SFN parsing available

10 views
Skip to first unread message

Matyáš Novák

unread,
Sep 21, 2026, 1:20:28 PMSep 21
to ase2s...@googlegroups.com

Dear all,

Thanks to Ridha's initial parser, I have now added support for reading, writing and visualizing SFN files (shape functions used in FULLPOT calculations).
The functionality is now available in the refactor branch. 

For the visualization, I took inspiration from several AI-generated suggestions and further improved them. I hope at least some of the visualization modes will prove useful. Feel free to experiment with them and let me know if you have any suggestions for improvement.

You can visualize an SFN file using:

ase2sprkkr plot file.sfn [--cell|--periodic|--shape|--slice|--radial]

The following visualization options are available:

--mesh       Show the distinct Voronoi meshes (the SFN default).
--cell       Place each distinct mesh in the crystallographic cell.
--periodic   Show the periodic tessellation clipped to one cell.
--shape      Reconstruct the S=0.5 shape-function isosurface.
--slice      Show a planar section of the reconstructed shape function.
--radial     Show the radial spherical-harmonic coefficients.

Best regards,

Matyáš

PS: Warning: if no bugs are reported in the refactor branch (though I suspect nobody has tried it yet :-)), I'll merge it 
into the default develop branch and make a new release.

Ridha Eddhib

unread,
Sep 22, 2026, 4:59:53 AMSep 22
to Matyáš Novák, ase2s...@googlegroups.com
Dear Matyáš,

Thanks for the update. Aki would be the best person to test it, since I tried to implement his logic; hopefully I succeeded.

The reason behind this parser is to handle shape functions in the OSCARpes database. I wanted all the parsers to live in one package rather than being scattered across different places.

There are two more parsers still in my develop branch: one for LMTO potentials and one for the *spec.out files from ARPES calculations.

Best,
Ridha
--
You received this message because you are subscribed to the Google Groups "ASE2SPRKKR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ase2sprkkr+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ase2sprkkr/7a30a054-3709-4410-868d-ce054fbe121e%40fzu.cz.

Matyáš Novák

unread,
Sep 22, 2026, 5:41:43 AMSep 22
to ase2s...@googlegroups.com
Dear Ridha,
Thanks for the update. Aki would be the best person to test it, since I tried to implement his logic; hopefully I succeeded.
I think you did. I want to allow not only reading, but writing, too, since there is CONTROL.SFNFIL, where you can supply your 
own modified SFN file. Thus, I put your parser to AI and ask her what is needed to reimplement it as ASE2SPRKKR grammar
and it succedded (after some collaborative work on shared functionality, e.g. NumpyArrays of given length, repetition of sections etc.),
 so your parser was apparently well :-).


I wanted all the parsers to live in one package rather than being scattered across different places.
Agree, it is the ultimate goal. 
There are two more parsers still in my develop branch: one for LMTO potentials and one for the *spec.out files from ARPES calculations.
Please, try to send me whatever you have. Preferably not in one package, but 
one functionality per one "package". Or just create github issue, where you can 
put your scripts, if you like (if you don't mind them being publicly visible).

Even if it is not fully working: AI is pretty cappable to take SPRKKR sources and extract the 
logic from it, so I will be probably able to fix, if there will be some issues. 
The much bigger problem is to force AI not to write "unsustainable bloatware", 
I keep having to drill it into her head that she needs to follow sound design principles
and not to duplicate the code :-)
Best,
Ridha
Best,
Matyáš
Reply all
Reply to author
Forward
0 new messages