Re: [Ifeffit] How to use Artemis with feff85L

[Bruce Ravel]

Jose,

I am unclear what you are sad about.  The version of Feff has little to 
do with the quality of the analysis.  See our recent paper:


  https://doi.org/10.1107/S1600577517011651

It is true that feff8l would simplify and streamline the code, making it 
easier to write and maintain.  That's a bit incentive /for/ /me/ to do 
that work.


But I think you are overstating the impact on the analysis you and 
others do with Artemis.


B


On 01/12/2018 07:15 AM, José Carlos Conesa wrote:

OK Bruce, I understand. (Sigh)

Good luck in your actual job.

JC


El 12/01/2018 a las 12:35, Bruce Ravel escribió:


Yes.  But I am pretty busy these with my new beamline.

"Beamline scientist" is my actual job.  "Software guy" is a side 
project.  So the software work gets done when it gets done.


B



On 01/12/2018 05:25 AM, José Carlos Conesa wrote:

Dear Bruce,

Do you foresee that (in a not far future) Artemis may be adapted so 
that it can use the results of feff8.5L?


All the best,


El 12/09/2017 a las 19:11, Bruce Ravel escribió:


Feff8L is not properly integrated into Artemis yet.  Sorry.  I have 
a new beamline these days.  It's hard to find time to work on the 
software.


This is on my to do list, but it's going to take a while.

B


On 09/12/2017 12:54 PM, José Carlos Conesa wrote:

Hi,

I should like to use feff85L within Artemis (from Demeter 0.9.25), 
just to check in my system whether selfconsistency has a role in 
unifying E0 values (with feff 6, allowing different E0 for each 
scatterer leads to E0 differences >10 eV, which seems too much to 
me). I find that after loading the EXAFS data (from Athena) and a 
model (a .cif file containing Cu, Ce and O atoms), if I click Run 
Atoms and then Run Feff with feff6 as executable I get the usual 
set of paths that allow me doing a fit of the data, but if I try to 
use the feff85L.exe that comes with the Demeter software (by 
changing the feff executable in the File/Edit preferences menu), 
after clicking the Run Feff button (without including the SCF card) 
a bit of computation happens and after finishing it the path list 
is empty.


¿Any idea of the possible reason why I get no path list?

Thanks in advance














--
 Bruce Ravel   bra...@bnl.gov

 National Institute of Standards and Technology
 Synchrotron Science Group at NSLS-II
 Building 743, Room 114
 Upton NY, 11973

 Homepage:http://bruceravel.github.io/home/
 Beamline:https://www.bnl.gov/ps/beamlines/beamline.php?r=6-BM
 Software:https://github.com/bruceravel
 Demeter: http://bruceravel.github.io/demeter/
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[Bruce Ravel]

Yes.  But I am pretty busy these with my new beamline.

"Beamline scientist" is my actual job.  "Software guy" is a side 
project.  So the software work gets done when it gets done.


B



On 01/12/2018 05:25 AM, José Carlos Conesa wrote:

Dear Bruce,

Do you foresee that (in a not far future) Artemis may be adapted so that 
it can use the results of feff8.5L?


All the best,


El 12/09/2017 a las 19:11, Bruce Ravel escribió:


Feff8L is not properly integrated into Artemis yet.  Sorry.  I have a 
new beamline these days.  It's hard to find time to work on the software.


This is on my to do list, but it's going to take a while.

B


On 09/12/2017 12:54 PM, José Carlos Conesa wrote:

Hi,

I should like to use feff85L within Artemis (from Demeter 0.9.25), 
just to check in my system whether selfconsistency has a role in 
unifying E0 values (with feff 6, allowing different E0 for each 
scatterer leads to E0 differences >10 eV, which seems too much to 
me). I find that after loading the EXAFS data (from Athena) and a 
model (a .cif file containing Cu, Ce and O atoms), if I click Run 
Atoms and then Run Feff with feff6 as executable I get the usual set 
of paths that allow me doing a fit of the data, but if I try to use 
the feff85L.exe that comes with the Demeter software (by changing the 
feff executable in the File/Edit preferences menu), after clicking 
the Run Feff button (without including the SCF card) a bit of 
computation happens and after finishing it the path list is empty.


¿Any idea of the possible reason why I get no path list?

