You are given $n$ types of balls, with colors numbered from $1$ to $n$. There are $a_i$ balls of the $i$-th color. Little L wants to arrange these balls into a sequence such that there is no non-empty prefix or non-empty suffix that is "neat". A prefix or suffix is defined as "neat" if and only if the occurrences of all $n$ colors within it are equal. Your task is to calculate the number of distinct sequences.
Some necessary clarifications provided by the kind temporaryDO:
- A prefix (or suffix) of length $len$ of a sequence refers to the $len$ balls at the leftmost (or rightmost) positions of the sequence. A non-empty prefix or suffix refers to a prefix or suffix where $len \neq 0$.
- For a sequence of balls $S$, let $S_i$ denote the color of the $i$-th ball from the left. Two sequences $S$ and $T$ are considered distinct if and only if there exists a position $i$ such that $S_i \neq T_i$.
To avoid the need for arbitrary-precision arithmetic, you only need to output the result modulo $998244353$.
Input
The first line contains an integer $n$, representing the number of color types.
The next line contains $n$ space-separated integers, where the $i$-th positive integer $a_i$ represents the number of balls of the $i$-th color.
Output
Output a single integer representing the number of distinct sequences modulo $998244353$.
Examples
Input 1
3
1 1 2
Output 1
2
Note 1
The two valid arrangements are: 1 3 3 2 and 2 3 3 1. Clearly, there are no other sequences that satisfy the condition.
Input 2
5
5 6 7 8 9
Output 2
322192262
Subtasks
| Test Case ID | $n=$ | $a_i\le$ |
|---|---|---|
| $1$ | $2$ | $500$ |
| $2$ | $2$ | $2\times 10^5$ |
| $3$ | $2$ | $2\times 10^5$ |
| $4$ | $50$ | $2\times 10^3$ |
| $5$ | $75$ | $2\times 10^3$ |
| $6$ | $100$ | $2\times 10^3$ |
| $7$ | $50$ | $10^5$ |
| $8$ | $75$ | $1.5\times 10^5$ |
| $9$ | $90$ | $1.75\times 10^5$ |
| $10$ | $100$ | $2\times 10^5$ |
For $100\%$ of the data, it is guaranteed that $n \le 100$ and all $a_i \le 200,000$.
After submitting your code, the pretest data will be evaluated and the results returned. This problem's pretest data consists of 10 test cases, where the $i$-th test case satisfies the constraints of the $i$-th row in the table above.
Note: The evaluation results of the pretest data have no relation to the final evaluation results.