A support-operations analyst merges a batch of customer ticket transcripts into one block of text and wants to know which word customers repeat the most. A few terms — internal codenames, canned-response boilerplate, and the like — are not meaningful signal and must be excluded from the count. Word matching should ignore case ("Refund" and "refund" are the same word) and should ignore punctuation, splitting the text into words wherever anything other than a letter appears.

Input format

• Line 1: a string text, the merged ticket transcript. It contains English letters, spaces, and the punctuation characters !?'.,;, mixing uppercase and lowercase letters.
• Line 2: an integer k, the number of blocked words.
• Line 3: k space-separated blocked words, each written in lowercase letters only (this line may be empty when k is 0).

Output format

Print, in lowercase, the word from text that occurs the most times among words that are not in the blocked list (matching case-insensitively, with punctuation stripped). It is guaranteed that at least one word survives blocking and that the word with the maximum count among non-blocked words is unique.

Constraints

• 1 <= length of text <= 1000
• 0 <= k <= 20
• Each blocked word has length between 1 and 10, consists of lowercase English letters only, and all blocked words are distinct.
• At least one word of text (case-insensitively, punctuation stripped) is not in the blocked list.
• The maximum-count word among the non-blocked words is unique.

Bob hit a ball, the hit BALL flew far after it was hit.
1
hit

ball

a, a, a, a, b,b,b,c, c
0

a