Iterate over corpus directory and check for corresponding results when benchmarking

[yaseminbridges]

Currently in the benchmarking process, the results present in the PhEval processed results directory are iterated and the corresponding phenopacket is found this way - this should be swapped as some tools fail on cases and so no output is produced. This allows fair analysis as then we can assess based on the tool's capability on a whole corpus instead on those that it could run out of that corpus