[ITensors] ITensorMPS submodule

ITensorGPU/test/runtests.jl

@@ -21,7 +21,6 @@ CUDA.allowscalar(false)
     "test_cucontract.jl",
     "test_cumpo.jl",
     "test_cumps.jl",
-    "test_cuiterativesolvers.jl",
     "test_cutruncate.jl",
     #"test_pastaq.jl",
   )

Project.toml

@@ -21,6 +21,7 @@ PackageCompiler = "9b87118b-4619-50d2-8e1e-99f35a4d4d9d"
 Pkg = "44cfe95a-1eb2-52ea-b672-e2afdf69b78f"
 Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+Reexport = "189a3867-3050-52da-a836-e630ba90ab69"
 Requires = "ae029012-a4dd-5104-9daa-d747884805df"
 SerializedElementArrays = "d3ce8812-9567-47e9-a7b5-65a6d70a3065"
 SimpleTraits = "699a6c99-e7fa-54fc-8d76-47d257e15c1d"

src/ITensorMPS/ITensorMPS.jl

@@ -0,0 +1,26 @@
+module ITensorMPS
+
+include("usings.jl")
+include("imports.jl")
+include("exports.jl")
+include("abstractmps.jl")
+include("mps.jl")
+include("mpo.jl")
+include("sweeps.jl")
+include("abstractprojmpo/abstractprojmpo.jl")
+include("abstractprojmpo/projmpo.jl")
+include("abstractprojmpo/diskprojmpo.jl")
+include("abstractprojmpo/projmposum.jl")
+include("abstractprojmpo/projmps.jl")
+include("abstractprojmpo/projmpo_mps.jl")
+include("observer.jl")
+include("dmrg.jl")
+include("adapt.jl")
+include("autompo/matelem.jl")
+include("autompo/qnmatelem.jl")
+include("autompo/opsum_to_mpo_generic.jl")
+include("autompo/opsum_to_mpo.jl")
+include("autompo/opsum_to_mpo_qn.jl")
+include("deprecated.jl")
+
+end # module ITensorMPS

src/mps/abstractmps.jl → src/ITensorMPS/abstractmps.jl

@@ -79,7 +79,7 @@ know if there is functionality not available for MPS/MPO you would like.
 """
 data(m::AbstractMPS) = m.data
 
-contract(ψ::AbstractMPS) = contract(data(ψ))
+ITensors.contract(ψ::AbstractMPS) = contract(data(ψ))
 
 leftlim(m::AbstractMPS) = m.llim
 
@@ -1076,7 +1076,7 @@ function deprecate_make_inds_match!(
     make_inds_match = false
   end
   if !hassameinds(siteinds, M1dag, M2) && make_inds_match
-    warn_once(inner_mps_mpo_mps_deprecation_warning(), :inner_mps_mps)
+    ITensors.warn_once(inner_mps_mpo_mps_deprecation_warning(), :inner_mps_mps)
     replace_siteinds!(M1dag, siteindsM2)
   end
   return M1dag, M2
@@ -1776,7 +1776,7 @@ function _fermionic_swap(s1::Index, s2::Index)
     if _isodd_fermionic_parity(s1, n1) && _isodd_fermionic_parity(s2, n2)
       dval = -1.0
     end
-    Tb = blockview(tensor(T), b)
+    Tb = ITensors.blockview(tensor(T), b)
     mat_dim = prod(dims(Tb)[1:2])
     Tbr = reshape(Tb, mat_dim, mat_dim)
     for i in diagind(Tbr)
@@ -1847,9 +1847,9 @@ function setindex!(
     sites = sites0[[perm...]]
     # Check if the site indices
     # are fermionic
-    if !using_auto_fermion() && any(anyfermionic, sites)
+    if !using_auto_fermion() && any(ITensors.anyfermionic, sites)
       if length(sites) == 2 && ψ isa MPS
-        if all(allfermionic, sites)
+        if all(ITensors.allfermionic, sites)
           s0 = Index.(sites0)
 
           # TODO: the Fermionic swap is could be diagonal,

src/mps/abstractprojmpo/abstractprojmpo.jl → src/ITensorMPS/abstractprojmpo/abstractprojmpo.jl

