! Copyright (C) 2008, 2009 Doug Coleman, Daniel Ehrenberg.
! See http://factorcode.org/license.txt for BSD license.
-USING: accessors kernel math math.order words combinators locals
-ascii unicode.categories combinators.short-circuit sequences
-fry macros arrays assocs sets classes ;
+USING: accessors kernel math math.order words combinators
+combinators.smart combinators.short-circuit locals
+unicode.categories sequences fry macros arrays assocs sets
+classes unicode.script unicode.data ;
+FROM: ascii => ascii? ;
+FROM: sets => members ;
IN: regexp.classes
-SINGLETONS: any-char any-char-no-nl
-letter-class LETTER-class Letter-class digit-class
+SINGLETONS: dot letter-class LETTER-class Letter-class digit-class
alpha-class non-newline-blank-class
ascii-class punctuation-class java-printable-class blank-class
control-character-class hex-digit-class java-blank-class c-identifier-class
unmatchable-class terminator-class word-boundary-class ;
-SINGLETONS: beginning-of-input ^ end-of-input $ end-of-file ;
+SINGLETONS: beginning-of-input ^ end-of-input $ end-of-file
+^unix $unix word-break ;
-TUPLE: range from to ;
-C: <range> range
+TUPLE: range-class { from read-only } { to read-only } ;
+C: <range-class> range-class
-GENERIC: class-member? ( obj class -- ? )
+TUPLE: primitive-class { class read-only } ;
+C: <primitive-class> primitive-class
+
+TUPLE: category-class { category read-only } ;
+C: <category-class> category-class
-M: t class-member? ( obj class -- ? ) 2drop t ;
+TUPLE: category-range-class { category read-only } ;
+C: <category-range-class> category-range-class
-M: integer class-member? ( obj class -- ? ) = ;
+TUPLE: script-class { script read-only } ;
+C: <script-class> script-class
+
+GENERIC: class-member? ( obj class -- ? )
-M: range class-member? ( obj class -- ? )
- [ from>> ] [ to>> ] bi between? ;
+M: t class-member? ( obj class -- ? ) 2drop t ; inline
-M: any-char class-member? ( obj class -- ? )
- 2drop t ;
+M: integer class-member? ( obj class -- ? ) = ; inline
-M: any-char-no-nl class-member? ( obj class -- ? )
- drop CHAR: \n = not ;
+M: range-class class-member? ( obj class -- ? )
+ [ from>> ] [ to>> ] bi between? ; inline
M: letter-class class-member? ( obj class -- ? )
- drop letter? ;
-
+ drop letter? ; inline
+
M: LETTER-class class-member? ( obj class -- ? )
- drop LETTER? ;
+ drop LETTER? ; inline
M: Letter-class class-member? ( obj class -- ? )
- drop Letter? ;
+ drop Letter? ; inline
M: ascii-class class-member? ( obj class -- ? )
- drop ascii? ;
+ drop ascii? ; inline
M: digit-class class-member? ( obj class -- ? )
- drop digit? ;
+ drop digit? ; inline
: c-identifier-char? ( ch -- ? )
{ [ alpha? ] [ CHAR: _ = ] } 1|| ;
M: c-identifier-class class-member? ( obj class -- ? )
- drop c-identifier-char? ;
+ drop c-identifier-char? ; inline
M: alpha-class class-member? ( obj class -- ? )
- drop alpha? ;
+ drop alpha? ; inline
: punct? ( ch -- ? )
"!\"#$%&'()*+,-./:;<=>?@[\\]^_`{|}~" member? ;
M: punctuation-class class-member? ( obj class -- ? )
- drop punct? ;
+ drop punct? ; inline
: java-printable? ( ch -- ? )
{ [ alpha? ] [ punct? ] } 1|| ;
M: java-printable-class class-member? ( obj class -- ? )
- drop java-printable? ;
+ drop java-printable? ; inline
M: non-newline-blank-class class-member? ( obj class -- ? )
- drop { [ blank? ] [ CHAR: \n = not ] } 1&& ;
+ drop { [ blank? ] [ CHAR: \n = not ] } 1&& ; inline
M: control-character-class class-member? ( obj class -- ? )
- drop control? ;
+ drop control? ; inline
: hex-digit? ( ch -- ? )
{
} 1|| ;
M: hex-digit-class class-member? ( obj class -- ? )
- drop hex-digit? ;
+ drop hex-digit? ; inline
: java-blank? ( ch -- ? )
{
CHAR: \s CHAR: \t CHAR: \n
- HEX: b HEX: 7 CHAR: \r
+ 0xb 0x7 CHAR: \r
} member? ;
M: java-blank-class class-member? ( obj class -- ? )
- drop java-blank? ;
+ drop java-blank? ; inline
M: unmatchable-class class-member? ( obj class -- ? )
- 2drop f ;
+ 2drop f ; inline
M: terminator-class class-member? ( obj class -- ? )
- drop "\r\n\u000085\u002029\u002028" member? ;
-
-M: ^ class-member? ( obj class -- ? )
- 2drop f ;
-
-M: $ class-member? ( obj class -- ? )
- 2drop f ;
-
-M: f class-member? 2drop f ;
+ drop "\r\n\u000085\u002029\u002028" member? ; inline
-TUPLE: primitive-class class ;
-C: <primitive-class> primitive-class
-
-TUPLE: or-class seq ;
-
-TUPLE: not-class class ;
+M: f class-member? 2drop f ; inline
-TUPLE: and-class seq ;
+: same? ( obj1 obj2 quot1: ( obj1 -- val1 ) quot2: ( obj2 -- val2 ) -- ? )
+ bi* = ; inline
-GENERIC: combine-and ( class1 class2 -- combined ? )
+M: script-class class-member?
