You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a problem? Please describe.
We're looking to generate some properties for the QM7-b dataset such as polarizabilities and maybe some other linear response properties like specific rotation. QCA would be a great place to put these as a start towards a machine-learning effort, and some groups have already shown important use cases for these data.
Describe the solution you'd like
A spot for response tensors like the polarizability or Rosenfeld tensors would be great. I've been pointed towards AtomicResultProperties as a place for these to be added (thanks @loriab !).
Describe alternatives you've considered
Currently looking into rolling out Fractal to run the calculations through QCEng, we could harvest these things using some home-baked scripts but it feels like a waste when we could use the QCA infrastructure with a little tweaking. I'm told @bennybp and others at the MolSSI rolled out quite a number of calculations on ARC to populate QCA, so there may be some particulars that I don't know yet.
Additional context
As we're still in the planning stages, we're open to suggestions on what types of properties to consider and how to best organize them. For example: since we're looking at some frequency-dependent properties, it may be advantageous to have a frequency (or wavelength) field, which can simply be set to 0 for static properties to keep things consistent. These sorts of considerations will generally come up as new properties are added, but I'd like to be as consistent as possible! Any feedback is appreciated!
EDIT: a sister-issue is now up on the QCSchema GH for additional consideration.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a problem? Please describe.
We're looking to generate some properties for the QM7-b dataset such as polarizabilities and maybe some other linear response properties like specific rotation. QCA would be a great place to put these as a start towards a machine-learning effort, and some groups have already shown important use cases for these data.
Describe the solution you'd like
A spot for response tensors like the polarizability or Rosenfeld tensors would be great. I've been pointed towards
AtomicResultProperties
as a place for these to be added (thanks @loriab !).Describe alternatives you've considered
Currently looking into rolling out Fractal to run the calculations through QCEng, we could harvest these things using some home-baked scripts but it feels like a waste when we could use the QCA infrastructure with a little tweaking. I'm told @bennybp and others at the MolSSI rolled out quite a number of calculations on ARC to populate QCA, so there may be some particulars that I don't know yet.
Additional context
As we're still in the planning stages, we're open to suggestions on what types of properties to consider and how to best organize them. For example: since we're looking at some frequency-dependent properties, it may be advantageous to have a frequency (or wavelength) field, which can simply be set to 0 for static properties to keep things consistent. These sorts of considerations will generally come up as new properties are added, but I'd like to be as consistent as possible! Any feedback is appreciated!
EDIT: a sister-issue is now up on the QCSchema GH for additional consideration.
The text was updated successfully, but these errors were encountered: