'''
Description: 
Autor: Au3C2
Date: 2020-12-15 12:49:52
LastEditors: Au3C2
LastEditTime: 2020-12-16 15:55:48
'''
class Solution:
    def longestSubarray(self, nums: List[int], limit: int) -> int:
        n = len(nums)
        left, right, max_len = 0, 0, 0
        # cur_min = cur_max = nums[0]
        cur_list = [nums[0]]
        max_len = cur_len = 1
        while right < n-1:
            right += 1
            bisect.insort_left(cur_list, nums[right])
            cur_len += 1

            while cur_list[-1] - cur_list[0] > limit:
                cur_list.pop(bisect.bisect_left(cur_list, nums[left]))
                left += 1
                cur_len -= 1
            max_len = max(max_len,cur_len)
        return max_len
# 滑窗，这个程序性能比较差,其实我维护有序表的目的是获取队列最大最小值，那么只需要维护一个最大堆，一个最小堆即可
# https://leetcode-cn.com/problems/longest-continuous-subarray-with-absolute-diff-less-than-or-equal-to-limit/