Ensure contiguous occupations

src/occupation.jl

@@ -30,6 +30,11 @@ end
 function compute_occupation(basis::PlaneWaveBasis{T}, eigenvalues::AbstractVector, εF::Number;
                             temperature=basis.model.temperature,
                             smearing=basis.model.smearing) where {T}
+    # Check that eigenvalues are increasing monotonically, for contiguous occupations
+    if !all(all(diff(εk) .≥ -eps(T)) for εk in eigenvalues)
+        error("Eigenvalues should be monotonically increasing.")
+    end
+
     # This is needed to get the right behaviour for special floating-point types
     # such as intervals.
     inverse_temperature = iszero(temperature) ? T(Inf) : 1/temperature
@@ -207,3 +212,18 @@ function check_full_occupation(basis::PlaneWaveBasis, occupation)
         all(occ_k .== filled_occ) || error("Only full occupation is supported, but $occ_k has partial occupation.")
     end
 end
+
+"""
+Return ranges of occupied elements based on a given occupation threshold
+"""
+function occupied_empty_masks(occupation, occupation_threshold)
+    n_occ = map(occupation) do occ
+        n = count(occ_i -> abs(occ_i) > occupation_threshold, occ)
+        # Check that all occupied elements are contiguous, otherwise range is wrong
+        @assert all(occ[1:n] .> occupation_threshold)
+        n
+    end
+    mask_occ  = [1:n_occ[ik] for ik in 1:length(occupation)]
+    mask_empty = [(n_occ[ik] + 1):length(occupation[ik]) for ik in 1:length(occupation)]
+    (; mask_occ, mask_empty)
+end

src/orbitals.jl

@@ -5,9 +5,9 @@ using Random  # Used to have a generic API for CPU and GPU computations alike: s
 # threshold is a parameter to distinguish between states we want to keep and the
 # others when using temperature. It is set to 0.0 by default, to treat with insulators.
 function select_occupied_orbitals(basis, ψ, occupation; threshold=0.0)
-    N = [something(findlast(x -> x > threshold, occk), 0) for occk in occupation]
-    selected_ψ   = [@view ψk[:, 1:N[ik]] for (ik, ψk)   in enumerate(ψ)]
-    selected_occ = [      occk[1:N[ik]]  for (ik, occk) in enumerate(occupation)]
+    mask_occ = occupied_empty_masks(occupation, threshold).mask_occ
+    selected_ψ   = [@view ψk[:, mask_occ[ik]] for (ik, ψk)   in enumerate(ψ)]
+    selected_occ = [      occk[mask_occ[ik]]  for (ik, occk) in enumerate(occupation)]
 
     # if we have an insulator, sanity check that the orbitals we kept are the
     # occupied ones

src/response/chi0.jl

@@ -409,9 +409,7 @@ to the Hamiltonian change `δH` represented by the matrix-vector products `δHψ
     # We then use the extra information we have from these additional bands,
     # non-necessarily converged, to split the Sternheimer_solver with a Schur
     # complement.
-    occ_thresh = occupation_threshold
-    mask_occ   = map(occk -> findall(occnk -> abs(occnk) ≥ occ_thresh, occk), occupation)
-    mask_extra = map(occk -> findall(occnk -> abs(occnk) < occ_thresh, occk), occupation)
+    (mask_occ, mask_extra) = occupied_empty_masks(occupation, occupation_threshold)
 
     ψ_occ   = [ψ[ik][:, maskk] for (ik, maskk) in enumerate(mask_occ)]
     ψ_extra = [ψ[ik][:, maskk] for (ik, maskk) in enumerate(mask_extra)]
@@ -563,7 +561,7 @@ function construct_bandtol(Bandtol::Type, basis::PlaneWaveBasis, ψ, occupation:
     Ω  = basis.model.unit_cell_volume
     Ng = prod(basis.fft_size)
     Nk = length(basis.kpoints)
-    mask_occ = map(ok -> findall(onk -> abs(onk) ≥ occupation_threshold, ok), occupation)
+    mask_occ = occupied_empty_masks(occupation, occupation_threshold).mask_occ
 
     # Including k-points the expression (3.11) in 2505.02319 becomes
     #   with Φk = (ψ_{1,k} … ψ_{n,k})_k  (Concatenation of all occupied orbitals for this k)

src/supercell.jl

@@ -112,6 +112,9 @@ function cell_to_supercell(scfres::NamedTuple)
     basis_supercell = cell_to_supercell(basis)
     ψ_supercell     = [cell_to_supercell(ψ, basis, basis_supercell)]
     eigs_supercell  = [reduce(vcat, scfres_unfold.eigenvalues)]
+    perms = [sortperm(eigs_supercell[ik]) for ik = 1:length(eigs_supercell)]
+    ψ_supercell     = [ψ_supercell[ik][:, perms[ik]] for ik = 1:length(ψ_supercell)]
+    eigs_supercell  = [eigs_supercell[ik][perms[ik]] for ik = 1:length(eigs_supercell)]
     occ_supercell   = compute_occupation(basis_supercell, eigs_supercell, scfres.εF).occupation
     ρ_supercell     = compute_density(basis_supercell, ψ_supercell, occ_supercell;
                                       scfres.occupation_threshold)