- Ifeffit - millenia.cars.aps.anl.gov

(no subject)
by Bruce Ravel 24 Jun '02

24 Jun '02

Hi, I fixed several bugs in Artemis last week and today I posted a new source code tarball. The changes include: 1. fixed several bugs in displaying and using different FT window choices 2. fixed some bugs affecting deletion and rereading of feff calculation. 3. added the ability to turn on and off the "include this path in fit" button for the set of selected paths (i.e. all the paths highlighted in orange in the list on the right). I also reorganized the pull-down menus for including/excluding paths from the fit. 4. paths included in the fit are listed in black text, paths excluded from the fit are listed in a different color. also things not intended for plotting (such as the GSD listing) are printed in italics. this all gives some visual cues to help use the contents of the list effectively. 5. fixed (I hope!) a serious problem which manifested itself by having the data appear as a bizarre starburst-like plot in k-space. Several people had noticed this already. These are all pretty significant usability improvements. Special thanks this week go to Paul Fons for helping me track down several of these problems. As always, grab the latest from http://feff.phys.washington.edu/~ravel/software/exafs/ Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Code 6134, Building 3, Room 222 Naval Research Laboratory phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

artemis source code update
by Bruce Ravel 18 Jun '02

18 Jun '02

Hi, I fixed the following four problems with artemis reported to me last week: 1. a really bad bug triggered by using capital letters in parameter names 2. a strange end-of-line problem when artemis was run on unix using feffNNNN.dat files made on windows 3. the "View paths.dat" button will look for "paths.dat" or "path00.dat", i.e. it will find the paths file from feff6 or feff8 4. several related bugs in how the functional form of the FT windows were chosen and displayed on the screen I updated the source tarball and posted it on my web site. http://feff.phys.washington.edu/~ravel/software/exafs/ Artemis currently has the unfortunate, complimetary problems of lacking documentation while being a complex and unintuitive program. The best suggestion I can make for now for learning to use it is to check out the two project files in the examples/Cu/ directory. They are the ones with the .apj extensions. If you are a feffit user, you can import your `feffit.inp' files. If you have never used feffit before, I would recommend reading Chapters 2 and 4 of the `notes.pdf' file found on the web page for my "EXAFS Analysis Using FEFF and FEFFIT" course: http://leonardo.phys.washington.edu/~ravel/course/ Between that and examining the project files, hopefully Artemis will make some kind of sense. Thanks especially to Mark Jensen and Michael DeLeon for their email in the last few days. Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Naval Research Laboratory, Code 6134 phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

2 2

Mark Jensen's artemis bug report
by Bruce Ravel 14 Jun '02

14 Jun '02

Hi, Yesterday, Mark Jensen alerted me to two problems in the current version of artemis. The first is a situation wherein artemis ignores certain guess parameters. The second is a situation in which artemis fails to display the header information from the feffNNNN.dat file as in this screenshot: http://leonardo.phys.washington.edu/~ravel/software/exafs/images/artemis/ar… The first problem is due to the fact that artemis gets confused by capital letters in parameter names. That is, "deltar" is ok, but "deltaR" is not. This will be fixed in the next release. The work-around for this problem is to avoid capital letters in parameter names. The second problem was for the somewhat strange situation of running artemis on a unix machine and having made the feffNNNN.dat files on a windows machine. Because of a mistake that I made in the code that fills in the text box in the picture above, artemis got confused by the different end-of-line conventions on windows and unix. How embarassing! This will also be fixed soon. The work-around is to run feff and artemis on the same kind of computer. I want to thank Mark for the excellent bug report and for his help in troubleshooting. Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu U.S. Naval Research Laboratory, Code 6134 phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

XAS GUIs
by Webb, Samuel M. 14 Jun '02

14 Jun '02

Hi All... Just wanted to post this to the audience here. I've been working on a set of GUIs for XAS preprocessing and fitting and I think they are just about ready for testing by a broader audience. The first program is SamView and performs basic averaging of scans, energy calibration, deadtime correction and viewing of raw data. At this point, it is specialized to read only certain file formats, but the capability will be expanded in the future. Since my focus has been at SSRL recently, it will read binary and ASCII formats from SSRL, as well as from DND-CAT at the APS. In summary, it is essentially a graphical version of MAVE, MCALIB, and MDEAD for those who have used EXAFSPAK at SSRL. The second program (SamXAS) is a XAS fitting routine. It has a GUI interface to IFEFFIT for FEFF fitting of XAS, a linear combination fitting routine for XAS and XANES, and an interface for creating FEFF single scattering paths (if Feff8 is installed on your system). It also has a fitting routine to account for fluoresence self absorption with out prior knowledge of the sample composition. These are all in their infancy, but I'm looking for feedback on how to improve them, with out replicating the magnificnet work of other contributors. The documentation for both is essentially non-existent, but there are a few sample workspaces and data files included to get you started. Familiarity with IFEFFIT is definitely a plus. The programs can be downloaded from www.stanford.edu/~swebb. Thanks! Sam Webb SSRL samwebb(a)slac.stanford.edu