@@ -33,7 +33,7 @@ function rproj(P::AbstractProjMPO)::Union{ITensor,OneITensor}
   return P.LR[P.rpos]
 end
 
-function contract(P::AbstractProjMPO, v::ITensor)::ITensor
+function ITensors.contract(P::AbstractProjMPO, v::ITensor)::ITensor
   itensor_map = Union{ITensor,OneITensor}[lproj(P)]
   append!(itensor_map, P.H[site_range(P)])
   push!(itensor_map, rproj(P))

src/mps/abstractprojmpo/projmposum.jl → src/ITensorMPS/abstractprojmpo/projmposum.jl

@@ -44,7 +44,7 @@ shorthand for `product(P,v)`.
 """
 product(A::AbstractSum, v::ITensor) = sum(t -> product(t, v), terms(A))
 
-contract(A::AbstractSum, v::ITensor) = sum(t -> contract(t, v), terms(A))
+ITensors.contract(A::AbstractSum, v::ITensor) = sum(t -> contract(t, v), terms(A))
 
 """
     eltype(P::ProjMPOSum)

src/physics/autompo/opsum_to_mpo_generic.jl → src/ITensorMPS/autompo/opsum_to_mpo_generic.jl

@@ -68,7 +68,7 @@ subtract!(os::OpSum, o::Tuple) = add!(os, -Ops.op_term(o))
 function isfermionic(t::Vector{Op}, sites)
   p = +1
   for op in t
-    if has_fermion_string(name(op), sites[site(op)])
+    if has_fermion_string(ITensors.name(op), sites[site(op)])
       p *= -1
     end
   end
@@ -207,7 +207,7 @@ function sorteachterm(os::OpSum, sites)
     filter!(!iszero, perm)
     # and account for anti-commuting, fermionic operators
     # during above sort; put resulting sign into coef
-    t *= parity_sign(perm)
+    t *= ITensors.parity_sign(perm)
     terms(os)[j] = t
   end
 
@@ -299,8 +299,8 @@ function MPO(os::OpSum, sites::Vector{<:Index}; splitblocks=true, kwargs...)::MP
   return M
 end
 
-function MPO(eltype::Type{<:Number}, os::OpSum, sites::Vector{<:Index}; kwargs...)
-  return NDTensors.convert_scalartype(eltype, MPO(os, sites; kwargs...))
+function MPO(elt::Type{<:Number}, os::OpSum, sites::Vector{<:Index}; kwargs...)
+  return NDTensors.convert_scalartype(elt, MPO(os, sites; kwargs...))
 end
 
 # Conversion from other formats

src/physics/autompo/opsum_to_mpo_qn.jl → src/ITensorMPS/autompo/opsum_to_mpo_qn.jl

@@ -99,7 +99,7 @@ function qn_svdMPO(
   llinks = Vector{QNIndex}(undef, N + 1)
   # Set dir=In for fermionic ordering, avoid arrow sign
   # <fermions>:
-  linkdir = using_auto_fermion() ? In : Out
+  linkdir = using_auto_fermion() ? ITensors.In : ITensors.Out
   llinks[1] = Index([QN() => 1, Hflux => 1]; tags="Link,l=0", dir=linkdir)
   for n in 1:N
     qi = Vector{Pair{QN,Int}}()
@@ -220,7 +220,7 @@ function qn_svdMPO(
         end
 
         b = loc(rq, cq)
-        T = BlockSparseTensor(ValType, [b], (dag(ll), rl))
+        T = ITensors.NDTensors.BlockSparseTensor(ValType, [b], (dag(ll), rl))
         T[b] .= M
 
         H[n] += (itensor(T) * Op)
@@ -230,13 +230,13 @@ function qn_svdMPO(
     # Put in ending identity operator
     Id = op("Id", sites[n])
     b = Block(1, 1)
-    T = BlockSparseTensor(ValType, [b], (dag(ll), rl))
+    T = ITensors.NDTensors.BlockSparseTensor(ValType, [b], (dag(ll), rl))
     T[b] = 1
     H[n] += (itensor(T) * Id)
 
