1 ! Copyright (C) 2008 Daniel Ehrenberg.
\r
2 ! See http://factorcode.org/license.txt for BSD license.
\r
3 USING: combinators.short-circuit sequences io.files
\r
4 io.encodings.ascii kernel values splitting accessors math.parser
\r
5 ascii io assocs strings math namespaces make sorting combinators
\r
6 math.order arrays unicode.normalize unicode.data locals
\r
7 unicode.syntax macros sequences.deep words unicode.breaks
\r
8 quotations combinators.short-circuit ;
\r
9 IN: unicode.collation
\r
14 TUPLE: weight primary secondary tertiary ignorable? ;
\r
16 : parse-weight ( string -- weight )
\r
17 "]" split but-last [
\r
18 weight new swap rest unclip CHAR: * = swapd >>ignorable?
\r
19 swap "." split first3 [ hex> ] tri@
\r
20 [ >>primary ] [ >>secondary ] [ >>tertiary ] tri*
\r
23 : parse-line ( line -- code-poing weight )
\r
24 ";" split1 [ [ blank? ] trim ] bi@
\r
25 [ " " split [ hex> ] "" map-as ] [ parse-weight ] bi* ;
\r
27 : parse-ducet ( stream -- ducet )
\r
28 lines filter-comments
\r
29 [ parse-line ] H{ } map>assoc ;
\r
31 "resource:basis/unicode/collation/allkeys.txt"
\r
32 ascii <file-reader> parse-ducet to: ducet
\r
34 ! Fix up table for long contractions
\r
35 : help-one ( assoc key -- )
\r
36 ! Need to be more general? Not for DUCET, apparently
\r
37 2 head 2dup swap key? [ 2drop ] [
\r
38 [ [ 1string swap at ] with { } map-as concat ]
\r
42 : insert-helpers ( assoc -- )
\r
43 dup keys [ length 3 >= ] filter
\r
44 [ help-one ] with each ;
\r
46 ducet insert-helpers
\r
48 : base ( char -- base )
\r
50 { [ dup HEX: 3400 HEX: 4DB5 between? ] [ drop HEX: FB80 ] } ! Extension A
\r
51 { [ dup HEX: 20000 HEX: 2A6D6 between? ] [ drop HEX: FB80 ] } ! Extension B
\r
52 { [ dup HEX: 4E00 HEX: 9FC3 between? ] [ drop HEX: FB40 ] } ! CJK
\r
53 [ drop HEX: FBC0 ] ! Other
\r
56 : AAAA ( char -- weight )
\r
57 [ base ] [ -15 shift ] bi + HEX: 20 2 f weight boa ;
\r
59 : BBBB ( char -- weight )
\r
60 HEX: 7FFF bitand HEX: 8000 bitor 0 0 f weight boa ;
\r
62 : illegal? ( char -- ? )
\r
63 { [ "Noncharacter_Code_Point" property? ] [ category "Cs" = ] } 1|| ;
\r
65 : derive-weight ( char -- weights )
\r
68 [ [ AAAA ] [ BBBB ] bi 2array ] if ;
\r
71 building get empty? [ 0 ] [ building get peek peek ] if ;
\r
73 : blocked? ( char -- ? )
\r
74 combining-class dup { 0 f } member?
\r
75 [ drop last non-starter? ]
\r
76 [ last combining-class = ] if ;
\r
78 : possible-bases ( -- slice-of-building )
\r
79 building get dup [ first non-starter? not ] find-last
\r
80 drop [ 0 ] unless* tail-slice ;
\r
82 :: ?combine ( char slice i -- ? )
\r
83 [let | str [ i slice nth char suffix ] |
\r
85 [ str i slice set-nth ] when
\r
89 dup blocked? [ 1string , ] [
\r
90 dup possible-bases dup length
\r
91 [ ?combine ] with with contains?
\r
92 [ drop ] [ 1string , ] if
\r
95 : string>graphemes ( string -- graphemes )
\r
96 [ [ add ] each ] { } make ;
\r
98 : graphemes>weights ( graphemes -- weights )
\r
100 dup weight? [ 1array ] ! From tailoring
\r
101 [ dup ducet at [ ] [ derive-weight ] ?if ] if
\r
102 ] { } map-as concat ;
\r
104 : append-weights ( weights quot -- )
\r
105 [ [ ignorable?>> not ] filter ] dip
\r
106 map [ zero? not ] filter % 0 , ; inline
\r
108 : variable-weight ( weight -- )
\r
109 dup ignorable?>> [ primary>> ] [ drop HEX: FFFF ] if , ;
\r
111 : weights>bytes ( weights -- byte-array )
\r
114 [ [ primary>> ] append-weights ]
\r
115 [ [ secondary>> ] append-weights ]
\r
116 [ [ tertiary>> ] append-weights ]
\r
117 [ [ variable-weight ] each ]
\r
122 : completely-ignorable? ( weight -- ? )
\r
123 [ primary>> ] [ secondary>> ] [ tertiary>> ] tri
\r
124 [ zero? ] tri@ and and ;
\r
126 : filter-ignorable ( weights -- weights' )
\r
128 tuck primary>> zero? and
\r
129 [ swap ignorable?>> or ]
\r
130 [ swap completely-ignorable? or not ] 2bi
\r
133 : collation-key ( string -- key )
\r
134 nfd string>graphemes graphemes>weights
\r
135 filter-ignorable weights>bytes ;
\r
138 : insensitive= ( str1 str2 levels-removed -- ? )
\r
140 [ collation-key ] dip
\r
141 [ [ 0 = not ] trim-right but-last ] times
\r
145 : primary= ( str1 str2 -- ? )
\r
148 : secondary= ( str1 str2 -- ? )
\r
151 : tertiary= ( str1 str2 -- ? )
\r
154 : quaternary= ( str1 str2 -- ? )
\r
158 : w/collation-key ( str -- {str,key} )
\r
159 [ collation-key ] keep 2array ;
\r
162 : sort-strings ( strings -- sorted )
\r
163 [ w/collation-key ] map natural-sort values ;
\r
165 : string<=> ( str1 str2 -- <=> )
\r
166 [ w/collation-key ] compare ;
\r