New: Remove websites in parentheses before parsing

This commit is contained in:
Mark McDowall 2024-06-21 17:10:58 -07:00 committed by Mark McDowall
parent 45fe585944
commit ea4fe392a0
2 changed files with 2 additions and 1 deletions

View file

@ -23,6 +23,7 @@ namespace NzbDrone.Core.Test.ParserTests
[TestCase("[www.test-hyphen.ca] - Series (2011) S01", "Series (2011)")]
[TestCase("test123.ca - Series Time S02 720p HDTV x264 CRON", "Series Time")]
[TestCase("[www.test-hyphen123.co.za] - Series Title S01E01", "Series Title")]
[TestCase("(seriesawake.com) Series Super - 57 [720p] [English Subbed]", "Series Super")]
public void should_not_parse_url_in_name(string postTitle, string title)
{

View file

@ -514,7 +514,7 @@ namespace NzbDrone.Core.Parser
// Valid TLDs http://data.iana.org/TLD/tlds-alpha-by-domain.txt
private static readonly RegexReplace WebsitePrefixRegex = new RegexReplace(@"^(?:\[\s*)?(?:www\.)?[-a-z0-9-]{1,256}\.(?<!Naruto-Kun\.)(?:[a-z]{2,6}\.[a-z]{2,6}|xn--[a-z0-9-]{4,}|[a-z]{2,})\b(?:\s*\]|[ -]{2,})[ -]*",
private static readonly RegexReplace WebsitePrefixRegex = new RegexReplace(@"^(?:(?:\[|\()\s*)?(?:www\.)?[-a-z0-9-]{1,256}\.(?<!Naruto-Kun\.)(?:[a-z]{2,6}\.[a-z]{2,6}|xn--[a-z0-9-]{4,}|[a-z]{2,})\b(?:\s*(?:\]|\))|[ -]{2,})[ -]*",
string.Empty,
RegexOptions.IgnoreCase | RegexOptions.Compiled);