asklyphe-frontend/src/main.rs

@@ -14,7 +14,6 @@
 pub mod searchbot;
 pub mod wikipedia;
 pub mod unit_converter;
-pub mod spellcheck;
 pub mod routes;
 
 use std::{env, process};

asklyphe-frontend/src/routes/search.rs

@@ -18,8 +18,6 @@ use crate::unit_converter;
 use crate::unit_converter::UnitConversion;
 use crate::wikipedia::WikipediaSummary;
 use crate::{wikipedia, Opts, ALPHA, BUILT_ON, GIT_COMMIT, VERSION, YEAR};
-use crate::spellcheck;
-use crate::spellcheck::SpellCheckResults;
 use askama::Template;
 use asklyphe_common::nats;
 use asklyphe_common::nats::bingservice::{
@@ -70,7 +68,6 @@ pub struct Complications {
     disabled: bool,
     wikipedia: Option<WikipediaSummary>,
     unit_converter: Option<UnitConversion>,
-    spellcheck: Option<SpellCheckResults>,
 }
 
 pub async fn search(
@@ -161,21 +158,18 @@ pub async fn search_js(
         let mut complications = Complications::default();
         // todo: better way of specifying that user doesn't want complications
         if !query.contains("-complications") {
-            /*let mut wikiquery = query.clone().to_lowercase();
+            let mut wikiquery = query.clone().to_lowercase();
             wikiquery.retain(|c| c.is_alphanumeric() || c.is_ascii_whitespace());
             wikiquery = wikiquery.replace(' ', "%20");
             // todo: proper url escaping
             let wikipedia_comp =
                 tokio::spawn(async move { wikipedia::get_wikipedia_page(&wikiquery, 20).await });
-            complications.wikipedia = wikipedia_comp.await.unwrap_or_default();*/
+            complications.wikipedia = wikipedia_comp.await.unwrap_or_default();
 
             let mut unit_query = query.clone().to_lowercase();
             unit_query = unit_query.replace("metre", "meter");
             let unit_comp = unit_converter::convert_unit(&unit_query);
             complications.unit_converter = unit_comp;
-
-            let corrections = spellcheck::check(&query);
-            complications.spellcheck = corrections;
         } else {
             complications.disabled = true;
             query = query.replace("-complications", "");

