Suggestion for heterogeneous reconstruction

Hi,

In heterogenous reconstruction, it would be useful to have an “input” option as an alternative to class identity, so that particle sets provided as separate inputs can be reconstructed separately within the same job.

This would I think be a flexible way of dealing with edge cases that are perhaps not worth accommodating individually.

For example, I would like to symmetry expand a particle set, and then reconstruct each symmetry operator subgroup, to generate a “symmetry expanded” set of volumes for classification - right now I have to run a separate reconstruction job for each symmetry operation, and it would be convenient to be able to do all of them in a single hetero reconstruct job.

Cheers
Oli

1 Like

Hi @olibclarke,

Would you be able to make a workflow where you generate X number of reconstructions from the output of a symmetry expand job? Or is it that to keep things less cluttered having these volumes being generated by a single job?

Additionally, what is the purpose of generating a volume for each symmetry group?

Cheers,
Kye

To keep things less cluttered, and also avoid all the repetitive clicking and dragging (which I think will also be involved in generating a workflow for this unless I am missing something)

In this specific case, we wanted to generate a volume of a pseudosymmetric entity in every possible orientation, to seed a classification. We could also do this using volume tools, but that would be equally painful.

Cheers
Oli

1 Like

Hi @olibclarke,

We have made note of this request.

Cheers,
Kye

1 Like

Just to add to this - it would also be useful to be able to specify a subset of classes from a single input. For example, let’s say I ran a Class3D job with 20 classes, but only the first 5 classes are good. I would like to reconstruct these first five classes with Het reconstruct, but right now I have to reconstruct all 20. Being able to provide either multiple individual classes as input, or a numerical class list as a parameter, would be useful I think.

Hi @olibclarke,

Thanks for the suggestion, we have noted this in our feature requests.

Thanks,
Kye