1 0

First attempts to use Artemis
by Mark Jensen 13 Jun '02

13 Jun '02

Dear Everyone: I'm not sure if this is a bug report or a dopey user report for Artemis, but I am having a problem with Artemis on two Red Hat 7.3 systems with Feff8. The data files load and are displayed in the PGPlot window. I can load and display FEFF paths in the windows too. (Nice touch Bruce. Ravel-ware is great!). However, :( , the path information is not displayed in the path text box like the last picture in Bruce's poster shows. That's not a killer problem in my mind. The real problem is that I can only fit 0 or 1 variable, no matter how many I set as guessed. It appears that Artemis/ifeffit is only varying the guessed variable associated with the first path variable on whichever tab (Basic parameters or Other parameters) is visible (i.e. S02, OR the third cummulant). It won't vary both S02 and the 3rd cummulant and it won't vary any guessed variables that appear in the other path parameters. It doesn;t appear to matter how many different paths I include. If there isn't a guessed variable in either the S02 or third cummulant path, none of the guessed variables is varied. The values of all guessed variables that are not varied always return from ifeffit as 0.0. Any thoughts Mark Jensen ------------------------------ Mark P. Jensen Chemistry Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439-4831 USA 630-252-3670 630-252-7501 FAX mjensen(a)anl.gov

1 0

rendering problem with Athena
by Carlo U. Segre 08 Jun '02

08 Jun '02

hello Bruce: I have noticed a rendering problem with Athena in Linux. I am attaching a PNG screenshot of the offending screen. As you can see, the lowest part of the "Background Removal" subwindow is showing only the top and it is impossible to read it. Since the window will not resize, it is not just a question of enlarging it. I use a Debian testing system and I would be happy to tell you what libraries are installed if you let me know which ones are relevant. Carlo -- Carlo U. Segre -- Professor of Physics Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 Carlo.Segre(a)iit.edu http://www.iit.edu/~segre

1 0

Ifeffit: spline with a standard
by Bruce Ravel 23 May '02

23 May '02

Hi, Earlier today Michael Green asked Matt a question of the variety "How do you do ______?" Athena (and artemis) may provide some help with regards to questions like that. If you look in the Edit menu, you will see a item that says "Display Ifeffit buffer". This shows a text box which displays all text strings that have been sent by athena to ifeffit. This is useful to folks who want to learn ifeffit's syntax. To use the example that was just posted to the mailing list, one could read some data and a feff chi.dat file. Then plot the data in mu(E) and the background without a standard. Then tell athena to use the chi.dat file as a standard and replot in energy. Because the standard was changed, the spline will be redone using the newly chosen standard. All of the ifeffit commands to do this will be written to the ifeffit buffer for your examination. Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu U.S. Naval Research Laboratory, Code 6134 phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

Ifeffit: spline with a standard
by Matt Newville 23 May '02

23 May '02

Hi Michael, This seems like a question that others might have, so I hope it's OK that I'm sending this reply to the ifeffit mailing list too. On Thu, 23 May 2002, Michael Green wrote: > I can't figure out how to use a standard for background > removal in G.I.FEFFIT or IFEFFIT. Is there a way to do > this? Table 4 of the Refman lists a chi_std data type, but > I didn't find any mention of it anywhere else. Yes. In ifeffit, it would go like this: #### # background-removal using known chi(k) data read_data(file= data.xmu, label= 'energy xmu', group= dat) read_data(file= standard.chi, label= 'k chi ', group= std) # without standard spline(energy= dat.energy, xmu= dat.xmu, kweight=1,rbkg=1) # with standard spline(energy= dat.energy, xmu= dat.xmu, kweight=1,rbkg=1, k_std=std.k, chi_std = std.chi) #### That is, you read in the spectra for the standard, and use it's "k" and "chi" arrays for "k_std" and "chi_std". Hope that helps, --Matt

