A Julia port and restructuring of the neighbourlist implemented in matscipy (with the authors' permission). Single-threaded, the Julia version is faster than matscipy for small systems, probably due to the overhead of dealing with Python, but on large systems it is tends to be slower (up to around a factor 2 for 100,000 atoms).
The package is can be used stand-alone, with JuLIP.jl, or with AtomsBase.jl.
Pkg.add("NeighbourLists")
using NeighbourLists
?PairList
using ASEconvert, NeighbourLists, Unitful
cu = ase.build.bulk("Cu") * pytuple((4, 2, 3))
sys = pyconvert(AbstractSystem, cu)
nlist = PairList(sys, 3.5u"Å")
neigs_1, Rs_1 = neigs(nlist, 1)
using JuLIP
at = bulk(:Cu) * (4, 2, 3)
nlist = neighbourlist(at, 3.5)
neigs_1, Rs_1 = neigs(nlist, 1)
Please also look at the tests on how to use this package. Or just open an issue and ask.