cmelab / morphct

GNU General Public License v3.0
0 stars 6 forks source link

Update `snap_molecule_indices` or use function from cmeutils? #33

Open JimmyRushing opened 2 years ago

JimmyRushing commented 2 years ago

https://github.com/cmelab/morphct/blob/31ddf263979b7200b7015fa16eb4d6af3ea01fca/morphct/mobility_kmc.py#L502

I changed this code to the code below to deal with an indexing artifact of how the xml was converted to gsd . Is this worth updating in morph? Or is this redundant with this cmeutils PR https://github.com/cmelab/cmeutils/pull/20

    system = freud.AABBQuery.from_system(snap)
    bonds = snap.bonds.group
    bonds = bonds[np.lexsort((bonds[:,1],bonds[:,0]))]
    snap.bonds.group = bonds
    n_query_pts = n_pts = snap.bonds.N
    query_pt_inds = snap.bonds.group[:,0]
    pt_inds = snap.bonds.group[:, 1]
    distances = system.box.compute_distances(
        system.points[query_pt_inds], system.points[pt_inds]
    )
    nlist = freud.NeighborList.from_arrays(
        n_query_pts, n_pts, query_pt_inds, pt_inds, distances
    )
    cluster = freud.cluster.Cluster()
    cluster.compute(system=system, neighbors=nlist)
    return cluster.cluster_idx
jennyfothergill commented 2 years ago

If we switch to using snap_molecule_indices from cmeutils, then yes. Probably that should be done.