Improved tag matching query efficiency, should fix some big scrape failures.

2026-10-21 05:20:19 +02:00
parent b8ad6bb10b
commit 40d56c515d
1 changed files with 5 additions and 4 deletions
@@ -79,13 +79,14 @@ function withRelations(queryBuilder, withMedia) {
 }

 async function matchTags(rawTags) {
-	const tags = rawTags
-		.map((tag) => tag?.trim().match(/[a-z0-9()-]+/ig)?.join(' ').toLowerCase())
-		.filter(Boolean);
+	const tags = Array.from(new Set(rawTags
+		.map((tag) => (typeof tag === 'string' ? tag : null)?.trim().match(/[a-z0-9()-]+/ig)?.join(' ').toLowerCase())
+		.filter(Boolean)));

+	// array binding, large batches can easily exceed the bindings limit otherwise
 	const tagEntries = await knex('tags')
 		.select('tags.id', 'tags.name', 'tags.alias_for', 'tags.implied_tag_ids')
-		.whereIn(knex.raw('lower(tags.name)'), tags);
+		.whereRaw('lower(tags.name) = ANY(:tags)', { tags });

 	const tagIdsBySlug = tagEntries
 		.reduce((acc, tag) => ({