Thanks in advance









--
 Bruce Ravel   bra...@bnl.gov

 National Institute of Standards and Technology
 Synchrotron Science Group at NSLS-II
 Building 743, Room 114
 Upton NY, 11973

 Homepage:http://bruceravel.github.io/home/
 Beamline:https://www.bnl.gov/ps/beamlines/beamline.php?r=6-BM
 Software:https://github.com/bruceravel
 Demeter: http://bruceravel.github.io/demeter/
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[Matt Newville]

Hi Jose,

On Thu, Sep 14, 2017 at 3:54 PM, José C. Conesa 
wrote:

> Hi Matt,
>
> Some comments (hopefully final):
>
> 1-A Windows version of feff85L (called FEFF8-Lite) is freely available,
> with some other codes, at
>
> http://leonardo.phys.washington.edu/feff/Bin-Source-
> feff8Lite/Bin-Source.html
>

OK, thanks for the link.  I was not aware that binaries were being
distributed there.  I would guess that they work, though I have not tried
these myself.

> One can also find the feff85L.f source code in other places.
>
> 2-Being able to generate automatically with Artemis an input file for
> feff8 is of course convenient, even if one would have to enter by hand
> some keywords (e.g. SCF) to get the full functionality. I did this, added
> into Artemis the so modified feff.inp file and tried to run feff8 with that
> file. I found a message saying "You are running Feff8 with self-consistent
> potentials.It WILL be time consuming and all interaction with Artemis will
> be blocked until the Feff calculation is done. Currently Artemis does not
> provide real-time feedback, so you will have to be very patient." I would
> say that gives the impression that Feff8 can indeed be used within Artemis.
> Indeed you say "Artemis has been able to use various versions of Feff8 for
> a very long time". Is it still possible?
>

Yes. Or, it should be.  Perhaps this Feff85L.exe did run for you, but just
not do everything you wanted?


> 3-I do have (properly licensed) feff 9, and indeed have its documentation
> independently of Demeter. On the other hand, unless one has an illegal copy
> of feff 9 I do not see the usefulness of having that documentation without
> having a license.
>

I don't know of anyone with an illegal copy of Feff9.  I've only seen it
run in classes, and do not have a copy myself.

The fact that Artemis points to the on-line Feff documentation is very
convenient.  That this is for Feff9 and not Feff8 or other versions is not
really in our control.   Artemis comes with the latest version of Feff that
it can.


> 4-Unfortunately I am not competent enough as programmer. I would not be
> able to contribute to make Artemis able to use feff8 calculations. If you
> have not time to devote to it, I fully understand it. Already what has been
> done in Demeter is a great work.
>

A great many people say the same thing.  If that "many" becomes "everyone"
you will not have to worry which version of Feff can be run by Artemis.

--Matt
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[José C. Conesa]

Hi Matt,

Some comments (hopefully final):

1-A Windows version of feff85L (called FEFF8-Lite) is freely available, 
with some other codes, at


http://leonardo.phys.washington.edu/feff/Bin-Source-feff8Lite/Bin-Source.html

One can also find the feff85L.f source code in other places.

2-Being able to generate automatically with Artemis an input file for 
feff8 is of course convenient, even if one would have to enterby hand 
some keywords (e.g. SCF) to get the full functionality. I did this, 
added into Artemis the so modified feff.inp file and tried to run feff8 
with that file. I found a message saying "You are running Feff8 with 
self-consistent potentials.It WILL be time consuming and all interaction 
with Artemis will be blocked until the Feff calculation is done. 
Currently Artemis does not provide real-time feedback, so you will have 
to be very patient." I would say that gives the impression that Feff8 
can indeed be used within Artemis. Indeed you say "Artemis has been able 
to use various versions of Feff8 for a very long time". Is it still 
possible?


3-I do have (properly licensed) feff 9, and indeed have its 
documentation independently of Demeter. On the other hand, unless one 
has an illegal copy of feff 9 I do not see the usefulness of having that 
documentation without having a license.


4-Unfortunately I am not competent enough as programmer. I would not be 
able to contribute to make Artemis able to use feff8 calculations. If 
you have not time to devote to it, I fully understand it. Already what 
has been done in Demeter is a great work.


