I start by defining a helper function digit_sum that calculates the sum of digits for a given number. Then, I create a dictionary sum_map to store the largest number encountered for each digit sum. I then iterate through the list nums, compute the digit sum for each number, and check if that digit sum already exists in sum_map. If it does, I update max_sum with the sum of the current number and the stored maximum number for that digit sum. I also update sum_map to ensure it always holds the largest value for each digit sum.
Finally, I return max_sum, which holds the largest sum found. If no valid pair is found, I return -1.
