Switchboard Peg Cipher Decoder

An antique telephone switchboard encodes a lowercase message as a log of peg presses, recorded as a string containing only the digits '0'-'9' and the hash character '#'. The log is read left to right and decodes one letter at a time using two possible token forms:

• A single digit token from '1' to '9' decodes to the letter that many positions after 'a' (so '1' decodes to 'a', '2' to 'b', ..., '9' to 'i').
• A two-digit token from "10" to "26" immediately followed by a '#' decodes to the letter that many positions after 'a' (so "10#" decodes to 'j', "11#" to 'k', ..., "26#" to 'z').

The log is guaranteed to represent a valid encoding of some lowercase message, meaning it can always be decoded unambiguously by scanning left to right and, at each position, checking whether a two-digit-plus-hash token is present before falling back to a single digit. Decode the log into its original lowercase message.

Input format

A single line: the log string s, 1 to 2000 characters long, consisting only of the characters '0'-'9' and '#'.

Output format

A single line: the decoded lowercase message.

Constraints

• 1 <= |s| <= 2000.
• s consists only of digits '0'-'9' and '#'.
• s is guaranteed to be a valid, unambiguous encoding of some non-empty lowercase message.

10#11#12#

jkl

1326#

acz