Practice Random Flip Matrix

Solution(m,n) initializes an all-zero matrix. flip() uniformly selects a currently zero cell, changes it to one, and returns [row,column]. reset() makes all cells zero again. Minimize random calls and avoid storing the entire potentially large matrix. Practice approach: Flatten indices and simulate sampling without replacement using a sparse swap map. Choose one remaining index, swap its mapped value with the last remaining slot, and shrink the available range. Reset clears the map and restores the count. This provides O(1) expected work per flip with storage proportional to flips.