aboutsummaryrefslogtreecommitdiff
path: root/src/Jellyfin.Extensions/StringExtensions.cs
diff options
context:
space:
mode:
authorClaus Vium <cvium@users.noreply.github.com>2022-10-07 09:57:16 +0200
committerGitHub <noreply@github.com>2022-10-07 09:57:16 +0200
commit81b04ddbb54201d2e07310e3c700cca8a94d8955 (patch)
treea4e9aeb70e35b3e47a7d8af17b328e88a1c77ed0 /src/Jellyfin.Extensions/StringExtensions.cs
parent6bf71c0fd355e9c95a1e142019d9bc5cce34200d (diff)
parent14027f962ce074623fd89967ca9565bbeb785066 (diff)
Merge branch 'master' into providermanager-cleanup
Diffstat (limited to 'src/Jellyfin.Extensions/StringExtensions.cs')
-rw-r--r--src/Jellyfin.Extensions/StringExtensions.cs41
1 files changed, 41 insertions, 0 deletions
diff --git a/src/Jellyfin.Extensions/StringExtensions.cs b/src/Jellyfin.Extensions/StringExtensions.cs
index 3a7707253..b19be071b 100644
--- a/src/Jellyfin.Extensions/StringExtensions.cs
+++ b/src/Jellyfin.Extensions/StringExtensions.cs
@@ -1,4 +1,7 @@
using System;
+using System.Globalization;
+using System.Text;
+using System.Text.RegularExpressions;
namespace Jellyfin.Extensions
{
@@ -7,6 +10,44 @@ namespace Jellyfin.Extensions
/// </summary>
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]))|((?<![\ud800-\udbff])[\udc00-\udfff])|(\ufffd)");
+
+ /// <summary>
+ /// Removes the diacritics character from the strings.
+ /// </summary>
+ /// <param name="text">The string to act on.</param>
+ /// <returns>The string without diacritics character.</returns>
+ 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);
+ }
+
+ /// <summary>
+ /// Checks whether or not the specified string has diacritics in it.
+ /// </summary>
+ /// <param name="text">The string to check.</param>
+ /// <returns>True if the string has diacritics, false otherwise.</returns>
+ public static bool HasDiacritics(this string text)
+ {
+ return !string.Equals(text, text.RemoveDiacritics(), StringComparison.Ordinal);
+ }
+
/// <summary>
/// Counts the number of occurrences of [needle] in the string.
/// </summary>