I gave a try to Larch plugin system and I enjoyed its simplicity. Here a first version of a simple plugin to read SPEC files:

https://gist.github.com/maurov/5284508

I plan to improve the class by adding more methods, but first I would validate if I'm going in the right direction for implementing new plugins.

Regarding the Python class to read SPEC files, I have evaluated:

(1) specfilewrapper in PyMca (rewritten in Python from the 'specfile' C class developed at ESRF) (2) pyspec (http://pythonhosted.org/pyspec/) (3) specparser (https://github.com/tpikonen/specparser) (4) the one included in TDL

To my opinion, (1) is the best choice for simplicity, versatility and maintenance. (2) is a nice project too but cannot handle motors names that include a dot, as 'Mono.Energy'.

After a fit, this is a symbol: fit.datasets[0].epsilon_k.

If I ask the larch server for that symbol via the get_data command, nothing is returned (the XMLRPC version of an empty response).

If, however, I do this right after the fit: foo = fit.datasets[0], then ask the server for foo.epsilon_k, I get the expected return value (a list of epsilon values for each k-weight).

Thus, the problem seems to be with the [0].

This isn't an immediate problem for Demeter -- Demeter gives another name to each of the fit.datasets[N]. However, it's either a bug or I am misunderstanding the syntax of using the server for that kind of symbol.

The packaging step of xraylarch fails on archlinux since version 0.9.38 (0.9.37 worked fine, 0.9.39 still fails) with:

running install_scripts
Traceback (most recent call last):
  File "setup.py", line 311, in <module>
    shutil.copy(src, dest)
  File "/usr/lib/python2.7/shutil.py", line 133, in copy
    copyfile(src, dst)
  File "/usr/lib/python2.7/shutil.py", line 97, in copyfile
    with open(dst, 'wb') as fdst:
IOError: [Errno 13] Permission denied: '/usr/bin/larch_server'
==> ERROR: A failure occurred in package().

The command executed (as non-root) is:

python2 setup.py install --skip-build --root="${pkgdir}" --prefix=/usr

My guess is, that --root=... is not honored in all places, thus /usr/bin/larch_server gets touched instead of ${pkgdir}/usr/bin/larch_server

regards, Erich

Hi I was wondering why the data group add attributes: pre1 pre2 norm1 norm2 after pre_edge() while kmin, kmax , kweight are not stored. Maybe will be nice make a dictionary with the parameter of EXAFS extraction in the group. What do you think about?

@newville I have observed an annoying behavior on xas_viewer (master on Linux) that seems a bug to me:

• click on a group -> it gets highlighted in the left panel
• push the plot button -> the highlight disappears

Can you reproduce this misbehavior from your side?

I find myself dissatisfied with show(), particularly for a group with a lot of members. This stab at an "improvement" adds flags for color, truncating lengthy lists and tuples, and suppressing bound methods. All this together makes it a bit easier -- at least for me -- to examine the group's contents.

Note that

display(group, truncate=False, with_color=False, with_methods=True)

behaves just like show(group).

I put this in plugins/local/ and in the _local group. It's not clear to me quite what the purpose of "local" is. plugins/std might have been more appropriate, but I did not want to put this in _builtin.

Does anybody is planning to port "atoms" to Larch? If not, I could probably invest some time in this. I never went through the source code of atoms itself, so the ideas I have in mind on how to implement it could result not compatible with Demeter then...

The Larch plugin I have in mind should simply do: 1) read a CIF file and some keyword parameters; 2) based on this, it will return a FEFF* or FDMNES input file.

For the templates, I think is enough the 'string.Template' included in the Python standard library (there are tons of powerful template systems available, but this will avoid to include a further dependence in Larch).

For the CIF parsing, well, I think we are obliged to include a dependence on an external library. Do you have any preference for this? I'm evaluating some of them and at the moment the one I like the most is in ASE (https://wiki.fysik.dtu.dk/ase/index.html)... well, this is using a sledgehammer to crack a nut.