All the best,

José Carlos

El 14/09/2017 a las 15:08, Matt Newville escribió:


Hi Jose,


On Thu, Sep 14, 2017 at 2:32 AM, José C. Conesa > wrote:


Hi,

Well, the information on this was somewhat confusing. Thus:

1-In the main Artemis window, after importing a .cif file to do a
Feff calculation, the Atoms tab includes in the Export button the
possibility of generating a .inp file with feff8 format; this is
mentioned as well in the corresponding part of the Help
documentation. Since no warning appears at that point about the
impossibility of using that .inp file in Demeter, one would
believe that such thing is indeed possible.


It's long been the case that Artemis / Atoms can generate an input 
file for Feff8, even though Feff8 is not distrubuted with Demeter.  
There have been several discussions on this list over the past several 
years about how to use a version of Feff other than feff6l included in 
the distribution.


For this reason, although indeed no feff8 executable comes with
Demeter (my excuses for having assumed otherwise), I managed to
obtain a feff85L.exe file and put it some time ago in the same
Demeter directory as feff6.exe.


If you have a working Feff8 executable, Artemis should be able to use 
it.  But, I am not aware of a distribution of feff85L.exe for 
Windows.  Bruce and I have been working on Feff85L for EXAFS, but 
there is not a single Windows executable in that work.  Maybe someone 
else has produced such a thing.


Then I forgot that feff85L.exe did not come with Demeter. And when
choosing the feff executable with the Edit preferences menu item,
I thought that the mention there that feff8 was not yet supported
was outdated.

Artemis has been able to use various versions of Feff8 for a very long 
time.  It does not come with a version of Feff8, at least not yet.


2-In that same Help documentation, section 4.2 The Feff tab gives
a link to the documentation of FEFF9. This gives the impression
that the documentation is not very consistent.


Feff9 is the latest version of Feff.  Isn't it awesome that the Feff 
tab gives you a link to the Feff documention?


All this said, one can now only wish that a Demeter version
allowing the use of feff8 is released some day.


Well, it's not exactly true that one can now only wish that Demeter 
allow use of Feff8.  One can also work on it. These are open source 
projects, desparate for help, including things as simple as 
encouragement, testing, and giving complete and actionable error 
reports. If you expect someone else to give you a free, pre-packaged 
solution that works for you without effort, you might be waiting a 
very long time.


--Matt



--
José C. Conesa
Instituto de Catálisis y Petroleoquímica, CSIC
Marie Curie 2, Cantoblanco
28049 Madrid, Spain
Tel. (+34)915854766

___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[Matt Newville]

Hi Jose,


On Thu, Sep 14, 2017 at 2:32 AM, José C. Conesa 
wrote:

> Hi,
>
> Well, the information on this was somewhat confusing. Thus:
>
> 1-In the main Artemis window, after importing a .cif file to do a Feff
> calculation, the Atoms tab includes in the Export button the possibility of
> generating a .inp file with feff8 format; this is mentioned as well in the
> corresponding part of the Help documentation. Since no warning appears at
> that point about the impossibility of using that .inp file in Demeter, one
> would believe that such thing is indeed possible.
>

It's long been the case that Artemis / Atoms can generate an input file for
Feff8, even though Feff8 is not distrubuted with Demeter.  There have been
several discussions on this list over the past several years about how to
use a version of Feff other than feff6l included in the distribution.

> For this reason, although indeed no feff8 executable comes with Demeter
> (my excuses for having assumed otherwise), I managed to obtain a
> feff85L.exe file and put it some time ago in the same Demeter directory as
> feff6.exe.
>

If you have a working Feff8 executable, Artemis should be able to use it.
But, I am not aware of a distribution of feff85L.exe for Windows.  Bruce
and I have been working on Feff85L for EXAFS, but there is not a single
Windows executable in that work.  Maybe someone else has produced such a
thing.

> Then I forgot that feff85L.exe did not come with Demeter. And when
> choosing the feff executable with the Edit preferences menu item, I thought
> that the mention there that feff8 was not yet supported was outdated.
>
Artemis has been able to use various versions of Feff8 for a very long
time.  It does not come with a version of Feff8, at least not yet.

2-In that same Help documentation, section 4.2 The Feff tab gives a link to
> the documentation of FEFF9. This gives the impression that the
> documentation is not very consistent.
>

Feff9 is the latest version of Feff.  Isn't it awesome that the Feff tab
gives you a link to the Feff documention?