asklyphe-frontend/src/spellcheck.rs

@@ -1,176 +0,0 @@
-use tracing::{debug, error};
-use std::{cmp, mem};
-// TODO: cache distances of strings/substrings
-// TODO: use binary search to find direct matches, and if that fails, calculate and cache the result in BTreeMap<word: String, closest_match: String>
-// TODO: limit by number of words and word length, not max chars, and use code more like this for better readability & async:
-/*
-	let words = prepare(query).split_whitespace()
-		.filter(|qword| qword.len() > 0)
-		.map(|qword| qword.to_lowercase());
-	for word in words { // it might need to be while let Some(word) = words.next()
-		tokio::spawn(levenshtein_distance(...))
-	}
- */
-
-include!("./words.txt");
-
-// max distance before no alternatives are considered
-const MAX_DISTANCE: usize = 6;
-// max input text size before spellcheck is not run. on my laptop 13,000 chars of input takes around 4 seconds so this should be fine
-// update: got a larger word database and it doesn't take 4 seconds anymore lmao
-const MAX_QUERY_SIZE: usize = 1024;
-
-pub type SpellCheckResults = Vec<SpellCheckResult>;
-
-#[derive(Debug)]
-pub struct SpellCheckResult {
-	pub orig: String,
-	pub correction: String,
-}
-
-pub fn check(query: &String) -> Option<SpellCheckResults> {
-	error!("Query: {}", query);
-	let query: &str = {
-		if query.len() > MAX_QUERY_SIZE {
-			error!("Query is too large to be spell checked, only checking first {} chars", MAX_QUERY_SIZE);
-			query.get(0..MAX_QUERY_SIZE).unwrap()
-			// return None;
-		} else {
-			query
-		}
-	};
-
-	let distances = prepare(query).split_whitespace()
-		.filter(|qword| qword.len() > 0)
-		.map(|qword| qword.to_lowercase())
-		.map(
-			|qword| {
-				let mut exact_match = false;
-				KNOWN_WORDS.iter()
-					.map(|kword| kword.to_lowercase())
-					.map(
-						|kword|
-						(qword.clone(), kword.clone(), levenshtein_distance(&qword, &kword)))
-							// totally isn't jank at all and is the best solution totally
-							.take_while(|val| if exact_match {false} else if val.2 == 0 {exact_match = true; true} else {true})
-							// .map(|val| {error!("Val: {:?}", val); val})
-							.min_by(|a, b| a.2.cmp(&b.2)).unwrap()
-			})/*.filter_map(|word| word)*/.filter(|word| word.2 > 0 && word.2 <= MAX_DISTANCE)/*.filter(|(_, _, dist)| *dist > 0 && *dist <= MAX_DISTANCE)*/.map(|word| SpellCheckResult{orig: word.0, correction: word.1.to_owned().to_owned()})/*.filter(|(_, _, d)| *d > 0)*/
-		.map(|word| {
-			debug!("instead of '{}' did you mean '{}'?", word.orig, word.correction);
-			word
-		})
-		.collect::<Vec<_>>();
-	/*for word in &distances {
-		debug!("instead of '{}' did you mean '{}'? (distance of )", word.0, word.1/*, word.2*/);
-	}*/
-
-	if distances.len() > 0 {
-		Some(distances)
-	} else {
-		None
-	}
-	// vec![]
-}
-
-// TODO: handle symbols better, probably with a regex
-fn prepare(s: &str) -> String {
-	s.replace("\"", "")
-		.replace(",", " ")
-		.replace(":", " ")
-		.replace(".", " ")
-		.replace("/", " ")
-		.replace("&", " ")
-		.replace("!", " ")
-		.replace("?", " ")
-		/*.replace("'", "")*/
-		.replace("0", "")
-		.replace("1", "")
-		.replace("2", "")
-		.replace("3", "")
-		.replace("4", "")
-		.replace("5", "")
-		.replace("6", "")
-		.replace("7", "")
-		.replace("8", "")
-		.replace("9", "")
-}
-
-// cost of 2 for add/remove, cost of 1 for replace
-fn levenshtein_distance(a: &str, other: &str) -> usize {
-	// debug!("Self: '{}', Other: '{}'", a, other);
-	let mut dist = vec![0usize; other.len() + 1];
-	let mut dist_prev = vec![0usize; other.len() + 1];
-
-	for i in 0..=other.len() {
-		dist_prev[i] = i;
-	}
-
-	for i in 1..=a.len() {
-		dist[0] = i;
-
-		for j in 1..=other.len() {
-			if a.get(i - 1..i).unwrap() == other.get(j - 1..j).unwrap() {
-				dist[j] = dist_prev[j - 1];
-			} else {
-				dist[j] = 1 + cmp::min(
-					dist.get(j - 1).unwrap(),
-					cmp::min(dist_prev.get(j).unwrap(), dist_prev.get(j - 1).unwrap()));
-			}
-		}
-		mem::swap(&mut dist, &mut dist_prev);
-	}
-	dist_prev[other.len()]
-
-	
-
-	/*let mut distances = vec![vec![0usize; other.len() + 1]; a.len() + 1];
-	for i in 1..=a.len() {
-		distances[i][0] = i;
-	}
-
-	for j in 1..=other.len() {
-		distances[0][j] = j;
-	}
-
-	/*unsafe {
-	for i in 1..=a.len() {
-		for j in 1..=other.len() {
-			if *a.get_unchecked(i - 1..i) == *other.get_unchecked(j - 1..j) {
-				// 0
-				distances[i][j] = *distances.get_unchecked(i - 1).get_unchecked(j - 1);
-			} else {
-				// 1
-				distances[i][j] = 1 + cmp::min(
-					(*distances.get_unchecked(i - 1).get_unchecked(j - 1)),
-					cmp::min(
-						(*distances.get_unchecked(i - 1).get_unchecked(j)),
-						(*distances.get_unchecked(i).get_unchecked(j - 1))
-					)
-				);
-			}
-		}
-	}
-	}*/
-
-
-	for i in 1..=a.len()  {
-		for j in 1..=other.len() {
-			if a.get(i - 1..i).unwrap() == other.get(j - 1..j).unwrap() {
-				// 0
-				distances[i][j] = *distances.get(i - 1).unwrap().get(j - 1).unwrap();
-			} else {
-				// 1
-				distances[i][j] = 1 + cmp::min(
-					(*distances.get(i - 1).unwrap().get(j - 1).unwrap()),
-					cmp::min(
-						(*distances.get(i - 1).unwrap().get(j).unwrap()),
-						(*distances.get(i).unwrap().get(j - 1).unwrap())
-					)
-				);
-			}
-		}
-	}
-	*distances.get(a.len()).unwrap().get(other.len()).unwrap()*/
-}
-