Anyway, your feedback is welcome.

Following issue #324 :

Jupyter notebook example using xraylarch and lmfit in Python to treat spectra at the Fe K-edge.

Let me know if it's good for you!

Hi,

I tried installing Larch using miniconda as well as the bash script on Linux x86-64. In both case I end with an error message related to wx when trying to import larch:

>>> import larch
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/charles/xraylarch/lib/python3.6/site-packages/larch/__init__.py", line 28, in <module>
    from .shell import shell
  File "/home/charles/xraylarch/lib/python3.6/site-packages/larch/shell.py", line 10, in <module>
    from .interpreter import Interpreter
  File "/home/charles/xraylarch/lib/python3.6/site-packages/larch/interpreter.py", line 18, in <module>
    from . import builtins
  File "/home/charles/xraylarch/lib/python3.6/site-packages/larch/builtins.py", line 22, in <module>
    from . import fitting
  File "/home/charles/xraylarch/lib/python3.6/site-packages/larch/fitting/__init__.py", line 6, in <module>
    import wx
  File "/home/charles/xraylarch/lib/python3.6/site-packages/wx/__init__.py", line 12, in <module>
    __version__ = wx.__version__.VERSION_STRING
AttributeError: module 'wx' has no attribute '__version__'

This issue is solved by installing wxpython with conda.

Should wxpython be added as a dependency?

Hi Matt,

This morning, with a freshly git pull-ed larch, I am unable to start the larch interpreter.

Below is the traceback along with the output of these print statements

                print "\n", mod, filelist
                for fname in filelist:
                    print fname, fname[:-3]

at this location: https://github.com/xraypy/xraylarch/blob/master/lib/builtins.py#L393

/home/bruce/.larch/plugins/wx ['scanviewer.py', 'periodictable.py', 'gui_utils.py', 'xrfdisplay_fitpeaks.py', 'plotter.py', 'xrfdisplay.py', 'mapviewer.py', '__init__.py', 'gse_dtcorrect.py', 'xrfdisplay_utils.py', 'parameter.py', 'mapimageframe.py']
scanviewer.py scanviewer
periodictable.py periodictable
gui_utils.py gui_utils
xrfdisplay_fitpeaks.py xrfdisplay_fitpeaks
  File "/usr/local/lib/python2.7/dist-packages/larch/builtins.py", line 411, in _plugin_file
    out = imp.load_module(plugin, fh, modpath, desc)
  File "/home/bruce/.larch/plugins/wx/xrfdisplay_fitpeaks.py", line 24, in <module>
    from parameter import ParameterPanel
  File "/usr/local/lib/python2.7/dist-packages/larch/builtins.py", line 397, in _plugin_file
    ret =  _plugin_file(fname[:-3], path=[mod])
  File "/usr/local/lib/python2.7/dist-packages/larch/builtins.py", line 419, in _plugin_file
    write(traceback.print_tb(tback))
Traceback (most recent call last):
  File "/usr/local/bin/larch", line 99, in <module>
    shell = larch.shell(banner_msg=BANNER, quiet=options.quiet)
  File "/usr/local/lib/python2.7/dist-packages/larch/shell.py", line 53, in __init__
    self.larch  = Interpreter()
  File "/usr/local/lib/python2.7/dist-packages/larch/interpreter.py", line 142, in __init__
    builtins._addplugin(pdir, _larch=self)
  File "/usr/local/lib/python2.7/dist-packages/larch/builtins.py", line 443, in _addplugin
    return _plugin_file(plugin)
  File "/usr/local/lib/python2.7/dist-packages/larch/builtins.py", line 400, in _plugin_file
    write(traceback.print_tb(tback))
TypeError: expected a character buffer object

The location of the beginning of the traceback seems to be the line in plugins/wx/xrfdisplay_fitpeaks.py that says

from parameter import ParameterPanel

and the problem seems to be that parameter.py has not been imported yet (it comes later in filelist).

