search

band-unfolding / banduppy

Python version ofthe BandUP code
GNU General Public License v3.0
16 stars 7 forks source link

AttributeError: module 'banduppy' has no attribute 'UnfoldingPath'. #30

Open josueclavijo opened 1 month ago

josueclavijo commented 1 month ago

First of all, thanks to Stepan and all collabs for developing this exciting BandUPpy code.

Excuse me if my question is a very basic one, but really i don't know what i'm missing to do in order to get past of this error:

Traceback (most recent call last): File "/home/zephyrus/Descargas/Si-for-banduppy/./run_banduppy.py", line 18, in unfold_path=banduppy.UnfoldingPath( AttributeError: module 'banduppy' has no attribute 'UnfoldingPath'. Did you mean: 'Unfolding'?

I'm trying to use banduppy to unfold supercell structures coming from DFT using Quantum Espresso 7.3., after having banduppy and irrep packages installed via pip.

Any advice will be greatly appreciated.

Regards,

Josh

stepan-tsirkin commented 1 month ago

Hi,

the structure of the code has chamged in recent versions, so, please try to follow the updated tutorials: https://github.com/band-unfolding/banduppy/tree/master/tutorials

It is generally the same for QE and Vasp, except the constructuin of the Bandstructure object by irrep.

josueclavijo commented 1 month ago

Respected Stepan, I'm honored, and really thanks for taking the time to answer my queries. The code now works like a charm.

Just to gain a deeper undestanding of your code, pardon me if i ask to you now the following things:

1 - Are critical to get meaningful results the following warnings, or should I just ignore them

Warning 1

DeprecationWarning: dict interface (SpglibDataset['origin_shift']) is deprecated.Use attribute interface ({self.class.name}.{key}) instead

Warning 2

Neither refUC nor shiftUC were specified in CLI and searchcell was False. Taking 3x3 identity matrix as refUC and shiftUC=(0,0,0). If you want to calculate the transformation to conventional cell, run IrRep with -searchcell ???

2 - Can the constraint over the supercell size (setting the transformation matrix to just being defined using integer numbers) be modified so the transformation matrix could be expressed using any rational or decimal numbers? this is because when relaxing supercells, the transformation matrix with respect to an also relaxed unit cell is not always an integer number relationship, due to the minute changes in forces acting over atoms that arise as a function of the different functional used in the supercell and unit cell relaxations calculations...