Design a Stream Top-K Class (Online Top-K Elements/Frequencies)

Problem: Design a Top-K Class for a Data Stream You are given an unbounded stream of incoming elements. Design and implement a class that can report the current Top K elements at any time. Implement the following interface (pseudocode): class TopKStream: TopKStream(int k) void add(int x) List<int> topk() Requirements TopKStream(k): initialize the structure with k. add(x): ingest an element x (duplicates allowed). topk(): return the k most frequent elements seen so far. Rules If fewer than k distinct elements exist, return all distinct elements. If frequencies tie, break ties by smaller element value first (unless the interviewer specifies a different rule). add() can be called very frequently; topk() can be called at arbitrary times. Constraints (to guide complexity) Up to 10^6 calls to add() Values are 32-bit signed integers Discuss scalable time/space complexity for both operations. Example For k = 2: Operations: add(1) add(1) add(2) add(3) add(2) topk() Frequencies: 1->2, 2->2, 3->1 Return: [1, 2] (frequency desc; value asc for ties). Example Input k=2 ops=add 1, add 1, add 2, add 3, add 2, topk Output [1, 2]