rs.rfree should be able to extend resolution range

[kmdalton]

It would be really helpful if rs.rfree could take as input a refinement output from PHENIX and output a new r-free file which is consistent with it but has an expanded resolution range. See this question on the rs-station forum as a starting point.

[DorisMai]

Motivation: help benchmark merging --> refinement. Abismal typically yield sharp map (e.g. rsICH apo), and we would like to make Rfree comparable to other methods for benchmarking purposes while pushing to higher resolution.

Method: I am using Mark Wilson's r-free-flag file as an example to start the implementation. A minimal example script is posted to the question on the rs-station forum.

Issue: Currently Mark generated the original flags using phenix.fmodel, which caps the maximum number of rfree-flags and disrepects his input of 5% R-free. Resampling needs to be implemented.