Active Communication Top Users

Process a continuous stream of user pairs such as `A-B` and `B-C`; each pair creates an active communication edge, and the system should output users ordered by current communication count. Follow-ups add self-communication handling, duplicate communication dedupe, and top-K users by count. Requirements Input arrives as a stream of user pairs, for example A-B, B-C. After a pair is accepted, the two users form an active communication. Maintain each user's active communication count. Output users ordered by descending communication count. Clarify tie-breaking before coding; the prompt emphasizes sorted output but does not fix a canonical tie rule. Follow-ups: Handle self-communication such as A-A. Deduplicate an already-existing communication pair. Return only the top K users, still ordered by count. Notes A Java solution used an ordered map / tree structure to keep counts sortable. The interviewer asked many clarification questions and expected dry runs. The first version did not require self-communication handling; it was added as a coded follow-up. Preparation Implement the pair-normalization layer separately from count maintenance. Practice maintaining user -> count, edge set, and count -> ordered users views without letting updates drift. Dry-run duplicate edges and self-edges before discussing top-K optimization.