first pass at setting cut parameters at runtime

[mjmottram]

I decided that I wanted to add some extra cuts (e.g. a simple evIndex==0) when extracting from a file, this isn't possible at the moment without first parsing through an ntuple to create a new one. This PR allows for users to specify their own cuts and also removes lots of repeated code from dsextract. That said, it sounds like people don't really like dsextract anyway so maybe it should be scrapped and people write their own scripts to run in a manner of their choosing.

I also updated the num_decays for ntuples to remove duplicated MC events.

[arushanova]

As expected, I agree with this statement "That said, it sounds like people don't really like dsextract anyway so maybe it should be scrapped and people write their own scripts to run in a manner of their choosing." For my studies I can't use existing extractors and it is much easier for me to simply create hdf5 file in the end of my code. But, I can write extractors, when I have time, if needed.

When you say "people", I'm not sure whom you are talking about. It seems, that Ashley and James are both using this method and, I suppose, find it useful.

[mjmottram]

Ah, but that's how this might help. You can now pass arbitrary lists of cuts (e.g. on classifier values) to the dsextract module classes. I know Ashley wasn't sure whether dsextract should remain in the codebase. Personally I think it's useful purely to ensure consistency when keeping track of total events before cuts are applied.

[ashleyrback]

@mjmottram you'll need to merge in from snoplusuk/master to resolve the conflicts.

[ashleyrback]

@snoplusuk/core any volunteers to review this?

[mjmottram]

Conflicts fixed.