Problem Narrative

You’re building a search deduplication service for a large e-commerce marketplace (10M+ daily searches). Users often misspell or reorder characters when searching for SKUs and promo codes (e.g., "SAVE10" vs "10EVAS"). To reduce redundant indexing and improve cache hit rates, you need a fast function that determines whether two strings are anagrams under strict normalization rules.

Task

Implement a function:

• Function: are_anagrams(s: str, t: str) -> bool

Return True if s and t are anagrams after applying the following rules:

1. Case-insensitive: treat uppercase and lowercase letters as the same.
2. Ignore non-alphanumeric characters: spaces, punctuation, and symbols should be skipped.
3. Count-based: strings are anagrams if they contain the same characters with the same frequencies after normalization.

Input/Output

• Input: two strings s and t
• Output: boolean indicating whether they are anagrams under the rules above

Examples

Example 1

• Input: s = "Dormitory", t = "Dirty room!!"
• Output: True
• Explanation:
  • Normalize: "dormitory" and "dirtyroom"
  • Both have identical character counts → anagrams.

Example 2

• Input: s = "promo-code-2026", t = "2026 code promo"
• Output: True
• Explanation:
  • Normalize: "promocode2026" and "2026codepromo"
  • Same multiset of characters → anagrams.

Constraints

• 0 <= len(s), len(t) <= 2 * 10^5
• Strings may contain any ASCII characters

Notes / Clarifications

• Two empty strings are anagrams.
• If, after normalization, the lengths differ, you can immediately return False.
• Aim for linear time to support high QPS traffic in production services.