I can make larch finish startup by manually prepending parameter.py to filelist.

Do you need to maintain a dependence tree so that plugins are imported in the right order?

Hi Matt @newville , I'm using Larch for some time. Everything worked fine until this weekend. I can use XASviewer to process data however I can't open any of the saved projects. When I try to open the saved project, I got error message in the terminal (see below) and the project has not been opened. I also tried to load new data and saved as a new project - l also couldn't open it. Any ideas how to fix it?

run in conda environment "base":

c:\users\annaw\xraylarch\Scripts\xas_viewer.exe

Traceback (most recent call last): File "c:\users\annaw\xraylarch\lib\site-packages\larch\wxxas\xasgui.py", line 799, in onReadDialog self.onRead(path) File "c:\users\annaw\xraylarch\lib\site-packages\larch\wxxas\xasgui.py", line 809, in onRead self.show_subframe('athena_import', AthenaImporter, File "c:\users\annaw\xraylarch\lib\site-packages\larch\wxxas\xasgui.py", line 754, in show_subframe self.subframes[name] = frameclass(self, **opts) File "c:\users\annaw\xraylarch\lib\site-packages\larch\wxlib\athena_importer.py", line 83, in init self.a_project = read_athena(self.filename, do_bkg=False, do_fft=False, File "c:\users\annaw\xraylarch\lib\site-packages\larch\io\athena_project.py", line 915, in read_athena aprj.read(filename, match=match, do_preedge=do_preedge, do_bkg=do_bkg, File "c:\users\annaw\xraylarch\lib\site-packages\larch\io\athena_project.py", line 780, in read data = parse_perlathena(text, self.filename) File "c:\users\annaw\xraylarch\lib\site-packages\larch\io\athena_project.py", line 353, in parse_perlathena raw[key] = np.array([float(x) for x in aeval(text2list(t))]) File "c:\users\annaw\xraylarch\lib\site-packages\asteval\asteval.py", line 314, in call return self.eval(expr, **kw) File "c:\users\annaw\xraylarch\lib\site-packages\asteval\asteval.py", line 324, in eval raise ValueError(msg) ValueError: length of text exceeds 50000 characters[]

Best, Anna

Hi there. In section 13.9 of the Larch Github manual, it refers to computing the f' and f" values as in Cross et al. (1998). I have managed to install the relevant modules and import the relevant libraries. However, I cannot work out how to plot the results. It is described in the example as:

''' newplot(dkk.energy, dkk.f2, label='f2', xlabel='Energy (eV)', ylabel='scattering factors', show_legend=True, legend_loc='lr') plot(dkk.energy, dkk.fpp, label='f"(E)') plot(dkk.energy, dkk.f1, label='f1') plot(dkk.energy, dkk.fp, label='f'(E)') '''

However this does not work. I have tried matplotlib.pyplot and also wxmplot but neither work. I don't think the wxpython library is compatible with my Spyder IDE, as Python crashes whenever I try to use Wxpython to load this result. Does anybody have a solution?

@newville a feature request I get usually from the users is to have the possibility to keep the selected zoom in an plot window when replotting. This would be something like the following:

• plot something (in plot window: plot_N)
• with the mouse zoom a region in plot_N
• when replotting in plot_N (current group or selected) keep the XY region as selected before
• to reset the plot region either right click on the plot window or select one of the standard ranges given in the combo box

What do you think?

@newville this is a follow-up of #357 . I have tried to fix it by myself, but I got lost in the Wx GUI stuff. If you could help with this soon, it would be great. The users at the beamline still find difficult to adjust the energy shifts with Larch. As a consequence, they go back using Athena...

