[Ifeffit] Path updating in Artemis

Bruce Ravel bravel at bnl.gov
Mon Dec 31 07:37:14 CST 2007

On Saturday 29 December 2007 00:05:45 Silvio Levy wrote:
> Now suppose I tweak my feff data (in the feff.inp tab of Artemis) and
> rerun feff.  If I DON'T import any paths at the end, as recommended by
> the dialog, the Reff, amp, etc. associated with each path in the "Data
> & Paths" panel remain unchanged, while the data for the paths in
> "Interpretation" panel get updated.  The two columns are now out of
> sync, and pressing the Fit button merely redoes the computation using
> the old paths.
> If I rerun feff and DO import paths at the end, I end up with
> duplicate paths; the new ones have the the right Reff etc., but they
> use the default variables rather than the ones that I laboriously
> associated with the earlier paths.
> How can I get paths that have the variable scheme I created, but new
> data for each feff run?  The only way I seem to be able to do this is
> by closing the project and reopening it.  But I could swear that
> earlier this was not the case.
> What am I missing?

Hi Silvio,

You're not missing a thing.  You have discovered one of the serious
shortcomings of Artemis.  Changing the structure in the feff.inp file
and rerunning feff is high likely to change the list of paths in some
way that artemis is not currently clever enough to track.

Quitting and restarting is one work-around.  Another would be to
re-read the feff.inp file, thus starting another Feff entry in the
paths list.  Modify the feff.inp file, run feff, import paths,
parametrize them.  Then, exclude *all* of the paths from the previous
Feff calculation.  

I confess that this is also a clunky work-around.  But it has an
advantage over stopping and restarting in that it maintains a single
fit history inside the project file.

For what it's worth, the problem you are asking about is something
that my Demeter project will be able to handle gracefully.  In fact, I
have spent much of the last week working on Demeter's Feff interface.
It is coming along nicely, but I've still got a long path ahead.


 Bruce Ravel  ------------------------------------ bravel at bnl.gov

 National Institute of Standards and Technology
 Synchrotron Methods Group at NSLS --- Beamlines U7A, X24A, X23A2
 Building 535A
 Upton NY, 11973

 My homepage:    http://xafs.org/BruceRavel
 EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/

More information about the Ifeffit mailing list