aiidalab / aiidalab/aiidalab-qe
Extend van der Waals support (vdW) and custom functional choice
- 主要语言
- Python
- 星标
- 33
- 派生
- 25
- 平均合并
- 6 小时 7 分钟
- 30 天内合并 PR
- 8
描述
# Van der Waals
* Check if it requires to change `input_dft` in the QE input, or it's just via the vdw_corr input flag (plus maybe one or two more, e.g. `dftd3_version`) (as this might conflict with adding `input_dft` as a code option, see later
* Support a few more, at least rvv10 and DF2 (check if they need a file in the pseudo folder and if AiiDA needs a node for that: see MC2D calculations, and check with Natasha
* (to decide) if supporting a few more is possible by just passing a different string to QE, maybe allow an "other/custom" option in the dropdown that shows a textbox?
# custom functional
* allow to pick a custom functional, setting `input_dft`: provide a short list of common options from the [funct.f90 of QE](https://gitlab.com/QEF/q-e/blob/master/Modules/funct.f90), and allow also here custom/other and people can type
* suggest to do this when pseudos have different functionals, and do not block anymore if the selected option is not "default/from pseudos" (which should be the default)
* double check this does not conflict with the vdW corrections
# Pseudo parsing
* double check the two ways of storing LDA (note that there are 2 LDA, PW and PZ), and PBE as a string in the pseudos, and make them equivalent in the check. E.g. " SLA PZ NOGX NOGC" and "functional="SLA PW NOGX NOGC" are the two LDA versions (check in funct.f90, only one of the two is "LDA" for QE), and functional = "functional=" SLA PW PBX PBC" should be PBE.
* In any case, if this is not done, but the `input_dft` can be customized, this is not a critical issue (but I would still fix it).
贡献指南
这个仓库没有索引到贡献指南
评估
这个 Issue 还没有评估数据。