1 ! Copyright (C) 2008, 2009 Daniel Ehrenberg.
2 ! See http://factorcode.org/license.txt for BSD license.
3 USING: combinators.short-circuit assocs math kernel sequences
4 io.files hashtables quotations splitting grouping arrays io
5 math.parser math.order byte-arrays namespaces math.bitwise
6 compiler.units parser io.encodings.ascii values interval-maps
7 ascii sets combinators locals math.ranges sorting make
8 strings.parser io.encodings.utf8 memoize simple-flat-file ;
9 FROM: namespaces => set ;
20 VALUE: compatibility-map
25 : >2ch ( a b -- c ) [ 21 shift ] dip + ;
26 : 2ch> ( c -- a b ) [ -21 shift ] [ 21 on-bits mask ] bi ;
32 : canonical-entry ( char -- seq ) canonical-map at ; inline
33 : combine-chars ( a b -- char/f ) >2ch combine-map at ; inline
34 : compatibility-entry ( char -- seq ) compatibility-map at ; inline
35 : combining-class ( char -- n ) class-map at ; inline
36 : non-starter? ( char -- ? ) combining-class { 0 f } member? not ; inline
37 : name>char ( name -- char ) name-map at ; inline
38 : char>name ( char -- name ) name-map value-at ; inline
39 : property? ( char property -- ? ) properties at interval-key? ; inline
40 : ch>lower ( ch -- lower ) simple-lower ?at drop ; inline
41 : ch>upper ( ch -- upper ) simple-upper ?at drop ; inline
42 : ch>title ( ch -- title ) simple-title ?at drop ; inline
43 : special-case ( ch -- casing-tuple ) special-casing at ; inline
45 ! For non-existent characters, use Cn
48 "Lu" "Ll" "Lt" "Lm" "Lo"
51 "Pc" "Pd" "Ps" "Pe" "Pi" "Pf" "Po"
58 MEMO: categories-map ( -- hashtable )
59 categories <enum> [ swap ] H{ } assoc-map-as ;
61 CONSTANT: num-chars HEX: 2FA1E
65 : category# ( char -- n )
66 ! There are a few characters that should be Cn
67 ! that this gives Cf or Mn
68 ! Cf = 26; Mn = 5; Cn = 29
69 ! Use a compressed array instead?
70 dup category-map ?nth [ ] [
71 dup HEX: E0001 HEX: E007F between?
73 HEX: E0100 HEX: E01EF between? 5 29 ?
77 : category ( char -- category )
78 category# categories nth ;
82 ! Loading data from UnicodeData.txt
84 : load-data ( -- data )
85 "vocab:unicode/data/UnicodeData.txt" data ;
87 : (process-data) ( index data -- newdata )
88 [ [ nth ] keep first swap ] with { } map>assoc
89 [ [ hex> ] dip ] assoc-map ;
91 : process-data ( index data -- hash )
92 (process-data) [ hex> ] assoc-map [ nip ] H{ } assoc-filter-as ;
94 : (chain-decomposed) ( hash value -- newvalue )
97 [ (chain-decomposed) ] [ 1array nip ] ?if
100 : chain-decomposed ( hash -- newhash )
101 dup [ swap (chain-decomposed) ] curry assoc-map ;
103 : first* ( seq -- ? )
104 second { [ empty? ] [ first ] } 1|| ;
106 : (process-decomposed) ( data -- alist )
107 5 swap (process-data)
108 [ " " split [ hex> ] map ] assoc-map ;
110 : exclusions-file ( -- filename )
111 "vocab:unicode/data/CompositionExclusions.txt" ;
113 : exclusions ( -- set )
114 exclusions-file utf8 file-lines
115 [ "#" split1 drop [ blank? ] trim-tail hex> ] map
118 : remove-exclusions ( alist -- alist )
119 exclusions [ dup ] H{ } map>assoc assoc-diff ;
121 : process-canonical ( data -- hash hash )
122 (process-decomposed) [ first* ] filter
124 [ second length 2 = ] filter remove-exclusions
125 [ first2 >2ch swap ] H{ } assoc-map-as
126 ] [ >hashtable chain-decomposed ] bi ;
128 : process-compatibility ( data -- hash )
130 [ dup first* [ first2 rest 2array ] unless ] map
131 [ second empty? not ] filter
132 >hashtable chain-decomposed ;
134 : process-combining ( data -- hash )
135 3 swap (process-data)
136 [ string>number ] assoc-map
137 [ nip zero? not ] assoc-filter
140 ! the maximum unicode char in the first 3 planes
142 : ?set-nth ( val index seq -- )
143 2dup bounds-check? [ set-nth ] [ 3drop ] if ;
145 :: fill-ranges ( table -- table )
146 name-map sort-values keys
147 [ { [ "first>" tail? ] [ "last>" tail? ] } 1|| ] filter
149 [ name>char ] bi@ [ [a,b] ] [ table ?nth ] bi
150 [ swap table ?set-nth ] curry each
153 :: process-category ( data -- category-listing )
154 num-chars <byte-array> :> table
155 2 data (process-data) [| char cat |
156 cat categories-map at char table ?set-nth
157 ] assoc-each table fill-ranges ;
159 : process-names ( data -- names-hash )
160 1 swap (process-data) [
161 >lower { { CHAR: \s CHAR: - } } substitute swap
162 ] H{ } assoc-map-as ;
164 : multihex ( hexstring -- string )
165 " " split [ hex> ] map sift ;
169 TUPLE: code-point lower title upper ;
171 C: <code-point> code-point
175 : set-code-point ( seq -- )
176 4 head [ multihex ] map first4
177 <code-point> swap first set ;
180 : parse-properties ( -- {{[a,b],prop}} )
181 "vocab:unicode/data/PropList.txt" data [
183 ".." split1 [ dup ] unless*
188 : properties>intervals ( properties -- assoc[str,interval] )
189 dup values members [ f ] H{ } map>assoc
190 [ [ push-at ] curry assoc-each ] keep
191 [ <interval-set> ] assoc-map ;
193 : load-properties ( -- assoc )
194 parse-properties properties>intervals ;
196 ! Special casing data
197 : load-special-casing ( -- special-casing )
198 "vocab:unicode/data/SpecialCasing.txt" data
199 [ length 5 = ] filter
200 [ [ set-code-point ] each ] H{ } make-assoc ;
203 [ process-names \ name-map set-value ]
204 [ 13 swap process-data \ simple-lower set-value ]
205 [ 12 swap process-data \ simple-upper set-value ]
206 [ 14 swap process-data simple-upper assoc-union \ simple-title set-value ]
207 [ process-combining \ class-map set-value ]
208 [ process-canonical \ canonical-map set-value \ combine-map set-value ]
209 [ process-compatibility \ compatibility-map set-value ]
210 [ process-category \ category-map set-value ]
213 : postprocess-class ( -- )
214 combine-map keys [ 2ch> nip ] map
215 [ combining-class not ] filter
216 [ 0 swap class-map set-at ] each ;
220 load-special-casing \ special-casing set-value
222 load-properties \ properties set-value
224 [ name>char [ "Invalid character" throw ] unless* ]
225 name>char-hook set-global