You have a list of tasks, tasks[i] hours each, and you work in sessions of at most session_time hours. A task must be finished inside a single session (no splitting), you can do tasks in any order, and a session may hold several tasks as long as their total fits.
Write min_sessions(tasks: list[int], session_time: int) -> int returning the fewest sessions needed to finish every task.
Constraints: 1 <= len(tasks) <= 14, 1 <= tasks[i] <= session_time <= 15.
min_sessions([2, 3, 4], 5) # 2: {2, 3} and {4}
min_sessions([4, 4, 3, 3, 2], 8) # 2: {4, 4} and {3, 3, 2}
min_sessions([6, 6, 6], 6) # 3
Greedy packing (largest first into the first session with room) is not always optimal. ⭐ Bonus: trying every assignment of tasks to sessions passes small cases; a speed test with 14 tasks earns a star at about O(2^n · n).
Show hint
With only 14 tasks, the set of tasks already finished fits in a bitmask. For each such set, it is enough to remember the best way to have finished it: the fewest sessions used, and among those, the least time used in the current session.