@extract_meica_ortvec        - project good MEICA components out of bad ones

The MEICA process, via tedana.py, creates a set of components:

     accepted        : components it things are good BOLD
     ignored         : components it decides not to bother with
     midk_rejected   : components it "borderline" rejects
     rejected        : components it more strongly rejects

Together, this full matrix is fit to the data, and the fit of the
rejected components is subtracted from the data.  But the rejected
components are correlated with accepted ones.

To more conservatively keep the entirety of the accepted components,
projection components are created here by projecting the good ones
out of the bad ones, and taking the result as more strictly bad ones,
which can be projected later.

This script (currently) relies on being run from a tedana.py output
directory, probably of name TED.XXX.

sample commands:

   @extract_meica_ortvec -prefix run_5_meica_orts.1D

   @extract_meica_ortvec -meica_dir tedana_r01/TED.r01   \
         -work_dir tedana_r01/work.orts                  \
         -prefix tedana_r01/meica_orts.1D


   -prefix         PREFIX    : name for output 1D ortvec file
   -meica_dir      MDIR      : directory for meica files
   -reject_ignored VAL       : VAL=0/1, do we reject ignored components
                               (default = 0, keep, do not reject)
                               (should probably never reject)
   -reject_midk    VAL       : VAL=0/1, do we reject midk components
                               (default = 1, reject)
                               (should probably default to keeping)
   -work_dir       WDIR      : sub-directory for work
   -verb           VLEVEL    : set verbosity level

More options will be added, but this is enough to get used by
afni_proc.py for now.


Author: R Reynolds  May, 2018