
-- Hamza
hamza : pattern Str = #("ء"|"؟") ;
-- IL: using this to reuse patterns for weak verbs, might be strange/wrong
rmSukun : Str -> Str = \s -> case s of {
x + "ْ" + y => x + y ;
_ => s
} ;
rectifyHmz: Str -> Str = \word ->
case word of {
l@(""|"ال") + ("أ"|"أَ") + #hamza + "ْ" + tail => l + "آ" + tail;
l@(""|"ال") + ("أ"|"أَ") + #hamza + tail => l + "آ" + tail;
l@(""|"ال") + #hamza + v@("َ"|"ُ") + tail => l + "أ" + v + tail;
l@(""|"ال") + #hamza + v@("ِ") + tail => l + "إ" + v + tail;
-- Hamza
hamza : pattern Str = #("ء"|"؟") ;
head + v1@("ِ"|"ُ"|"َ"|"ْ"|"ا"|"ي"|"و") + #hamza + v2@(""|"ُ"|"َ"|"ْ"|"ِ") => head + v1 + (tHmz v1) + v2;
head + #hamza + tail => head + (bHmz (dp 2 head) (take 2 tail)) + tail; --last head , take 1 tail
_ => word
};
rectifyHmz : Str -> Str = \word ->
case word of {
l@(""|"ال") + ("أ"|"أَ") + #hamza + "ْ" + tail => l + "آ" + tail;
l@(""|"ال") + ("أ"|"أَ") + #hamza + tail => l + "آ" + tail;
l@(""|"ال") + #hamza + v@("َ"|"ُ") + tail => l + "أ" + v + tail;
l@(""|"ال") + #hamza + v@("ِ") + tail => l + "إ" + v + tail;
head + v1@("ِ"|"ُ"|"َ"|"ْ"|"ا"|"ي"|"و")
+ #hamza + v2@(#vow|"ْ") + tail =>
case v2 of { "ْ" => head + v1 + tHmz v1 + tail ; -- unsure about this /IL
_ => head + v1 + tHmz v1 + v2 + tail } ;
--hamza at beginning of word (head)
hHmz : Str -> Str = \d ->
case d of {
"ِ" => "إ";
_ => "أ"
};
head + #hamza + tail => head + (bHmz (dp 2 head) (take 2 tail)) + tail; --last head , take 1 tail
_ => word
};
--hamza in middle of word (body)
bHmz : Str -> Str -> Str = \d1,d2 ->
case <d1,d2> of {
<"ِ",_> | <_,"ِ"> => "ئ";
<"ُ",_> | <_,"ُ"> => "ؤ";
<"َ",_> | <_,"َ"> => "أ";
_ => "ء"
};
--hamza at beginning of word (head)
hHmz : Str -> Str = \d ->
case d of {
"ِ" => "إ";
_ => "أ"
};
--hamza carrier sequence
tHmz : Str -> Str = \d ->
case d of {
"ِ" => "ئ";
"ُ" => "ؤ";
"َ" => "أ";
"ْ"|"ا"|"و"|"ي" => "ء"
};
--hamza in middle of word (body)
bHmz : Str -> Str -> Str = \d1,d2 ->
case <d1,d2> of {
<"ِ",_> | <_,"ِ"> => "ئ";
<"ُ",_> | <_,"ُ"> => "ؤ";
<"َ",_> | <_,"َ"> => "أ";
_ => "ء"
};
--hamza carrier sequence
tHmz : Str -> Str = \d ->
case d of {
"ِ" => "ئ";
"ُ" => "ؤ";
"َ" => "أ";
"ْ"|"ا"|"و"|"ي" => "ء"
};
}