e2evalrefine

usage: prog [options] [refine_xx]
	This program performs various assessments of e2refine_easy (or similar) runs, and operates in one of several possible modes.
	A refine_xx folder name must always be provided, except for resolution_all or resolution_vsref, which operate on all 
	refine_xx folders:

	--timing
		will report how long each refinement took to complete as well as individual tasks within the refinement
		
	--timingbypath
		will report total timing information for each refine_xx folder, along with useful refinement parameters
		
	--resolution
		Computes per-iteration FSCs for a single refine_xx folder
		
	--resolution_all
		Computes FSC curves for the final iteration of every refine_xx folder
		
	--resolution_vsref
		Computes FSC curve for the final iteration of each refine_xx folder vs a provided reference volume. 
		
	--evalptclqual
		This provides the ability to assess the agreement between individual particles used for a refinement and the 
		final refinement itself in several critical ways. Using this option will automatically generate two new sets/ 
		containing "good" and "bad" particles from the population used to run the specified refinement. The assessment
		is made using a set of heuristics. A multicolumn text file is also produced containing detailed results of
		the per-particle comparison (ptclfsc_XX.txt). The first line of this text file contains details about the
		meaning of each column:
		0 - 100-30 it1 (integrated FSC from 100-30 A resolution for second to last iteration)
		1 - 30-18 it1
		2 - 18-10 it1
		3 - 10-4 it1
		4 - 100-30 it2 (integrated FSC from 100-30 A resolution for last iteration)
		5 - 30-18 it2
		6 - 18-10 it2
		7 - 10-4 it2
		8 - it12rmsd (RMSD between it1 and it2 integrated FSCs)
		9-11 - alt1, az1, cls1  (orientation parameters for second to last iteration)
		12-14 - alt2, az2, cls2 ( " for last iteration)
		15 - defocus
		
		e2display.py --plot ptclfsc*txt		(then adjust columns the plot displays, eg- 4 vs 5, 4 vs 6, 0 vs 4, 1 vs 5)

	--evalclassqual
		similar analysis to evalptclqual but for class-averages vs projections
		
	--anisotropy
		Assesses the amount and direction of any magnification anisotropy present in a raw data set by considering
		particles in a range of different orientations. Works best with large particles. Specify a class-average number
		for a class containing many particles (within a particular refine_xx folder). It is a good idea to compare results
		among classes in different highly occupied orientations.

	--extractorientptcl <output name>
		provides a way to extract particles with the alignment information for each particle stored in its header. Ostensibly such a data set could
		be run directly through e2make3d.py, but its primary use is with e2gmm.py and similar programs.
	

Option Type Description
--version None show program's version number and exit
--timing None Report on the time required for each step of each refinement run
--timingbypath None Report on the CPU time required in each refine_xx folder
--resolution None generates a resolution and convergence plot for a single refinement run.
--resolution_all None generates resolution plot with the last iteration of all refine_xx directories
--resolution_vsref str Computes the FSC between the last iteration of each refine_xx directory and a specified reference map. Map must be aligned, but will be rescaled if necessary.
--evalptclqual None Evaluates the particle-map agreement using the refine_xx folder name. This may be used to identify bad particles.
--evalclassqual None Evaluates the class-average-projection agreement using the refine_xx folder name.
--extractorientptcl str Extracts the particles from a refinement with per-particle orientation information. If HDF output, will store as xform_align3d in header.
--anisotropy int Specify a class-number (more particles better). Will use that class to evaluate magnification anisotropy in the data.
--evalclassdetail None If specified with evalclassqual, will generate individual FRC curves for each class average in the even subset
--includeprojs None If specified with --evalptclqual, projections will be written to disk for easy comparison.
--iter int If a refine_XX folder is being used, this selects a particular refinement iteration. Otherwise the last complete iteration is used.
--mask str Mask to be used to focus --evalptclqual and other options. May be useful for separating heterogeneous data.
--sym str Symmetry to be used in searching adjacent unit cells, default from refine_xx parms
--threads int Number of threads to run in parallel on a single computer when multi-computer parallelism isn't useful
--ppid int Set the PID of the parent process, used for cross platform PPID
--verbose, -v int verbose level [0-9], higher number means higher level of verboseness

For more information go to emanwiki/EMAN2/Programs/e2evalrefine.