12.77 text.segmented-match - Segmented string matching ¶

Module: text.segmented-match ¶

Sometimes you want to search for a group of words concatenated with a delimiter, using only the prefixes of each words. For example: you want to search call-with-current-continuation using c-w-c-c. One of its popular applications is symbol completion, which Gauche provides on REPL (see Input editing).

This module provides basic parts to perform such searching.

Function: segmented-prefix? pattern word :optional separator ¶

{text.segmented-match} The pattern and word arguments must be a string, and separator must be either a character, a character set, a regexp, a predicate taking a character, or a string. When separator is omitted, #\- is assumed.

First, pattern and word are splitted into segments by separator, using string-split (see String utilities). If every segments from pattern consist a prefix of the corresponding segments from word, it returns #t. It is allowed for word to have more segments than pattern.

(segmented-match? "a-b-c" "alpha-bravo-charlie") ⇒ #t
(segmented-match? "a-b" "alpha-bravo-charlie") ⇒ #t
(segmented-match? "al-br-char" "alpha-bravo-charlie") ⇒ #t
(segmented-match? "a-b" "alpha-delta-charlie") ⇒ #t

Function: make-segmented-prefix-matcher pattern separator ¶

{text.segmented-match} This is a subfunction of segmented-match?.

Returns a procedure that takes a string word, and sees if each segment of pattern is a prefix of corresponding segment of word. If any of the pattern’s segment is not a prefix of the corresponding word’s segment, #f is returned. Otherwise, it returns a true value: If the number of word’s segments is the same as the pattern’s, #t is retured; if word has more segments, the remaining part of the word (exclusing leading separator) is returned.

((make-segmented-prefix-matcher "a-b-c" #\-) "alpha-bravo-charlie")
 ⇒ #t
((make-segmented-prefix-matcher "a-b" #\-) "alpha-bravo-charlie-delta")
 ⇒ "charlie-delta"
((make-segmented-prefix-matcher "a-b" #\-) "alpha-charlie-delta")
 ⇒ #f

This preprocesses pattern, so if you want to match many words against a single pattern, it is faster to get predicates with this procedure than applying segmented-prefix? repeatedly.