ABINIT Discussion Forums

Dear all

I used script to plot full GW band structure by python. In this script needed *_EIG.nc + *_GW file
after that I used command by type
# python abinit_eignc_to_bandstructure.py cubo2_EIG.nc + cubo2_GW
I got error like this usage: abinit_eignc_to_bandstructure.py [-h] files
abinit_eignc_to_bandstructure.py: error: unrecognized arguments: cubo2_GW

Please suggestion

Abdulmutta Thatribud
Department of Physics, Faculty of Science
Prince of Songkla University, Thailand

Dear Mutta,


The script you have to use is ab_extended_scissors_guess.py.

You would then do :

python ab_extended_scissors_guess.py cubo2_EIG.nc cubo2_GW (without the +)

Note that you need numpy and matplotlib installed in your python path.


David

Dear David

After as you suggest, I got ERROR see below
ERROR: the number of spins/kpoints is not the same in the GW and DFT files used to make the interpolation ... exiting now

For DFT band I used kpoints 4 4 4 and shifted 0 0 0
For GW file I also used kpoints 4 4 4 and shifted 0 0 0
but the error is the same

Please help
Mutta

Dear Mutta,


Could you please post all your input files that you used to generate the *GW and *EIG.nc file as well as these two files ? Otherwise, its difficult to tell you what's wrong ...


David

Dear David

I sent you my input file for LDA, band_LDA.in and for GW, cubo2_G0W0.in
and also output file *.nc and *_GW

For *.nc and *_GW, I tried to attach two files but the extension is not allowed
Can you tell me how can I do it


Mutta

Hello,

In order to attach these file, one workaround would be to rename them as cubo2_EIG.nc.in and cubo2_GW.in.

For the interpolation, are you using the EIG.nc from the second dataset of your cubo2 input file ? If not, that might be the error that you get.

Indeed, you should do :

python ab_extended_scissors_guess.py cubo2_DS2_EIG.nc cubo2_DS4_GW

(or something like that, depending on the name scheme you use, as defined in the *files file)

David

Dear David

Yes, I used the second dataset.

I sent you again for output file coming from LDA with second dataset and G0W0


Thank you
Mutta

Dear David

Now, I got full gw band structure and you are right for *DS2_EIG.nc and *_GW should be from dataset 2 as GW run
and then used *EIG.nc coming from LDA.

Thank you for your help
Mutta

Dear David

Hope it's ok to append my question to this post.

Following the above discussion, I've tried using the script with silicon and got the gw-interpolated band structure (*.bandstructure file) at the end.
I noticed that the energies in the *.bandstructure file seems to be given in "Hartree". Is this true? I would like to be sure that I got things right.
If so, is there any way to get the energy in the *.bandstructure file in "eV" directly from running the script?

Regards,
Thanusit