map: Allow using the Levenshtein algorithm to search

This commit is contained in:
Kerollmops 2018-04-22 20:06:56 +02:00
parent 0581b296bb
commit 96d2fbcd3d
No known key found for this signature in database
GPG key ID: 016ACC0DC3ADC318
4 changed files with 78 additions and 6 deletions

View file

@ -1,4 +1,6 @@
extern crate env_logger;
extern crate fst;
extern crate fst_levenshtein;
extern crate futures;
extern crate raptor;
extern crate tokio_minihttp;
@ -8,6 +10,8 @@ extern crate url;
use std::io;
use fst_levenshtein::Levenshtein;
use fst::{IntoStreamer, Streamer};
use futures::future;
use tokio_minihttp::{Request, Response, Http};
use tokio_proto::TcpServer;
@ -34,8 +38,18 @@ impl Service for MainService {
if let Some((_, key)) = url.query_pairs().find(|&(ref k, _)| k == "q") {
let key = key.to_lowercase();
let values = self.map.get(&key).map(|a| &a[..10]);
resp.body(&format!("{:?}", values));
let lev = Levenshtein::new(&key, 2).unwrap();
let mut body = String::new();
let mut stream = self.map.search(lev).into_stream();
while let Some((key, values)) = stream.next() {
let values = &values[..values.len().min(10)];
body.push_str(&format!("{:?} {:?}\n", key, values));
}
resp.body(&body);
}
future::ok(resp)

View file

@ -5,6 +5,7 @@ extern crate serde_json;
#[macro_use] extern crate serde_derive;
extern crate smallvec;
use std::ops::{Deref, DerefMut};
use std::io::Write;
use std::fs::File;
use std::path::Path;
@ -50,6 +51,45 @@ impl MultiMap {
pub fn get<K: AsRef<[u8]>>(&self, key: K) -> Option<&[u64]> {
self.map.get(key).map(|i| &*self.values[i as usize])
}
pub fn search<A: fst::Automaton>(&self, aut: A) -> StreamBuilder<A> {
StreamBuilder {
inner: self.map.search(aut),
values: &self.values,
}
}
}
pub struct StreamBuilder<'a, A: fst::Automaton> {
inner: fst::map::StreamBuilder<'a, A>,
values: &'a [SmallVec32<u64>],
}
impl<'a, A: fst::Automaton> Deref for StreamBuilder<'a, A> {
type Target = fst::map::StreamBuilder<'a, A>;
fn deref(&self) -> &Self::Target {
&self.inner
}
}
impl<'a, A: fst::Automaton> DerefMut for StreamBuilder<'a, A> {
fn deref_mut(&mut self) -> &mut Self::Target {
&mut self.inner
}
}
impl<'a, A: fst::Automaton> fst::IntoStreamer<'a> for StreamBuilder<'a, A> {
type Item = (&'a str, &'a [u64]);
type Into = Stream<'a, A>;
fn into_stream(self) -> Self::Into {
Stream {
inner: self.inner.into_stream(),
values: self.values,
}
}
}
pub struct Stream<'a, A: fst::Automaton = fst::automaton::AlwaysMatch> {