Design top-K tracker with linked lists

Design a data structure that tracks the top-K most frequent keys in a stream. Support operations increment(key), decrement(key), and topK(k) -> list of k keys with highest frequencies. Achieve O( 1) amortized updates by combining a hash map with a doubly linked list of frequency buckets. Specify tie-breaking, deletion behavior when counts drop to zero, memory usage, and complexity.