Top-K trending search terms with per-user de-duplication

Design and implement a function/class to return the top k trending search terms from a stream of search events. Rules: Each event contains: user_id and term. Multiple searches of the same term by the same user count only once (deduplicate by (user_id, term) before counting). Popularity is defined as the number of distinct users who searched the term (or equivalently the deduplicated count). Requirements: Implement APIs: record(user_id: str, term: str) -> None: record a search event. top_k(k: int) -> List[str]: return the current top k terms. Explain your data structures and analyze time/space complexity for both APIs. Handle edge cases: What to return if k exceeds the number of distinct terms. Tie-breaking when multiple terms share the same popularity (e.g., lexicographic order or arbitrary; must be specified). Scale: Total number of events N can be large; consider efficiency. Example: Events: (u1, "apple"), (u1, "apple"), (u2, "apple"), (u2, "banana") After dedup: apple=2, banana=1 top_k(1) returns ["apple"]. Example Input record u1 apple record u1 apple record u2 apple top_k 1 Output apple