+ [ script-of ] [ script>> ] same? ; inline
-: replace-if-= ( object object -- object ? )
- over = ;
+M: category-class class-member?
+ [ category ] [ category>> ] same? ; inline
-M: object combine-and replace-if-= ;
+M: category-range-class class-member? inline
+ [ category first ] [ category>> ] same? ; inline
-M: t combine-and
- drop t ;
+TUPLE: not-class { class read-only } ;
-M: f combine-and
- nip t ;
+PREDICATE: not-integer < not-class class>> integer? ;
-M: not-class combine-and
- class>> 2dup = [ 2drop f t ] [
- dup integer? [
- 2dup swap class-member?
- [ 2drop f f ]
- [ drop t ] if
- ] [ 2drop f f ] if
- ] if ;
+UNION: simple-class
+ primitive-class range-class dot ;
+PREDICATE: not-simple < not-class class>> simple-class? ;
-M: integer combine-and
- swap 2dup class-member? [ drop t ] [ 2drop f t ] if ;
-
-GENERIC: combine-or ( class1 class2 -- combined ? )
+M: not-class class-member?
+ class>> class-member? not ; inline
-M: object combine-or replace-if-= ;
+TUPLE: or-class { seq read-only } ;
-M: t combine-or
- nip t ;
+M: or-class class-member?
+ seq>> [ class-member? ] with any? ; inline
-M: f combine-or
- drop t ;
+TUPLE: and-class { seq read-only } ;
-M: not-class combine-or
- class>> = [ t t ] [ f f ] if ;
+M: and-class class-member?
+ seq>> [ class-member? ] with all? ; inline
-M: integer combine-or
- 2dup swap class-member? [ drop t ] [ 2drop f f ] if ;
+DEFER: substitute
: flatten ( seq class -- newseq )
'[ dup _ instance? [ seq>> ] [ 1array ] if ] map concat ; inline
-: try-combine ( elt1 elt2 quot -- combined/f ? )
- 3dup call [ [ 3drop ] dip t ] [ drop swapd call ] if ; inline
-
-DEFER: answer
-
-:: try-cancel ( elt1 elt2 empty -- combined/f ? )
- [ elt1 elt2 empty answer dup elt1 = not ] try-combine ;
-
-:: prefix-combining ( seq elt quot: ( elt1 elt2 -- combined/f ? ) -- newseq )
- f :> combined!
