tmQM dataset splitting according to metal elements

[MarshallYan]

We would like to have a feature in the dataset module (read from dataset.toml) to split tmQM dataset according to metal elements. e.g. Load all molecules with Pt as the metal center.

We would like this module to support binary operators. e.g. Load molecules containing Pt or Cu as the metal center.

More features shall be discussed and added to this module. We would like to split other datasets as well, but details need to be discussed.

[chrisiacovella]

As an additional workaround, I'll add some additional datasets that limit to:

• Pd, Zn, Fe, Cu: 27685 molecules
• Pd, Zn, Fe, Cu and then exclude any molecules that contain elements not in the set: C, H, P, S, O, N, F, Cl, Br: 18515 molecules
• Pd, Zn, Fe, Cu, Ni, Pt, Ir, Rh, Cr, Ag: 60637
• Pd, Zn, Fe, Cu, Ni, Pt, Ir, Rh, Cr, Ag and then exclude any molecules that contain elements not in the set: C, H, P, S, O, N, F, Cl, Br::: 40866

Putting in the data filter is a more generally useful and dynamic approach we should definitely implement as well.