Your question is Unique Users in Sliding Buffer. Start with the requirements on the right.
Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.
Interface.ai's conversational banking assistant emits an event stream as (userId, time) pairs. Given a buffer size k, determine the unique users represented in the buffer after each event arrives, where the buffer contains the most recent k events.
Implement unique_users_in_buffers(events, k). Return one result for every input event. Before the buffer reaches size k, use all events seen so far. Each result must contain unique user IDs in their first appearance order within the current buffer.
events, a list of pairs [userId, time], and positive integer k.userId is a string, and time is an integer. Events are provided in nondecreasing timestamp order, but timestamps may repeat.i contains the distinct user IDs in the buffer after processing events[i], ordered by first appearance in that buffer.def unique_users_in_buffers(events, k):