Bingqing Cheng

We'll need to discuss a bit about this. Do you mean that you want to import ASAP as a library? IMO, given the current design, it makes more sense to...

Okay. I see. I'll dig up on how to do this.

Fantastic. I tried it, seems to be working great!

It looks to me that the dataset.xyz doesn’t contain descriptors, which are needed for the sparsification. Please try to use asap gen_desc to generate the descriptors first. On Thu 3....

None of these are actually the Ref.33 ones, although it is possible to have it. Now the reducers are documented: https://bingqingcheng.github.io/gen_desc.html

So you mean PCA for a big dataset that is a collection of subsets, and then plot the PCA of the subsets on different panels?

I think you are right. I'm a bit snowed under these couple of days but will fix it before the end of the week.

This is done now. In summary, one can just compute atomic_descriptors using, e.g. # compute atomic descriptors only asapxyz.compute_atomic_descriptors(desc_spec_dict=soap_spec, sbs=[], tag='tio2-atomic', n_process=1) and fetch it by: dm = asapxyz.fetch_computed_atomic_descriptors(['soap1']) Will...

I tried and can replicate the output. The problem is with the chemiscope output. When I do asap map -f ASAP-desc.xyz -o xyz -dm '[*]' -ua pca -d 2 there's...

It has been a while since I looked into the updates of chemiscope, but back then I recall that it can only handle the per-atom output for a single frame....