> All this said, one can now only wish that a Demeter version allowing the
> use of feff8 is released some day.
>

Well, it's not exactly true that one can now only wish that Demeter allow
use of Feff8.  One can also work on it. These are open source projects,
desparate for help, including things as simple as encouragement, testing,
and giving complete and actionable error reports. If you expect someone
else to give you a free, pre-packaged solution that works for you without
effort, you might be waiting a very long time.

--Matt
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[José C. Conesa]

Hi,

Well, the information on this was somewhat confusing. Thus:

1-In the main Artemis window, after importing a .cif file to do a Feff 
calculation, the Atoms tab includes in the Export button the possibility 
of generating a .inp file with feff8 format; this is mentioned as well 
in the corresponding part of the Help documentation. Since no warning 
appears at that point about the impossibility of using that .inp file in 
Demeter, one would believe that such thing is indeed possible. For this 
reason, although indeed no feff8 executable comes with Demeter (my 
excuses for having assumed otherwise), I managed to obtain a feff85L.exe 
file and put it some time ago in the same Demeter directory as 
feff6.exe. Then I forgot that feff85L.exe did not come with Demeter. And 
when choosing the feff executable with the Edit preferences menu item, I 
thought that the mention there that feff8 was not yet supported was 
outdated.


2-In that same Help documentation, section 4.2 The Feff tab gives a link 
to the documentation of FEFF9. This gives the impression that the 
documentation is not very consistent.


All this said, one can now only wish that a Demeter version allowing the 
use of feff8 is released some day.


Thanks anyway for the clarifications.

José Carlos


El 14/09/2017 a las 5:08, Bruce Ravel escribió:

On 09/13/2017 09:35 PM, Matt Newville wrote:
Then again, maybe Demeter does come with a working feff8l, and I just 
wasn't aware of it.


Not yet. :(
B



--
José C. Conesa
Instituto de Catálisis y Petroleoquímica, CSIC
Marie Curie 2, Cantoblanco
28049 Madrid, Spain
Tel. (+34)915854766

___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[Bruce Ravel]

On 09/13/2017 09:35 PM, Matt Newville wrote:
Then again, maybe Demeter does come with a working feff8l, and I just 
wasn't aware of it.


Not yet. :(
B

--
 Bruce Ravel   bra...@bnl.gov

 National Institute of Standards and Technology
 Synchrotron Science Group at NSLS-II
 Building 743, Room 114
 Upton NY, 11973

 Homepage:http://bruceravel.github.io/home/
 Software:https://github.com/bruceravel
 Demeter: http://bruceravel.github.io/demeter/
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit

Re: [Ifeffit] How to use Artemis with feff85L

[Matt Newville]

Hi Jose,


On Tue, Sep 12, 2017 at 11:54 AM, José Carlos Conesa 
wrote:

> Hi,
>
> I should like to use feff85L within Artemis (from Demeter 0.9.25), just to
> check in my system whether selfconsistency has a role in unifying E0 values
> (with feff 6, allowing different E0 for each scatterer leads to E0
> differences >10 eV, which seems too much to me). I find that after loading
> the EXAFS data (from Athena) and a model (a .cif file containing Cu, Ce and
> O atoms), if I click Run Atoms and then Run Feff with feff6 as executable I
> get the usual set of paths that allow me doing a fit of the data, but if I
> try to use the feff85L.exe that comes with the Demeter software (by
> changing the feff executable in the File/Edit preferences menu), after
> clicking the Run Feff button (without including the SCF card) a bit of
> computation happens and after finishing it the path list is empty.
>
>
Does demeter come with a Feff85L executable?  I think it may not.

In fact, I just today added a standalone "feff8l" program to Larch and on a
development branch of the feff85exafs code (wherein I am tring to improve
the Python interface for "calculated an XAFS path on the fly").   But none
of that is released yet.   And it may have bugs, especially on Windows.

So, it's great that you're interested in trying to run Feff8L. But I
suspect that you may have to build it yourself, especially if you're hoping
to run it on Windows.

Then again, maybe Demeter does come with a working feff8l, and I just
wasn't aware of it.

--Matt
___
Ifeffit mailing list
Ifeffit@millenia.cars.aps.anl.gov
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Unsubscribe: http://millenia.cars.aps.anl.gov/mailman/options/ifeffit