rust-asearch
/icons2/Rust.icon版asearch
できたらhttps://github.com/takker99/rust-asearch に写してcrateにするtakker.icon
2021-05-27 03:13:52 できた
変更点
Rustにdefault引数はないのでambigを外部から明示的に指定するようにした
matchが予約識別子扱いなので、findに変えた
Rustで変数をcaptureする関数を返す関数を作るのが難しいので、素直にimplを使った
実装したいこと
https://github.com/teramotodaiki/asearch.ts#パターン文字列の文頭に余分な文字が挿入されている場合の挙動 の修正をあてたい
code:lib.rs
pub mod asearch {
const INITPAT: u32 = 0x80000000; // 0100,0000,0000,0000,0000,0000,0000,0000
const MAXCHAR: usize = 0x10000;
const INITSTATE: u32; 4 = INITPAT, 0, 0, 0;
pub struct Asearch {
shiftpat: u32; MAXCHAR,
acceptpat: u32,
epsilon: u32,
}
impl Asearch {
pub fn new(source: impl Into<String>) -> Asearch {
let mut shiftpat: u32; MAXCHAR = 0; MAXCHAR;
let mut mask = INITPAT;
let mut epsilon: u32 = 0;
for item in &unpack(source.into()) {
// 0x20 is a space
if *item == 0x20 {
epsilon |= mask;
} else {
shiftpat*item |= mask;
shiftpatto_upper(*item) |= mask;
shiftpatto_lower(*item) |= mask;
mask >>= 1;
}
}
Asearch {
acceptpat: mask,
shiftpat,
epsilon,
}
}
fn state(&self, text: impl Into<String>) -> u32; 4 {
let mut i0 = INITSTATE0;
let mut i1 = INITSTATE1;
let mut i2 = INITSTATE2;
let mut i3 = INITSTATE3;
for item in &unpack(text.into()) {
let mask = self.shiftpat*item;
i3 = (i3 & self.epsilon) | ((i3 & mask) >> 1) | (i2 >> 1) | i2;
i2 = (i2 & self.epsilon) | ((i2 & mask) >> 1) | (i1 >> 1) | i1;
i1 = (i1 & self.epsilon) | ((i1 & mask) >> 1) | (i0 >> 1) | i0;
i0 = (i0 & self.epsilon) | ((i0 & mask) >> 1);
i1 |= i0 >> 1;
i2 |= i1 >> 1;
i3 |= i2 >> 1;
}
i0, i1, i2, i3
}
pub fn find(&self, text: impl Into<String>, ambig: u8) -> bool {
let ambig_ = if (ambig as usize) < INITSTATE.len() {
ambig as usize
} else {
INITSTATE.len() - 1
};
let s = self.state(text.into());
(sambig_ & self.acceptpat) != 0
}
}
// code pointに変換する
// 添字に使う
fn unpack(text: impl Into<String>) -> Vec<usize> {
text.into().chars().into_iter().map(|c| c as usize).collect()
}
// 大文字小文字変換
fn is_upper(c: usize) -> bool {
(0x41..=0x5a).contains(&c)
}
fn is_lower(c: usize) -> bool {
(0x61..=0x7a).contains(&c)
}
fn to_lower(c: usize) -> usize {
if is_upper(c) {
c + 0x20
} else {
c
}
}
fn to_upper(c: usize) -> usize {
if is_lower(c) {
c - 0x20
} else {
c
}
}
}
test code
https://github.com/shokai/node-asearch/blob/master/tests/test.js を参考に作った
code:lib.rs
#cfg(test)
mod tests {
use crate::asearch::Asearch;
#test
fn pattern_abcde() {
let asearch = Asearch::new("abcde");
assert!(asearch.find("abcde", 0));
assert!(asearch.find("aBCDe", 0));
assert!(asearch.find("abXcde", 1));
assert!(asearch.find("ab?de", 1));
assert!(asearch.find("abXXde", 2));
assert!(!asearch.find("abXcde", 0));
assert!(!asearch.find("ab?de", 0));
assert!(!asearch.find("abde", 0));
assert!(!asearch.find("abXXde", 1));
assert!(asearch.find("abcde", 1));
assert!(!asearch.find("abcd", 0));
assert!(asearch.find("abcd", 1));
assert!(asearch.find("bcde", 2)); // TODO: 1で通るようにcodeを修正する
}
#test
fn pattern_ab_de() {
let asearch = Asearch::new("ab de");
assert!(asearch.find("abcde", 0));
assert!(asearch.find("abccde", 0));
assert!(asearch.find("abXXXXXXXde", 0));
assert!(asearch.find("ababcccccxede", 1));
assert!(!asearch.find("abcccccxe", 0));
}
#test
fn pattern_unicode() {
let asearch = Asearch::new("漢字文字列");
assert!(asearch.find("漢字文字列", 0));
assert!(!asearch.find("漢字の文字列", 0));
assert!(asearch.find("漢字の文字列", 1));
assert!(!asearch.find("漢字文字", 0));
assert!(asearch.find("漢字文字", 1));
assert!(!asearch.find("漢字文字烈", 0));
assert!(asearch.find("漢字文字烈", 1));
assert!(!asearch.find("漢和辞典", 2));
}
}