Optimize command is really slow

[osma]

There is something wrong with the optimize command. It used to be fast, but now it can take several minutes just to produce the table of results. I suspect this has to do with converting analysis results from lists to vectors (or vice versa) when evaluating different limit and threshold settings.

[osma]

It also seems to take a lot of memory - 4GB even for Maui projects where there should be no need to load models etc.

[osma]

Main issue seems to be that running all the evaluation measures just takes a lot of processing. The easiest fix would be to disable the nonessential eval measures (e.g. Prec@1, LRAP, true positives...) when running in optimize mode, as they are not really needed for the optimize command.

[osma]

High memory usage is still a problem with the optimize command. The intermediate AnalysisResults are not being freed, so memory usage grows during execution and can reach several GB after all parameter combinations have been tested.