May 2005 - Ifeffit - millenia.cars.aps.anl.gov

Re: athena for OSX.4
by Paul Fons 10 May '05

10 May '05

Please excuse me for forwarding a copy to the list in case there are others out there that have run into the same problem. For all those who want to be Tigers (Mac OS 10.4), here is a brief summary of the installation procedure. I should add that there is not an installer that will do everything at once, but the procedure is really pretty easy. I have not tried installing horae on a brand new system, I have only tried (and succeeded) in making a previously functional horae-051 under Mac OS 10.3 work properly under Mac OS (Tiger) 10.4. The errors you see are generated because there is a new version of perl installed (5.8.6) which doesn't know about Perl/Tk, the graphics system that Bruce uses for all of his packages. The solution is simple, you must download and install Perl/Tk and then reinstall horae. After that everything just works again. The build procedure is nearly automatic, but you need to have installed the developer tools (which are included on the Tiger installation DVD). First you need to download Tk from CPAN (the server where free Perl software is hosted). Here is the procedure. 0. I assume you had an installed and functional horae running under 10.3 and did an upgrade to 10.4 1. Download the source for Tk by a. make a new (temporary) scratch directory mkdir temp; cd temp b. download Tk going in your browser of choice go to the web page: http://search.cpan.org/~ni-s/Tk-804.027/ and hitting the download button on that page (the file will be named Tk-804-2.027.tar.gz) c. uncompress the file in your scratch directory tar zxf /the-path-to-your-download-directory/ Tk-804-2.027.tar.gz d. change to the Tk directory, compile, and install the new Tk cd Tk-804.027 perl Makefile.PL make sudo make install e. now you must download and install the latest version of horae via Bruce's web page go to the site http://leonardo.phys.washington.edu/~ravel/ software/downloads.html#unix and use the download link for horae-051.tar.gz f. decompress and install the new horae source cd ../temp (move up from the Tk directory to the top level of your new temporary directory) tar zxf /the-path-to-your-download-directory/ horae-051.tar.gz cd horae-051 perl Makefile.PL make sudo make install That's it. You are done. I just popped up a X11 window (you did remember to install the X11 program which is an optional install on the Tiger DVD accessible using the "Customize" button in the installer, right?) and typed artemis & and poof, artemis appeared. Note that if you did a clean install you should still be able to get everything working by downloading Matt's installer and installing the 10.3 copy (it won't work out of the box) and then following the above procedure. I have not verified this, however, so caveat empor as the Romans said... Hope this helps. Paul On May 10, 2005, at 3:08 AM, denauwer(a)tele2.fr wrote: > Dear Paul, > > We have contacted Bruce because we want to use Athena on Mac OSX.4 > and it is not working. We have tried from the terminal and we got > the following message: > > "Can't load '/Library/Perl/5.8.1/darwin-thread-multi-2level/auto/Tk/ > Event/Event.b > undle' for module Tk::Event: dlopen(/Library/Perl/5.8.1/darwin- > thread-multi-2lev > el/auto/Tk/Event/Event.bundle, 1): Library not loaded: /System/ > Library/Perl/5.8. > 1/darwin-thread-multi-2level/CORE/libperl.dylib > Referenced from: /Library/Perl/5.8.1/darwin-thread-multi-2level/ > auto/Tk/Event/ > Event.bundle > Reason: image not found at /System/Library/Perl/5.8.6/darwin- > thread-multi-2lev > el/XSLoader.pm line 68. > at /Library/Perl/5.8.1/darwin-thread-multi-2level/Tk/Event.pm line 10 > Compilation failed in require at /Library/Perl/5.8.1/darwin-thread- > multi-2level/ > Tk.pm line 13. > BEGIN failed--compilation aborted at /Library/Perl/5.8.1/darwin- > thread-multi-2le > vel/Tk.pm line 13. > Compilation failed in require at /Applications/Ifeffit/bin/athena > line 53. > BEGIN failed--compilation aborted at /Applications/Ifeffit/bin/ > athena line 53. > Undefined subroutine &Tk::Event::CleanupGlue called at /Library/ > Perl/5.8.1/darwi > n-thread-multi-2level/Tk/Event.pm line 3. > END failed--call queue aborted at /Applications/Ifeffit/bin/athena > line 53." > > Could you if you have time, please, explain what we have to do to > fix this problem? > > Thank you for your help, > Christophe Den Auwer > >

1 0

Horae in Mac OS 10.4 (Tiger)
by Paul Fons 09 May '05

09 May '05

Being one who likes going where few have gone before, I went and installed Tiger (which has perl 5.8.6 installed by default by the way). I then installed the latest Tk and followed that up by the usual incantations for the horae-051 and all was well. It works without trouble. This just in case anyone else out there wants to use the new search feature spotlight (very handy for those pdfs, source code etc.) I can search my 20 gigabyte home directory in about 10 seconds for content (not just filename). Neat stuff. Paul

