Initial data commit

This commit is contained in:
Ingy döt Net 2023-07-01 11:58:00 -04:00
parent 72d218235f
commit f23f22d71c
199087 changed files with 3378941 additions and 0 deletions

View file

@ -0,0 +1,16 @@
USING: assocs fry io.encodings.utf8 io.files kernel sequences
sets splitting vectors ;
IN: rosettacode.inverted-index
: file-words ( file -- assoc )
utf8 file-contents " ,;:!?.()[]{}\n\r" split harvest ;
: add-to-file-list ( files file -- files )
over [ swap [ adjoin ] keep ] [ nip 1vector ] if ;
: add-to-index ( words index file -- )
'[ _ [ _ add-to-file-list ] change-at ] each ;
: (index-files) ( files index -- )
[ [ [ file-words ] keep ] dip swap add-to-index ] curry each ;
: index-files ( files -- index )
H{ } clone [ (index-files) ] keep ;
: query ( terms index -- files )
[ at ] curry map [ ] [ intersect ] map-reduce ;

View file

@ -0,0 +1,6 @@
( scratchpad ) { "f1" "f2" "f3" } index-files
--- Data stack:
H{ { "a" ~vector~ } { "is" ~vector~ } { "what" ~vector~ } { ...
( scratchpad ) { "what" "is" "it" } swap query .
V{ "f1" "f2" }