(<[^/]([^>]*[^/>])?>) # xml start tag, e.g. <tag>
|(</[^>]*>) # xml end tag, e.g. </tag>
|(<[^>]*/>) # xml empty tag, e.g. <tag />
|
# Get text at least two characters long that begins with a letter, 
# number, or CJK ideograph, that may also contain some 
# punctuation, and that ends with a letter, number, or CJK 
# ideograph.
  (
    (
      \p{Lu}                # Uppercase Letter
      |\p{Ll}               # Lowercase Letter
      |\p{Lt}               # Titlecase Letter
      |\p{Nd}               # Decimal Digit Number, i.e. 0-9
      |\p{Nl}               # Letter Number, e.g. Roman numerals
      |\p{No}               # Other Number
      |[\x{4e00}-\x{9fa5}]  # CJK Unified Ideographs
      |\x{3007}             # Ideographic number zero
      |[\x{3021}-\x{3029}]  # Hangzhou numerals 1-9
    )
    (
      (
        \p{Lu}
        |\p{Ll}
        |\p{Lt}
        |\p{Nd}
        |\p{Nl}
        |\p{No}
        |[-._:']            # some punctuation
        |[\x{4e00}-\x{9fa5}]
        |\x{3007}
        |[\x{3021}-\x{3029}]
      )*
      (
        \p{Lu}
        |\p{Ll}
        |\p{Lt}
        |\p{Nd}
        |\p{Nl}
        |\p{No}
        |[\x{4e00}-\x{9fa5}]
        |\x{3007}
        |[\x{3021}-\x{3029}]
      )
    )?
  )