Build an in-memory event aggregator for chat events. Each event has `user_id`, `chat_id`, `timestamp`, and `event_type`. First, support querying the number of events in the last five minutes. Then support querying, for each user, how many chats are active in the last five minutes. A chat is active for a user if that user has a `react` event for the chat in the last five minutes and there is no later `end_chat` event for that same user and chat. Function signature: ```python class MessageEventAggregator: def add_event(self, user_id: str, chat_id: str, timestamp: int, event_type: str) -> None: pass def recent_event_count(self, now: int) -> int: pass def active_chat_count(self, user_id: str, now: int) -> int: pass ``` Constraints: - Timestamps are integer seconds. - Events may arrive in increasing timestamp order for the base version. - A follow-up may allow out-of-order events. - Memory should not grow without bound as old chats and events age out. - `event_type` is either `react` or `end_chat`. Examples: ```text add react u1 c1 at 100, react u1 c2 at 120, end_chat u1 c1 at 130; active_chat_count('u1', 150) returns 1. ``` ```text At now = 1000, events older than 700 are outside the five-minute window. ```