1 0

divorce and OSX
by Bruce Ravel 23 May '02

23 May '02

Hi, There are two more issues that I wanted to comment on, but my last message was really long, so I am writing a new one. 1. Matt suggested an amicable divorce as a way of trying to avoid many of the installation problems that we have been having recently. I would suggest a separation rather than a divorce. Ifeffit and the apps might resolve their marital issues in the future and choose to get back together. But for now, for the sake of the kids, I think a separation is a good idea. I freely confess that many of the installation problems are either due to programming mistakes I have made or due to inherent problems in how perl modules get built. It seems imprudent that Matt and Ifeffit should suffer for those reasons. And, as Matt pointed out, my codes are evolving and releasing faster than Ifeffit itself. While it is certainly less convenient for the end user to have to install several packages than a single one, it is even less convenient if the whole kit 'n' kaboodle fails because of an error in a single app. A separation is thus a very good idea. The separation works well for unix users, where it is a simple matter to grab, say, the athena source tarball and install it on top of a functional ifeffit. It is a little less convenient for the windows users. I suppose a solution is that *I* build a new windows installer package based off of Matt most recent ifeffit release when I release new source code tarball. That's not a complete solution because Sam will certainly have a different release schedule than I will. I vote that Matt, Sam, and I put off dealing with that until dealing is unavoidable. 2. Could Paul, Francois, or Chuck summarize for us the status of ifeffit + X + perl + tk + my codes on OSX? (Although I chuckled when I read it, I cannot bring myself to actually use the word Ravelware ;-) Specifically I am interested in a clear statement of how difficult it is to get my codes up an running on a virgin system and how much time it takes. I am also interested in the prospects for providing a click-n-install package for OSX users. I just gave a poster at the NSLS users meeting on my software and showed one of Francois's pictures of Atoms running on OSX. I presume that it will spark some interest. Also, NRL's beamlines at NSLS are Mac shops (thanks to Johnny Kirkland, deacon of the church of the mac) and I'll want to be running the codes there. Thanks, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu U.S. Naval Research Laboratory, Code 6134 phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

installation issues
by Bruce Ravel 23 May '02

23 May '02

Wow! I was gone for three days and a whole flurry of interesting issues came up. ;-) First off, I must thank Mark Jensen for these kind words: MJ> Ever since the Windows version of Athena was released, I've been MJ> hooked. I'm blushin'. Then Mark said: MJ> After an apparently successful installation, neither Athena nor MJ> Artemis ran, saying that they couldn't locate FileDialog.pm, MJ> widgets.pm, or Chemistry/Elements.pm (depending on the MJ> program). I was unsure where to get these files, or why they MJ> weren't where they were supposed to be, but the MJ> Chemistry/Elements.pm file made me wonder if that wasn't part of MJ> Atoms3.0. Indeed it is. After installing Atoms, Athena and MJ> Artemis both appear to be working correctly. So, Bruce might want MJ> to make a change in the intall.unix file to say that Athena (or MJ> Artemis) requires PGPlot, IFeffit, and Atoms. and Paul followed up with PF> Although I had "cleaned" and reinstalled several times athena and PF> Tk, I hadn't given a second thought to tkatoms -- it had no PF> relation to athena, right? wrong! Upon reintalling tkatoms, PF> athena worked. For starters, I am pleased that you both were able to solve your immediate problems. I am lucky to have such smart users. This exchange, though, underscores a serious problem. Eventually I intend to tie atoms, athena, and artemis (and a feff gui, should I or someone else ever get around to writing it) into a more cogent package. At the current stage both of the software development and of my understanding of the needs and challanges of GUI EXAFS software, it is a great convenience for them to be distinct programs. Still, as Paul pointed out, there are some ties between the codes. I use some object inheritance in athena and artemis, so they do share some code already. And one small feature of athena is only enabled if one of modules from Atoms is installed. So I have put myself in an awkward state of having neither a unified suite of programs nor three disjoint programs. Clearly what I need to do is to make a "clean room" for myself. That is, I need a place where I do not have any of my own software installed. Then I can install the tarballs individually or in any combination. If I do this, I should be able to avoid the problems that Paul and Mark reproted to this list. Thanks you both for the excellent feedback. I have said it before and I'll say it many times in the future -- this stuff is much better thanks to you guys providing feedback. Why having an out-of-date atoms resulted in the error that Paul described is a bizarre mystery to me. And why lacking atoms gave Mark an error about widget.pm is equally strange. I'll look into all of it and certainly will work on the installation scripts. Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu U.S. Naval Research Laboratory, Code 6134 phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0