From 2b1a915eadbf26320f2c0baf988abc434102816d Mon Sep 17 00:00:00 2001 From: Bond-009 Date: Thu, 5 May 2022 19:59:17 +0200 Subject: Merge pull request #7604 from Jellifi007/fixes-diactritics Co-authored-by: Cody Robibero (cherry picked from commit 8d1d9734381472b301deb0118bbb8da2a769a65e) Signed-off-by: crobibero --- src/Jellyfin.Extensions/StringExtensions.cs | 42 +++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+) (limited to 'src') diff --git a/src/Jellyfin.Extensions/StringExtensions.cs b/src/Jellyfin.Extensions/StringExtensions.cs index 3a77072539..dadc9f1d5c 100644 --- a/src/Jellyfin.Extensions/StringExtensions.cs +++ b/src/Jellyfin.Extensions/StringExtensions.cs @@ -1,4 +1,8 @@ using System; +using System.Diagnostics; +using System.Globalization; +using System.Text; +using System.Text.RegularExpressions; namespace Jellyfin.Extensions { @@ -7,6 +11,44 @@ namespace Jellyfin.Extensions /// public static class StringExtensions { + // Matches non-conforming unicode chars + // https://mnaoumov.wordpress.com/2014/06/14/stripping-invalid-characters-from-utf-16-strings/ + private static readonly Regex _nonConformingUnicode = new Regex("([\ud800-\udbff](?![\udc00-\udfff]))|((? + /// Removes the diacritics character from the strings. + /// + /// The string to act on. + /// The string without diacritics character. + public static string RemoveDiacritics(this string text) + { + string withDiactritics = _nonConformingUnicode + .Replace(text, string.Empty) + .Normalize(NormalizationForm.FormD); + + var withoutDiactritics = new StringBuilder(); + foreach (char c in withDiactritics) + { + UnicodeCategory uc = CharUnicodeInfo.GetUnicodeCategory(c); + if (uc != UnicodeCategory.NonSpacingMark) + { + withoutDiactritics.Append(c); + } + } + + return withoutDiactritics.ToString().Normalize(NormalizationForm.FormC); + } + + /// + /// Checks wether or not the specified string has diacritics in it. + /// + /// The string to check. + /// True if the string has diacritics, false otherwise. + public static bool HasDiacritics(this string text) + { + return !string.Equals(text, text.RemoveDiacritics(), StringComparison.Ordinal); + } + /// /// Counts the number of occurrences of [needle] in the string. /// -- cgit v1.2.3