Background

"Whether you gain or lose, just laugh it off."

Problem Description

You are given $q$ ciphertexts. Each ciphertext consists of lowercase letters and “garbled codes”. There are $26$ kinds of “garbled codes”, which are the $26$ Arabic numerals from $1$ to $26$. The letter corresponding to $1$ is $a$, $2$ corresponds to $b$, and so on.

Now you are given an operation value $p$:

• When $p$ equals $1$, you need to convert all “garbled codes” in the ciphertext into lowercase letters. Specifically, you should treat each maximal consecutive digit segment as one “garbled code” and translate it into a single letter.
• When $p$ equals $2$, you need to convert all lowercase letters in the ciphertext into “garbled codes”.

Input Format

The first line contains an integer $q$, meaning there are $q$ ciphertexts.

In the next $q$ lines, each line contains an integer $p$ and a string $s_i$, representing the operation value and the ciphertext, respectively.

Output Format

Output $q$ lines. Each line contains the ciphertext after performing the required operation.

3
1 a9y3tf2p25p
2 jzc8ph4ph4ph
2 ph12pjzc21noi20ag26

aiyctfbpyp
10263816841684168
1681216102632114159201726

Hint

Constraints

This problem uses bundled testdata.

::cute-table{tuack}

For all data, it holds that:

• $1 \le q \le 2 \times 10^5$
• $1 \le \lvert s_i \rvert \le 2 \times 10^5$
• $\sum \lvert s_i \rvert \le 2 \times 10^5$
• $p \in \{1,2\}$
• It is guaranteed that the ciphertext contains only lowercase letters and digit characters.
• When $p=1$, the number represented by every maximal consecutive digit segment in the ciphertext is within $[1,26]$.

Translated by ChatGPT 5