TB2J-OPENMX（0.4.1） does not match the version ( 0.6.3) of TB2J

[gswylq gswylq]

Hi,

I run the SrMoO3 examples (given by TB2J-openmx) using OpenMX(0.4.1) and TB2J( 0.6.3), and an error occurs as follow. The OpenMX(0.3.3) and TB2J( 0.6.1) can work good for the SrMoO3 example of TB2J-openmx to extract the multibinit files.

################ errors ####

 #(python39_openmx) ggs:~/openmx_soft/EXAMPLE_python39_openmx$ ./openmx2J.py  --prefix openmx --elements Mn 

Traceback (most recent call last):

  File "/home/gswylq/openmx_soft/EXAMPLE_python39_openmx/./openmx2J.py", line 4, in <module>

    from TB2J_OpenMX.gen_exchange import gen_exchange

  File "/home/gswylq/miniconda3/envs/python39_openmx/lib/python3.9/site-packages/TB2J_OpenMX/gen_exchange.py", line 1, in <module>

    from TB2J_OpenMX.ffiparser import OpenmxWrapper

  File "/home/gswylq/miniconda3/envs/python39_openmx/lib/python3.9/site-packages/TB2J_OpenMX/ffiparser.py", line 10, in <module>

    from TB2J_OpenMX.cmod._scfout_parser import ffi, lib

ModuleNotFoundError: No module named 'TB2J_OpenMX.cmod._scfout_parser'

################

The error file is attached in this email.

Please give me some suggestions. 

Thank you.

Q LI

[Xu He]

Hi,

I don't think this is related to the new version. It seems TB2J-OpenMX is not properly installed. Could you try to reinstall it?

pip uninstall TB2J-OpenMX

pip install TB2J-OpenMX

and see if there is any error message during the installation.

Best regards,

HeXu

--
You received this message because you are subscribed to the Google Groups "TB2J" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tb2j+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tb2j/f1520663-6db3-4919-9acb-12a70835e43dn%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[gswylq gswylq]

Hi,

I use the commands (pip uninstall TB2J-OpenMX ,pip install TB2J-OpenMX) , but the  errors  remain the same.

I download the  TB2J-OpenMX master from github, and use the commands (python3 setup build and python3 setup install) solve this problem.

Thank you.

Q Li

[Xu He]

Hi,

Thanks!

This is mostly likely due to a multiple python/pip version problem. Can you check

- Is the pip used corresponding to the python used? Perhaps you can use "python3 -m pip" instead of pip to make sure that you're using the "right" pip.

- is the pip version up to date?

- Did you install TB2J-OpenMX previously with pip or setup.py? Is the uninstallation successful?

You can try to remove this directory manually with:

"rm -r /home/gswylq/miniconda3/envs/python39_openmx/lib/python3.9/site-packages/TB2J_OpenMX"

and install with pip again.

I do not recommend the use of "python setup.py install". Because with that it will be difficult to uninstall and upgrade.  This can cause troubles: when a new version is installed, sometimes the previous one is not properly cleaned, which will lead to problems such as modules not found.

Best regards,

HeXu

To view this discussion on the web visit https://groups.google.com/d/msgid/tb2j/f709bb9d-ea42-4ce1-9be3-3dc0a9acfd03n%40googlegroups.com.

[gswylq gswylq]

Hi,

Thank you for you reply.

I create a new env(python39) with a new pip using conda command, and install openmx with pip install TB2J-openmx and uninstall with pip uninstall  TB2J-openmx.

I think that conda may be incompatible with new TB2J-openmx. I try to install  openmx using pip, but it not unsuccessful. The setupy.py installation is good and also be successful uninstallation with the command ( python setup.py install --record log, and  cat log ｜ xagrs rm －rf).

However, the exchange.xml from TB2J 0.6.3 can be run by mulitibinit 9.4.2. If you have free time, please help me check the problem.

The install, uninstall and running process files is attached in email.

best

Q Li

[Xu He]

Hi,

Thanks for the file! I notice that with pip, it is using the cached .whl package. Could you try with the --no-cache-dir option to see if there is an error message? I have tried the latest conda/python3.9 and there is no problem in my local computer.

There is a bug in Multibinit 9.4: the spin_sia_add option cannot be used. The bug is already fixed but will only be publicly available in the next version (perhaps Multibinit 9.6). Could you try to set it to 0 and try again?

Best wishes,

HeXu

To view this discussion on the web visit https://groups.google.com/d/msgid/tb2j/0f8d2ee6-2387-4c6a-b7b9-e343cbddb043n%40googlegroups.com.