Tests fail if results of a prior test are still around
make check
seems to always give an error if run after make check-keep
.
frame['CL{}'.format(stat_type)] = [
> likelihood_point.combined_exclusion_dict[stat_type] for likelihood_point in self.points]
E KeyError: 'DATABG'
I think this is because of confusion between the outputs of the different tests. Running one more time sorts it out (passes test).
Could probably be fixed by just making the check delete any old results before starting, but maybe there is a better way.