Compress Repeated Letters

Run-length encoding shortens text by writing each run of the same character as the character followed by the length of the run: "aabcccccaaa" becomes "a2b1c5a3". Write compress(text) that returns the encoded string.

There is one catch: if the encoded string is not strictly shorter than text, return text unchanged, because compressing it would not save anything. Uppercase and lowercase letters are different characters, and a run can be long enough to need a count with several digits.

Examples

compress("aabcccccaaa")   # returns 'a2b1c5a3'
compress("abc")           # returns 'abc'
compress("zzzzzzzzzzzz")  # returns 'z12'

"abc" would turn into "a1b1c1", which is longer, so it stays as it is.

Constraints

• text has 0 to 100,000 characters.
• text contains only the letters a-z and A-Z.