# Summarize Transcript Events in a Sliding Window Implement `transcript_window_summaries(events, window_size)`. `events` is an arrival-ordered list of `(user_id, transcript)` pairs. After ingesting each event, return a summary for the most recent `window_size` events, or all events seen so far if fewer have arrived. Each summary is `(maximum_transcript_length, unique_user_count)`. Transcript length is Python's `len(transcript)`. If the input list is empty, return an empty list. Duplicate users and duplicate transcript strings are allowed. ## Constraints - `1 <= window_size <= 200,000` - `0 <= len(events) <= 200,000` - `user_id` and `transcript` are strings. - The total number of transcript characters is at most 2,000,000. - The output must contain one summary per input event. ## Example With a window size of three, the fourth summary covers events two through four. Its first component is the largest transcript length in those three events, while its second component counts distinct user IDs in that same window. ## Candidate clarifications Confirm that the window is based on event count rather than elapsed time, that the newest event is included, and that a user's repeated events count once toward the distinct-user total.