XML to QuantumEspresso Wave Function Conversion Tool?

[kgreeves]

Hello,

I am interested in using the QuantumEspresso post-processing tools to perform an additional analysis on wave function data that I generate from QBox simulations.

Are there currently any tools or any plans to create tools that convert the Qbox *.XML file to the QuantumEspresso *.wfc{N} files? If not, are there any workarounds to interface the two codes?

Thanks for your help!

Kyle Reeves
kgreeves@email.unc.edu
UNC-CH, Chemistry Dept.

[mgoldey]

Hi,

I've actually done close to the reverse out of convenience - taking cube files from QE post processing code and making Qbox inputs from them.

If you save the wavefunction (save command), you can give it the -text argument, which writes it in human-readable format, rather than binary. Each <grid_function> section (1-particle wfn) has an X,Y,Z (should be opposite order from a cube file, I think) ordering determined in line 1592 of SlaterDet.C (as of 1.60.4)

I don't know the .wfc file format, but you should be able to identify its syntax, translate the files to that format, and use iotk convert tool to make a binary that would be readable in some fashion by QE. Might require some significant testing to sort out parallel issues - see wf_collect keyword from PWSCF man pages.

Caveats
(1) The cell location differs between QE and Qbox
(2) The grid size does not necessarily line up - be careful

Matthew

[fgygi]

You may also consider using the Qbox "plot" command which generates a file in "cube" format compatible with e.g. the VMD visualization program.

For conversion to the QE wfc format, the real space grid data must be converted to Fourier coefficients as I believe the wfc files contain Fourier coefficients. In order to perform the conversion, the plane wave basis must be generated in the same way QE does (possibly using QE functions). Note that the plane wave basis generated by the Qbox Basis object is not ordered in the same way as QE's basis.

The rationale for using a real-space representation in Qbox restart files is that real space data is unambiguous and does not depend on the ordering of a plane wave basis set that can be code-specific in complicated ways (consider e.g. the ordering of G vectors having the same norm).

Depending on the post-processing tool that you would like to use, we could try to implement it within Qbox. Please let me know what functionality would be needed.

Francois

[lesheng]

Hi Matthew,

I am recently trying to convert a cube file to XML that can be read by QBox. Could you give me a more detailed description on how to do this please?

Thanks in advance.
Best,
Lesheng Li

Hi,

I've actually done close to the reverse out of convenience - taking cube files from QE post processing code and making Qbox inputs from them.

If you save the wavefunction (save command), you can give it the -text argument, which writes it in human-readable format, rather than binary. Each <grid_function> section (1-particle wfn) has an X,Y,Z (should be opposite order from a cube file, I think) ordering determined in line 1592 of SlaterDet.C (as of 1.60.4)

I don't know the .wfc file format, but you should be able to identify its syntax, translate the files to that format, and use iotk convert tool to make a binary that would be readable in some fashion by QE. Might require some significant testing to sort out parallel issues - see wf_collect keyword from PWSCF man pages.

Caveats
(1) The cell location differs between QE and Qbox
(2) The grid size does not necessarily line up - be careful

Matthew