StringExtensions.cs 1.6 KB

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152
  1. #pragma warning disable CS1591
  2. using System;
  3. using System.Globalization;
  4. using System.Linq;
  5. using System.Text;
  6. using System.Text.RegularExpressions;
  7. namespace MediaBrowser.Controller.Extensions
  8. {
  9. /// <summary>
  10. /// Class BaseExtensions.
  11. /// </summary>
  12. public static class StringExtensions
  13. {
  14. public static string RemoveDiacritics(this string text)
  15. {
  16. var chars = Normalize(text, NormalizationForm.FormD)
  17. .Where(ch => CharUnicodeInfo.GetUnicodeCategory(ch) != UnicodeCategory.NonSpacingMark);
  18. return Normalize(string.Concat(chars), NormalizationForm.FormC);
  19. }
  20. private static string Normalize(string text, NormalizationForm form, bool stripStringOnFailure = true)
  21. {
  22. if (stripStringOnFailure)
  23. {
  24. try
  25. {
  26. return text.Normalize(form);
  27. }
  28. catch (ArgumentException)
  29. {
  30. // will throw if input contains invalid unicode chars
  31. // https://mnaoumov.wordpress.com/2014/06/14/stripping-invalid-characters-from-utf-16-strings/
  32. text = Regex.Replace(text, "([\ud800-\udbff](?![\udc00-\udfff]))|((?<![\ud800-\udbff])[\udc00-\udfff])", string.Empty);
  33. return Normalize(text, form, false);
  34. }
  35. }
  36. try
  37. {
  38. return text.Normalize(form);
  39. }
  40. catch (ArgumentException)
  41. {
  42. // if it still fails, return the original text
  43. return text;
  44. }
  45. }
  46. }
  47. }