fix calculation of attached runs in print_nl() function

[nldoc]

print_nl() provides an estimation of expected runs and reports the actual number of runs of the attached simoutput file. However this calculation is wrong. It does not consider that multiple rows can exist for the same siminputrow/seed combination. Thus, things like this can happen: number of siminputrows: 598 number of random seeds: 5 estimated number of runs: 2990 simoutput results attached: ✓ number of runs calculated: 1731645

This can be easily fixed by calculating the number of attached runs by using unique values vectors of the siminputrow and random-seed columns.