From 20601f70b2156e35767c1f3e2a466ae6c939ce28 Mon Sep 17 00:00:00 2001
From: Fuwn <contact@fuwn.me>
Date: Mon, 8 Jan 2024 22:10:40 -0800
Subject: feat(match): better schedule match algorithm

---
 bun.lockb                                     | Bin 146194 -> 146958 bytes
 package.json                                  |   2 ++
 src/lib/Media/Anime/Airing/Subtitled/match.ts |  23 +++++++++++++++++------
 3 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/bun.lockb b/bun.lockb
index e31b637b..0612850c 100755
Binary files a/bun.lockb and b/bun.lockb differ
diff --git a/package.json b/package.json
index 57f8e9b4..73b159a3 100644
--- a/package.json
+++ b/package.json
@@ -16,6 +16,7 @@
     "@sveltejs/kit": "^1.20.4",
     "@types/fast-levenshtein": "^0.0.4",
     "@types/jsdom": "^21.1.6",
+    "@types/string-similarity": "^4.0.2",
     "@typescript-eslint/eslint-plugin": "^5.45.0",
     "@typescript-eslint/parser": "^5.45.0",
     "eslint": "^8.28.0",
@@ -40,6 +41,7 @@
     "lz-string": "^1.5.0",
     "modern-screenshot": "^4.4.33",
     "rss-parser": "^3.13.0",
+    "string-similarity": "^4.0.4",
     "wanakana": "^5.3.1"
   }
 }
diff --git a/src/lib/Media/Anime/Airing/Subtitled/match.ts b/src/lib/Media/Anime/Airing/Subtitled/match.ts
index 1a0c6b2d..1bf31b6f 100644
--- a/src/lib/Media/Anime/Airing/Subtitled/match.ts
+++ b/src/lib/Media/Anime/Airing/Subtitled/match.ts
@@ -3,6 +3,7 @@ import type { Media } from '../../../../AniList/media';
 import settings from '$stores/settings';
 import type { SubsPlease } from '$lib/Media/Anime/Airing/Subtitled/subsPlease';
 import levenshtein from 'fast-levenshtein';
+import stringSimilarity from 'string-similarity';
 
 export interface Time {
 	title: string;
@@ -33,7 +34,7 @@ const secondsUntil = (targetTime: string, targetDay: string) => {
 const normalizeTitle = (title: string | null) => {
 	return (title || '')
 		.toLowerCase()
-		.replace(/season \d+|s\d+|\W/g, '')
+		.replace(/season \d+|s\d+/g, '')
 		.replace(/\b(\d)(st|nd|rd|th)\b/g, '$1')
 		.replace(/\b(part|pt)\b/gi, '')
 		.trim();
@@ -67,14 +68,24 @@ export const findClosestMedia = (media: Media[], matchFor: string) => {
 	if (!matchFor) return null;
 
 	let bestFitMedia: Media | null = null;
-	let smallestDistance = Infinity;
+	let smallestDistance = -Infinity;
 
 	media.forEach((m) => {
-		[m.title.romaji, m.title.english, ...m.synonyms].filter(Boolean).forEach((title) => {
-			const normalizedItemTitle = normalizeTitle(title);
-			const distance = levenshtein.get(normalizeTitle(matchFor), normalizedItemTitle);
+		const titles = [m.title.romaji, m.title.english, ...m.synonyms].filter(Boolean);
+
+		titles.forEach((title) => {
+			const normalisedTitle = normalizeTitle(title);
+			const normalisedMatchFor = normalizeTitle(matchFor);
+			const distance = stringSimilarity.compareTwoStrings(normalisedMatchFor, normalisedTitle);
 
-			if (distance < smallestDistance && distance < Math.max(3, normalizedItemTitle.length * 0.4)) {
+			if (
+				distance > smallestDistance &&
+				(normalisedMatchFor
+					.split(' ')
+					.filter((word) => titles.some((title) => normalizeTitle(title).includes(word))).length >=
+					normalisedMatchFor.split(' ').length ||
+					titles.some((title) => normalizeTitle(title).includes(normalisedMatchFor)))
+			) {
 				smallestDistance = distance;
 				bestFitMedia = m;
 			}
-- 
cgit v1.2.3