From e5b23bfd165461be012c8a64f37bcbb3205a9447 Mon Sep 17 00:00:00 2001
From: JOLIMAITRE Matthieu <matthieu@imagevo.fr>
Date: Tue, 30 Apr 2024 16:38:45 +0200
Subject: [PATCH] fix incorrect prunning of composite words

---
 src/lib/dict.ts | 23 +++++++++++++++++++++--
 1 file changed, 21 insertions(+), 2 deletions(-)

diff --git a/src/lib/dict.ts b/src/lib/dict.ts
index 8e08a95..8bc4c5d 100644
--- a/src/lib/dict.ts
+++ b/src/lib/dict.ts
@@ -17,8 +17,8 @@ export class Dict {
 		for (const word of content.split("\n")) {
 			const word_ = word.trim().toLowerCase();
 			if (word_.length !== length) continue;
-			for (const forbidden of [" ", "-", "."]) if (word_.includes(forbidden)) continue;
-			words.add(remove_accent(word_));
+			if (contains_any(word_, [" ", "-", "."])) continue;
+			words.add(remove_accents(word_));
 		}
 		return new Dict(words, length);
 	}
@@ -27,3 +27,22 @@ export class Dict {
 		return `Dict { ${this.words.size} words }`;
 	}
 }
+
+export function contains_any(text: string, words: string[]) {
+	for (const word of words) if (text.includes(word)) return true;
+	return false;
+}
+
+export function remove_accents(text: string) {
+	const accents = [
+		...[["à", "a"], ["â", "a"], ["ä", "a"]],
+		...[["ç", "c"]],
+		...[["é", "e"], ["è", "e"], ["ê", "e"], ["ë", "e"]],
+		...[["î", "i"], ["ï", "i"]],
+		...[["ô", "o"], ["ö", "o"]],
+		...[["û", "u"]],
+	];
+	let result = text;
+	for (const [accent, alternative] of accents) result = result.replaceAll(accent, alternative);
+	return result;
+}