Duplicate Detection I / II / III

A three-part duplicate-detection series often restyled as watch-history episode IDs or show names. The third part may require bucketization or a longest-unique-subarray variant. Requirements Part 1: given a list of IDs, return whether any duplicate exists. Part 2: given K, return whether the same ID appears at indices i and j with |i - j| <= K. Part 3A: given threshold T, return whether two IDs in the current window have value distance <= T. Part 3B: show-name variant asks for the longest contiguous subarray with all unique names. A phone-screen string-array trio often runs as: (1) longest contiguous run of identical strings, (2) longest contiguous subarray with no repeated string, then (3) the disjoint-character pair count (its own card). The interviewer may ask for strings instead of integers; explain when bitmasking works and when it does not. Examples has_duplicate([55, 66, 77, 88, 99]) == False has_duplicate([55, 66, 77, 88, 66]) == True near_duplicate([1, 2, 3, 1, 2, 3], K=3) == True same_series([1, 5, 100], T=4) == True Notes Part 1 is a hash set. Part 2 is either a last-seen index map or a sliding window set of size K. Part 3 with |nums[i] - nums[j]| <= T is the canonical value-window-duplicate problem. Bucket width T + 1 gives O(n) expected time: compare the current bucket and adjacent buckets, then evict indices older than K if the K-window version is active. If the variant removes the index-window constraint and only asks for any pair within value threshold, sorting also works in O(n log n). Longest unique subarray uses a sliding window map from item to last index. Bitmasking only applies when each string can be represented as a fixed alphabet set, such as lowercase letters. It is not a general solution for arbitrary show names. Preparation Implement the three duplicate-detection versions back-to-back. Be able to derive bucket ID for negative integers correctly: use floor division semantics or normalize values. Prepare tests for K = 0, T = 0, duplicates exactly at distance K, and negative IDs.