a few small bugfixes for crystal refinement: cartesianMD now works; optionally specify minimum weight on xtal data