3 2

Comments on the new version
by Gerrit Schmithals 06 May '05

06 May '05

Hello Bruce, thank you for your constant efforts on improving this great analysis software. I just want to report a few observations on the new versions of Athena and Artemis. I use Windows 2000, Ifeffit + Update 014. - The file http://cars9.uchicago.edu/ifeffit/updates/update.dat still indicates an update level of 012. As discussed before it is easy to update the windows version by downloading the updater zip-files and extract them to the program folder. - The csv-file exported from Athena as two extra commas in the title lines so that title and data is misaligned. - When exporting data from Artemis choosing "Save selected group as" the x-value column is only exported if the original dataset is also selected. - I was often not able to export data via File | Save Data/Fit/Residual As | chi(k) (Artemis message below). I figured out, that it works when the group of the original chi-data is selected in the Data & Paths-window, but not when Fit or Residual is chosen. Another bug: if one of the FEFF-paths is selected while exporting via "Save data as" then the values of the original chi-data will be exported, but the filename and header will reflect the chosen path. I was initially trying to export some FEFF-paths using this method and was quite confused that they all look the same :). A tip that I can give to other users: it something seems to be wrong with the exported data, it is always possible to extract the correct data by saving the project to disc and opening it with an unzip program. Cheers, Gerrit Artemis message: Ifeffit checked for updates recently. Tk::Error: Can't call method "param_summary" on unblessed reference at script/ar temis line 7878. [\&main::save_data,fit,k] (menu invoke) main::__ANON__('Tk::Error: Can\'t call method "param_summary" on unbless ed refer...') called at blib\lib\Tk.pm (autosplit into blib\lib\auto\Tk\Error.al ) line 402 Tk::Error('MainWindow=HASH(0x219c69fc)', 'Can\'t call method "param_summ ary" on unblessed reference at scr...', '[\&main::save_data,fit,k]', '(menu invo ke)') called at Tk.pm line 340 eval {...} called at Tk.pm line 340 Tk::MainLoop() called at script/artemis line 1798 require main called at ../blib/lib/PAR.pm line 260 PAR::_run_member('Archive::Zip::ZipFileMember=HASH(0x11c0104)', 1) calle d at script/main.pl line 4 require main called at ../blib/lib/PAR.pm line 260 PAR::_run_member('Archive::Zip::ZipFileMember=HASH(0x11c0254)') called a t ../blib/lib/PAR.pm line 226 PAR::import('PAR') called at -e line 824 __par_pl::BEGIN() called at script/artemis line 0 eval {...} called at script/artemis line 0

3 4

Need an advice
by Steven S. Lim 06 May '05

06 May '05

Dear all, I have a question to ask all of you as follows. We have been investigating the EXAFS of some partially reduced Co materials (by Co EXFAS) and would wish to obtain some qualitative information on the Co metal particles. If, in fact, the Co had some average coordination environment because we have a distribution of Co particle sizes with oxygen adsorption, analyzing the windowed Co-O and Co-Co peaks together as a linear combination and assessing the average oxygen and Co coordination from the Co edge absorption (what we do with Athena) would appear to give a useful answer. However, if what we have is closer to a physical mixture of small Co metal particles and Co oxide particles, it would seem that it would be more appropriate to analyze separately the windowed Co-O and Co-Co peaks in R-space, to somehow normalize the Co-Co to the fraction of metal in the sample and use this as a qualitative measure of metal particle size. Is there a way to do this with Athena or other method? Or do you know of a reference where this has been discussed? Of course, we can (and will) make experimental mixtures as references but wonder if there is a different way of doing the analysis or if there is some literature on this problem (which we have not been able to locate). Thank you. Steven

4 9

new updates for windows
by Matt Newville 05 May '05

05 May '05

Hi Folks, As Gerrit S mentioned earlier, there is now a new updater for Windows that will install the latest versions of Athena, Artemis, Diana, and Hephaestus. In addition, there is an upgrade of the Ifeffit library to 1.2.7... a source release will be available soon -- if you're interested, it's already under http://cars.uchicago.edu/ifeffit/src/ but I haven't updated all the web pages or sourceforge yet. I think Bruce has discussed most of the new features in his programs. One thing that we should warn you about is that these versions of Bruce's programs use a new method to make Windows executables. Really, this is better for many reasons, but it currently means a noticeable lag in startup time, even before a splash screen pops up. We're looking into this, but please be patient! Thanks, --Matt

1 0

How is the path data exported?
by Gerrit Schmithals 03 May '05

03 May '05

Hello, I just started to use Artemis again and at the moment I am a bit confused about the results. In one Artemis experiment I just select two FEFF-paths and choose a "sum of paths" which produces the result shown in the appended picture (SumOfPaths2). Shouldn't the fit be more or less look like the sum of the two paths? I also exported both paths as chi(k)-data and manually added them which produced the same result as the chi(k) of the fit. But when I use more paths I get a different result: the manually summed chi has a greater amplitude (see chi.png). Could that be related to the fact that the fit parameters are applied to the sum of path but not to the exported paths? Cheers Gerrit

2 1

Reporting math expressions in scientific papers
by MauRo 02 May '05

02 May '05

Hi folks, I would like to ask you if it exist one common convention, apart the XAFS society one, to report IFEFFIT results when you have unusual math expressions. In particular, I have used a two-phase model (two FEFF calculations) to fit my sample and for the amplitude variable I have these math expressions: amp_feff1 = amp * x amp_feff2 = amp * y = amp * (1-x) In this way I can estimate the global amplitude variable and introduce a "percentage" of the two chemical compounds (correlation is kept under control). How to explain what I have done to a non-ifeffiter? Mauro --- Mauro Rovezzi <laluvedda(a)gmail.com> Physics student @ GILDA BM08 c/o ESRF (Grenoble, France)

4 3

???
by Elena Ion 02 May '05

02 May '05

Hello, Questions from a beginner in EXAFS...using Athena and Arthemis >From where to get the error bars for number of neighbours, distance and sigsqr (ss)? What are Chi-square,Reduced Chi-square and R-factor and how should be their values in a fit? Sometimes it happen that "R-factor for this data set" does not have the same value as the first R-factor from results. Why? many thanks, Elena

3 4