Add more words to formatter and start of word list

ajayyy · Dec 14, 2024 · 626bf42 · 626bf42
1 parent 6c5ce21
commit 626bf42
Show file tree

Hide file tree

Showing 2 changed files with 46 additions and 4 deletions.
diff --git a/src/titles/titleFormatter.ts b/src/titles/titleFormatter.ts
@@ -1,7 +1,7 @@
 import { VideoID } from "../../maze-utils/src/video";
 import Config, { TitleFormatting } from "../config/config";
 import { getTitleFormatting, shouldCleanEmojis } from "../config/channelOverrides";
-import { acronymBlocklist, allowlistedWords, fancyTextConversions, notStartOfSentence, titleCaseDetectionNotCapitalized, titleCaseNotCapitalized } from "./titleFormatterData";
+import { acronymBlocklist, allowlistedStartOfWords, allowlistedWords, fancyTextConversions, notStartOfSentence, titleCaseDetectionNotCapitalized, titleCaseNotCapitalized } from "./titleFormatterData";
 import { chromeP } from "../../maze-utils/src/browserApi";
 import type { LanguageIdentifier } from "cld3-asm";
 
@@ -316,12 +316,14 @@ function isFirstLetterCapital(word: string): boolean {
 
 function forceKeepFormatting(word: string, ignorePunctuation = true): boolean {
     let result = !!word.match(/^>/)
-        || listHasWord(allowlistedWords, word);
+        || listHasWord(allowlistedWords, word)
+        || listHasStartOfWord(allowlistedStartOfWords, word);
 
     if (ignorePunctuation) {
         const withoutPunctuation = word.replace(/[:?.!+\]]+$|^[[+:/]+/, "");
         if (word !== withoutPunctuation) {
-            result ||= listHasWord(allowlistedWords, withoutPunctuation);
+            result ||= listHasWord(allowlistedWords, withoutPunctuation)
+                || listHasStartOfWord(allowlistedStartOfWords, word);
         }
     }
 
@@ -582,6 +584,17 @@ function listHasWord(list: Set<string>, word: string): boolean {
     return list.has(word.replace(/[[「〈《【〔⦗『〖〘<({:〙〗』⦘〕】》〉」)}\]]/g, ""))
 }
 
+function listHasStartOfWord(list: Set<string>, word: string): boolean {
+    word = word.replace(/[[「〈《【〔⦗『〖〘<({:〙〗』⦘〕】》〉」)}\]]/g, "");
+
+    for (const item of list) {
+        if (word.startsWith(item)) {
+            return true;
+        }
+    }
+
+    return false;
+}
 
 export async function localizeHtmlPageWithFormatting(): Promise<void> {
     // Localize by replacing __MSG_***__ meta tags

diff --git a/src/titles/titleFormatterData.ts b/src/titles/titleFormatterData.ts
@@ -163,7 +163,6 @@ export const allowlistedWords = new Set([
     "GTA",
     "bell hooks",
     "TOOOL",
-    "PCECD",
     "drawholic",
     "TF2",
     "L4D",
@@ -322,6 +321,36 @@ export const allowlistedWords = new Set([
     "WWF",
     "WWI",
     "XML",
+    "LGBTQIA2S+",
+    "NVFBC",
+    "DALL-E",
+    "DALL·E",
+    "iilluminaughtii",
+    "TOML",
+    "DJ",
+    "SPTV",
+    "VST",
+    "EQ"
+]);
+
+// Can be switched to a trie structure if it grows
+export const allowlistedStartOfWords = new Set([
+    "osu!",
+    "de_",
+    "cs_",
+    "ar_",
+    "as_",
+    "es_",
+    "dz_",
+    "fy_",
+    "aim_",
+    "zm_",
+    "jail_",
+    "bb_",
+    "gg_",
+    "awp_",
+    "df_",
+    "deathrun_",
 ]);
 
 export const acronymBlocklist = new Set([