You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

68 lines
2.3 KiB
Racket

#lang debug racket/base
(require racket/file
racket/fasl
racket/runtime-path)
(provide (all-defined-out))
(define-runtime-path wordidx-file "compiled/words/words-index.rktd")
(struct word-rec (word charint length) #:prefab)
(define (make-word-recs)
(define reverse-string (compose1 list->string reverse string->list))
(define omit-words (map reverse-string (file->lines "data/omit.rktd")))
(for/vector ([w (in-lines (open-input-file "data/words.rktd"))]
#:when (and (not (regexp-match "'" w)) ; no apostrophes
(regexp-match #rx"^[A-Za-z]+$" w) ; no accented letters
(not (member w omit-words)))) ; no bad words
(word-rec w
(word->charidx w)
(string-length w))))
(define (regenerate-word-index!)
(make-parent-directory* wordidx-file)
(s-exp->fasl
(make-word-recs)
(open-output-file wordidx-file #:exists 'replace)))
(define wordrecs
(fasl->s-exp (open-input-file (and
(unless (file-exists? wordidx-file)
(regenerate-word-index!))
wordidx-file))))
(define (char->bitindex c)
;; 64-bit layout
;; __________ZYXWVUTSRQPONMLKJIHGFEDCBA______zyxwvutsrqponmlkjihgfedcba
(cond
[(char<=? #\a c #\z) (- (char->integer c) 97)] ; 97 = (char->integer #\a)
[(char<=? #\A c #\Z) (- (char->integer c) 33)] ; 65 = (char->integer #\A)
[else 0]))
(define (word->charidx word)
(apply bitwise-ior
(for/list ([c (in-string word)])
(expt 2 (char->bitindex c)))))
(define (bitindex->char i)
(cond
[(<= 0 i 26) (integer->char (+ i 97))]
[(<= 32 i 59) (integer->char (+ i 33))]
[else (error 'bong)]))
(define (charidx->chars int)
(for/list ([i (in-range 64)]
#:when (bitwise-bit-set? int i))
(bitindex->char i)))
(define (contains-char? charidx-entry c)
(bitwise-bit-set? charidx-entry (char->bitindex c)))
(define capitalized-mask
(for/sum ([i (in-range 32 59)])
(expt 2 i)))
(define (capitalized? charidx-entry)
;; a cap only appears at the beginning of a word,
;; so it's sufficient to test whether a cap exists in the idx
(positive? (bitwise-and charidx-entry capitalized-mask)))