Simplify the API some more

This commit is contained in:
Dirkjan Ochtman 2021-02-10 13:23:37 +01:00
parent 4338ff2c0c
commit 9dd1cf089d
4 changed files with 13 additions and 14 deletions

View File

@ -1,6 +1,6 @@
[package]
name = "instant-segment"
version = "0.6.1"
version = "0.7.0"
authors = ["Dirkjan Ochtman <dirkjan@ochtman.nl>"]
edition = "2018"
license = "Apache-2.0"

View File

@ -8,5 +8,7 @@ benchmark_main!(benches);
fn short(bench: &mut Bencher) {
let segmenter = instant_segment::test_data::segmenter();
let mut search = instant_segment::Search::default();
bench.iter(|| segmenter.segment("thisisatest", &mut search));
bench.iter(|| {
let _ = segmenter.segment("thisisatest", &mut search);
});
}

View File

@ -56,11 +56,13 @@ impl Segmenter {
/// Requires that the input `text` consists of lowercase ASCII characters only. Otherwise,
/// returns `Err(InvalidCharacter)`. The `search` parameter contains caches that are used
/// segmentation; passing it in allows the callers to reuse the cache allocations.
///
/// The segmentation result can be retrieved through the `Search::split()` method.
pub fn segment(&self, input: &str, search: &mut Search) -> Result<(), InvalidCharacter> {
pub fn segment<'a>(
&self,
input: &str,
search: &'a mut Search,
) -> Result<impl Iterator<Item = &'a str> + ExactSizeIterator, InvalidCharacter> {
SegmentState::new(Ascii::new(input)?, &self, search).run();
Ok(())
Ok(search.result.iter().map(|v| v.as_str()))
}
fn score(&self, word: &str, previous: Option<&str>) -> f64 {
@ -200,11 +202,6 @@ impl Search {
}
self.result.clear();
}
/// Get the segmentation result
pub fn split(&self) -> impl Iterator<Item = &str> + ExactSizeIterator {
self.result.iter().map(|v| v.as_str())
}
}
type MemoKey = (Range<usize>, Range<usize>);

View File

@ -10,14 +10,14 @@ pub fn run(segmenter: &Segmenter) {
}
pub fn assert_segments(s: &[&str], search: &mut Search, segmenter: &Segmenter) {
segmenter.segment(&s.join(""), search).unwrap();
let cmp = search.split().collect::<Vec<_>>();
let words = segmenter.segment(&s.join(""), search).unwrap();
let cmp = words.collect::<Vec<_>>();
assert_eq!(cmp, s);
}
pub fn check_segments(s: &[&str], search: &mut Search, segmenter: &Segmenter) -> bool {
match segmenter.segment(&s.join(""), search) {
Ok(()) => s == search.split().collect::<Vec<_>>(),
Ok(words) => s == words.collect::<Vec<_>>(),
Err(_) => false,
}
}