The following are checked on the latest master version (xraylarch-0.9.65.post133+g5c834772)

1. bug in copying energy reference groups
   • load two data groups (group1, group2), plus an energy reference one (group_eref)
   • select group1 and group2
   • click on group1 and set group_eref as energy reference
   • click copy in the energy reference
   • click on group2 -> BUG the energy reference is not group_eref (expected) but group2
2. energy_shift is not kept
   • click on group1
   • put 1 in "energy shift from original" and press enter (just in case)
   • click on group2
   • click back on group1 -> BUG energy shift from original is 0
3. energy_shift is not applied
   • click on group1
   • right click to copy group1 into group1_1
   • select group1 and group1_1
   • plot selected groups in dmu/dE -> are identical
   • click on group1_1 and put 1 into energy shift from original
   • replot selected groups -> BUG the two groups are not shifted
4. applying energy_shift to reference does not propagate to groups linked to it
   • click on group1, put energy_shift to 0 and select group_eref as energy reference
   • click on group_eref and add 1 to energy_shift
   • click group1 -> BUG energy_shift is still at 0

The workaround to fix bugs 2 and 3 and get the expected behavior is to deselect plus select again the "auto?" next to E0 selection. I do not understand why this solves bug 2. For bug 3, we may apply directly the energy_shift value also to E0, or even better, E0 should simply refer to group.energy not to group.energy_orig.

Please, just to clarify my understanding of the energy_shift behavior in the GUI, could you check that the following is correct?

• read a group (g=Group()) -> g.energy = g.energy_orig + energy_shift, energy_shift = 0
• when setting energy_shift to a value: g.energy = g.energy_orig + energy_shift
• everything else (E0, plots, exports, whatever) uses g.energy
• when read energy reference group (g_ref = Group()) -> g.energy_ref = g_ref or g.energy_ref = g_ref.name

I hope this is clear, if not, please, let me know.

Dear @newville

I have been trying to use python (in Spyder) for path fitting. I just encountered a new (it worked yesterday) error that I could not see the plots of fitting results, but the fitting itself worked out with a ".out" file: " plot_chifit(dset, title='Ti66_fitting', rmax=6)

File "C:\ProgramData\xrayLarch\lib\site-packages\larch\wxlib\xafsplots.py", line 683, in plot_chifit _plot(dataset.data.k, data_chik+offset, xmin=kmin, xmax=kmax,

File "C:\ProgramData\xrayLarch\lib\site-packages\larch\wxlib\plotter.py", line 486, in _plot plotter = _getDisplay(wxparent=wxparent, win=win, size=size,

File "C:\ProgramData\xrayLarch\lib\site-packages\larch\wxlib\plotter.py", line 339, in _getDisplay display, isnew = _get_disp(symname, creator, win, display_dict, wxparent,

File "C:\ProgramData\xrayLarch\lib\site-packages\larch\wxlib\plotter.py", line 327, in _get_disp display = creator(window=win, wxparent=wxparent,

File "C:\ProgramData\xrayLarch\lib\site-packages\larch\wxlib\plotter.py", line 114, in init PlotFrame.init(self, parent=None, size=size,

File "C:\ProgramData\xrayLarch\lib\site-packages\wxmplot\plotframe.py", line 18, in init BaseFrame.init(self, parent=parent, title=title,

File "C:\ProgramData\xrayLarch\lib\site-packages\wxmplot\baseframe.py", line 58, in init wx.Frame.init(self, parent, -1, title, **kws)

PyNoAppError: The wx.App object must be created first!"

Could you please help me with this?
Thank you in advance, Best regards

@newville

In view of having Larch more independent from the Wx GUI, and have the possibility to build more Jupyter friendly examples (or applications) I would propose the following actions:

• [ ] create a new module larch.controller
• [ ] move all the non-Wx part from larch.wxxas.xas_controller.XASController to larch.controller.LarchController
• [ ] XASController will inherit from LarchController
• [ ] assign _larch.symtable as model for LarchController
• [ ] dedicate _larch.symtable._data as the main data container for all data groups created during a session
• [ ] add in LarchController all the logic for handling data exchange from the model to the view (= GUI)

Does this sounds good to you?