Skip to content

Commit

Permalink
turkish: Adjust stringdefs to match other uses
Browse files Browse the repository at this point in the history
Use {sc} for s-cedilla and {i} for dotless-i.
  • Loading branch information
ojwb committed Nov 1, 2023
1 parent e2ae50c commit b97e9cd
Showing 1 changed file with 25 additions and 25 deletions.
50 changes: 25 additions & 25 deletions algorithms/turkish.sbl
Original file line number Diff line number Diff line change
Expand Up @@ -80,22 +80,22 @@ stringescapes { }
/* Special characters in Unicode Latin-1 and Latin Extended-A */
stringdef cc '{U+00E7}' // LATIN SMALL LETTER C WITH CEDILLA
stringdef g~ '{U+011F}' // LATIN SMALL LETTER G WITH BREVE
stringdef i' '{U+0131}' // LATIN SMALL LETTER I WITHOUT DOT
stringdef i '{U+0131}' // LATIN SMALL LETTER I WITHOUT DOT
stringdef o" '{U+00F6}' // LATIN SMALL LETTER O WITH DIAERESIS
stringdef s, '{U+015F}' // LATIN SMALL LETTER S WITH CEDILLA
stringdef sc '{U+015F}' // LATIN SMALL LETTER S WITH CEDILLA
stringdef u" '{U+00FC}' // LATIN SMALL LETTER U WITH DIAERESIS

booleans ( continue_stemming_noun_suffixes )

groupings ( vowel U vowel1 vowel2 vowel3 vowel4 vowel5 vowel6)

define vowel 'ae{i'}io{o"}u{u"}'
define U '{i'}iu{u"}'
define vowel 'ae{i}io{o"}u{u"}'
define U '{i}iu{u"}'

// the vowel grouping definitions below are used for checking vowel harmony
define vowel1 'a{i'}ou' // vowels that can end with suffixes containing 'a'
define vowel1 'a{i}ou' // vowels that can end with suffixes containing 'a'
define vowel2 'ei{o"}{u"}' // vowels that can end with suffixes containing 'e'
define vowel3 'a{i'}' // vowels that can end with suffixes containing 'i''
define vowel3 'a{i}' // vowels that can end with suffixes containing 'i''
define vowel4 'ei' // vowels that can end with suffixes containing 'i'
define vowel5 'ou' // vowels that can end with suffixes containing 'o' or 'u'
define vowel6 '{o"}{u"}' // vowels that can end with suffixes containing 'o"' or 'u"'
Expand All @@ -113,7 +113,7 @@ backwardmode (
(
('a' goto vowel1) or
('e' goto vowel2) or
('{i'}' goto vowel3) or
('{i}' goto vowel3) or
('i' goto vowel4) or
('o' goto vowel5) or
('{o"}' goto vowel6) or
Expand Down Expand Up @@ -162,8 +162,8 @@ backwardmode (
)

define mark_possessives as (
among ('m{i'}z' 'miz' 'muz' 'm{u"}z'
'n{i'}z' 'niz' 'nuz' 'n{u"}z' 'm' 'n')
among ('m{i}z' 'miz' 'muz' 'm{u"}z'
'n{i}z' 'niz' 'nuz' 'n{u"}z' 'm' 'n')
(mark_suffix_with_optional_U_vowel)
)

Expand All @@ -174,7 +174,7 @@ backwardmode (
)

define mark_lArI as (
among ('leri' 'lar{i'}')
among ('leri' 'lar{i}')
)

define mark_yU as (
Expand All @@ -185,12 +185,12 @@ backwardmode (

define mark_nU as (
check_vowel_harmony
among ('n{i'}' 'ni' 'nu' 'n{u"}')
among ('n{i}' 'ni' 'nu' 'n{u"}')
)

define mark_nUn as (
check_vowel_harmony
among ('{i'}n' 'in' 'un' '{u"}n')
among ('{i}n' 'in' 'un' '{u"}n')
(mark_suffix_with_optional_n_consonant)
)

Expand Down Expand Up @@ -243,23 +243,23 @@ backwardmode (

define mark_yUm as (
check_vowel_harmony
among ('{i'}m' 'im' 'um' '{u"}m')
among ('{i}m' 'im' 'um' '{u"}m')
(mark_suffix_with_optional_y_consonant)
)

define mark_sUn as (
check_vowel_harmony
among ('s{i'}n' 'sin' 'sun' 's{u"}n' )
among ('s{i}n' 'sin' 'sun' 's{u"}n' )
)

define mark_yUz as (
check_vowel_harmony
among ('{i'}z' 'iz' 'uz' '{u"}z')
among ('{i}z' 'iz' 'uz' '{u"}z')
(mark_suffix_with_optional_y_consonant)
)

define mark_sUnUz as (
among ('s{i'}n{i'}z' 'siniz' 'sunuz' 's{u"}n{u"}z')
among ('s{i}n{i}z' 'siniz' 'sunuz' 's{u"}n{u"}z')
)

define mark_lAr as (
Expand All @@ -269,24 +269,24 @@ backwardmode (

define mark_nUz as (
check_vowel_harmony
among ('n{i'}z' 'niz' 'nuz' 'n{u"}z')
among ('n{i}z' 'niz' 'nuz' 'n{u"}z')
)

define mark_DUr as (
check_vowel_harmony
among ('t{i'}r' 'tir' 'tur' 't{u"}r' 'd{i'}r' 'dir' 'dur' 'd{u"}r')
among ('t{i}r' 'tir' 'tur' 't{u"}r' 'd{i}r' 'dir' 'dur' 'd{u"}r')
)

define mark_cAsInA as (
among ('cas{i'}na' 'cesine')
among ('cas{i}na' 'cesine')
)

define mark_yDU as (
check_vowel_harmony
among ('t{i'}m' 'tim' 'tum' 't{u"}m' 'd{i'}m' 'dim' 'dum' 'd{u"}m'
't{i'}n' 'tin' 'tun' 't{u"}n' 'd{i'}n' 'din' 'dun' 'd{u"}n'
't{i'}k' 'tik' 'tuk' 't{u"}k' 'd{i'}k' 'dik' 'duk' 'd{u"}k'
't{i'}' 'ti' 'tu' 't{u"}' 'd{i'}' 'di' 'du' 'd{u"}')
among ('t{i}m' 'tim' 'tum' 't{u"}m' 'd{i}m' 'dim' 'dum' 'd{u"}m'
't{i}n' 'tin' 'tun' 't{u"}n' 'd{i}n' 'din' 'dun' 'd{u"}n'
't{i}k' 'tik' 'tuk' 't{u"}k' 'd{i}k' 'dik' 'duk' 'd{u"}k'
't{i}' 'ti' 'tu' 't{u"}' 'd{i}' 'di' 'du' 'd{u"}')
(mark_suffix_with_optional_y_consonant)
)

Expand All @@ -298,7 +298,7 @@ backwardmode (

define mark_ymUs_ as (
check_vowel_harmony
among ('m{i'}{s,}' 'mi{s,}' 'mu{s,}' 'm{u"}{s,}')
among ('m{i}{sc}' 'mi{sc}' 'mu{sc}' 'm{u"}{sc}')
(mark_suffix_with_optional_y_consonant)
)

Expand Down Expand Up @@ -428,7 +428,7 @@ backwardmode (
define append_U_to_stems_ending_with_d_or_g as (
[] ('d' or 'g') goto vowel

(('a' or '{i'}') <- '{i'}')
(('a' or '{i}') <- '{i}')
or
(('e' or 'i') <- 'i')
or
Expand Down

0 comments on commit b97e9cd

Please sign in to comment.