- seq [ elt quot call swap combined! ] find drop
- [ seq remove-nth combined prefix ]
- [ seq elt prefix ] if* ; inline
-
-: combine-by ( seq quot -- new-seq )
- { } swap '[ _ prefix-combining ] reduce ; inline
-
-:: seq>instance ( seq empty class -- instance )
+:: sequence>instance ( seq empty class -- instance )
seq length {
{ 0 [ empty ] }
{ 1 [ seq first ] }
- [ drop class new seq >>seq ]
+ [ drop seq { } like class boa ]
} case ; inline
-:: combine ( seq quot: ( elt1 elt2 -- combined/f ? ) empty class -- newseq )
- seq class flatten
- [ quot try-combine ] combine-by
- ! [ empty try-cancel ] combine-by ! This makes the algorithm O(n^4)
- empty class seq>instance ; inline
+TUPLE: class-partition integers not-integers simples not-simples and or other ;
+
+: partition-classes ( seq -- class-partition )
+ members
+ [ integer? ] partition
+ [ not-integer? ] partition
+ [ simple-class? ] partition
+ [ not-simple? ] partition
+ [ and-class? ] partition
+ [ or-class? ] partition
+ class-partition boa ;
+
+: class-partition>sequence ( class-partition -- seq )
+ {
+ [ integers>> ]
+ [ not-integers>> ]
+ [ simples>> ]
+ [ not-simples>> ]
+ [ and>> ]
+ [ or>> ]
+ [ other>> ]
+ } cleave>array concat ;
+
+: repartition ( partition -- partition' )
+ ! This could be made more efficient; only and and or are effected
+ class-partition>sequence partition-classes ;
+
+: filter-not-integers ( partition -- partition' )
+ dup
+ [ simples>> ] [ not-simples>> ] [ or>> ] tri
+ 3append and-class boa
+ '[ [ class>> _ class-member? ] filter ] change-not-integers ;
+
+: answer-ors ( partition -- partition' )
+ dup [ not-integers>> ] [ not-simples>> ] [ simples>> ] tri 3append
+ '[ [ _ [ t substitute ] each ] map ] change-or ;
+
+: contradiction? ( partition -- ? )
+ {
+ [ [ simples>> ] [ not-simples>> ] bi intersects? ]
+ [ other>> f swap member? ]
+ } 1|| ;
+
+: make-and-class ( partition -- and-class )
+ answer-ors repartition
+ [ t swap remove ] change-other
+ dup contradiction?
+ [ drop f ]
+ [ filter-not-integers class-partition>sequence members t and-class sequence>instance ] if ;
: <and-class> ( seq -- class )
- [ combine-and ] t and-class combine ;
+ dup and-class flatten partition-classes
+ dup integers>> length {
+ { 0 [ nip make-and-class ] }
+ { 1 [ integers>> first [ '[ _ swap class-member? ] all? ] keep and ] }
+ [ 3drop f ]
+ } case ;
+
+: filter-integers ( partition -- partition' )
+ dup
+ [ simples>> ] [ not-simples>> ] [ and>> ] tri
+ 3append or-class boa
+ '[ [ _ class-member? ] reject ] change-integers ;
+
+: answer-ands ( partition -- partition' )
+ dup [ integers>> ] [ not-simples>> ] [ simples>> ] tri 3append
+ '[ [ _ [ f substitute ] each ] map ] change-and ;
+
+: tautology? ( partition -- ? )
+ {
+ [ [ simples>> ] [ not-simples>> ] bi intersects? ]
+ [ other>> t swap member? ]
+ } 1|| ;
-M: and-class class-member?
- seq>> [ class-member? ] with all? ;
+: make-or-class ( partition -- and-class )
+ answer-ands repartition
+ [ f swap remove ] change-other
+ dup tautology?
+ [ drop t ]
+ [ filter-integers class-partition>sequence members f or-class sequence>instance ] if ;
: <or-class> ( seq -- class )
- [ combine-or ] f or-class combine ;
-
-M: or-class class-member?
- seq>> [ class-member? ] with any? ;
+ dup or-class flatten partition-classes
+ dup not-integers>> length {
+ { 0 [ nip make-or-class ] }
+ { 1 [
+ not-integers>> first
+ [ class>> '[ _ swap class-member? ] any? ] keep or
+ ] }
+ [ 3drop t ]
+ } case ;
GENERIC: <not-class> ( class -- inverse )
M: t <not-class> drop f ;
M: f <not-class> drop t ;
-M: not-class class-member?
- class>> class-member? not ;
+: <minus-class> ( a b -- a-b )
+ <not-class> 2array <and-class> ;
-M: primitive-class class-member?
- class>> class-member? ;
+: <sym-diff-class> ( a b -- a~b )
+ 2array [ <or-class> ] [ <and-class> ] bi <minus-class> ;
-UNION: class primitive-class not-class or-class and-class range ;
+M: primitive-class class-member?
+ class>> class-member? ; inline
TUPLE: condition question yes no ;
C: <condition> condition
M: not-class answer
[ class>> ] 2dip answer <not-class> ;
+GENERIC# substitute 1 ( class from to -- new-class )
+M: object substitute answer ;
+M: not-class substitute [ <not-class> ] bi@ answer ;
+
: assoc-answer ( table question answer -- new-table )
- '[ _ _ answer ] assoc-map
- [ nip ] assoc-filter ;
+ '[ _ _ substitute ] assoc-map sift-values ;
: assoc-answers ( table questions answer -- new-table )
'[ _ assoc-answer ] each ;
: condition-states ( condition -- states )
dup condition? [
[ yes>> ] [ no>> ] bi
- [ condition-states ] bi@ append prune
+ [ condition-states ] bi@ union
] [ 1array ] if ;
: condition-at ( condition assoc -- new-condition )