     # Put in starting identity operator
     b = Block(nblocks(ll), nblocks(rl))
-    T = BlockSparseTensor(ValType, [b], (dag(ll), rl))
+    T = ITensors.NDTensors.BlockSparseTensor(ValType, [b], (dag(ll), rl))
     T[b] = 1
     H[n] += (itensor(T) * Id)
   end # for n in 1:N

src/mps/deprecated.jl → src/ITensorMPS/deprecated.jl

@@ -1,3 +1,37 @@
+# mps/abstractmps.jl
+@deprecate orthoCenter(args...; kwargs...) orthocenter(args...; kwargs...)
+@deprecate store(m::AbstractMPS) data(m)
+@deprecate replacesites!(args...; kwargs...) ITensors.replace_siteinds!(args...; kwargs...)
+@deprecate applyMPO(args...; kwargs...) contract(args...; kwargs...)
+@deprecate applympo(args...; kwargs...) contract(args...; kwargs...)
+@deprecate errorMPOprod(args...; kwargs...) error_contract(args...; kwargs...)
+@deprecate error_mpoprod(args...; kwargs...) error_contract(args...; kwargs...)
+@deprecate error_mul(args...; kwargs...) error_contract(args...; kwargs...)
+@deprecate multMPO(args...; kwargs...) contract(args...; kwargs...)
+@deprecate sum(A::AbstractMPS, B::AbstractMPS; kwargs...) add(A, B; kwargs...)
+@deprecate multmpo(args...; kwargs...) contract(args...; kwargs...)
+@deprecate set_leftlim!(args...; kwargs...) ITensors.setleftlim!(args...; kwargs...)
+@deprecate set_rightlim!(args...; kwargs...) ITensors.setrightlim!(args...; kwargs...)
+@deprecate tensors(args...; kwargs...) ITensors.data(args...; kwargs...)
+@deprecate primelinks!(args...; kwargs...) ITensors.prime_linkinds!(args...; kwargs...)
+@deprecate simlinks!(args...; kwargs...) ITensors.sim_linkinds!(args...; kwargs...)
+@deprecate mul(A::AbstractMPS, B::AbstractMPS; kwargs...) contract(A, B; kwargs...)
+
+# mps/mpo.jl
+@deprecate MPO(A::MPS; kwargs...) outer(A', A; kwargs...)
+
+# mps/mps.jl
+@deprecate randomMPS(sites::Vector{<:Index}, linkdims::Integer) randomMPS(
+  sites; linkdims=linkdims
+)
+@deprecate randomMPS(ElType::Type, sites::Vector{<:Index}, linkdims::Integer) randomMPS(
+  ElType, sites; linkdims=linkdims
+)
+@deprecate randomMPS(sites::Vector{<:Index}, state, linkdims::Integer) randomMPS(
+  sites, state; linkdims=linkdims
+)
+
+# Pass throughs of old name to new name:
 
 unique_siteind(A::AbstractMPS, B::AbstractMPS, j::Integer) = siteinds(uniqueind, A, B, j)
 

src/mps/dmrg.jl → src/ITensorMPS/dmrg.jl

@@ -1,7 +1,4 @@
-
-#function IndexSet_ignore_missing(is::Union{Index,Nothing}...)
-#  return IndexSet(filter(i -> i isa Index, is))
-#end
+using KrylovKit: eigsolve
 
 function permute(
   M::AbstractMPS, ::Tuple{typeof(linkind),typeof(siteinds),typeof(linkind)}

src/ITensorMPS/exports.jl

@@ -0,0 +1,128 @@
+export
+  # dmrg.jl
+  dmrg,
+
+  # abstractmps.jl
+  # Macros
+  @preserve_ortho,
+  # Methods
+  AbstractMPS,
+  add,
+  common_siteind,
+  common_siteinds,
+  findfirstsiteind,
+  findfirstsiteinds,
+  findsite,
+  findsites,
+  firstsiteind,
+  firstsiteinds,
+  logdot,
+  loginner,
+  lognorm,
+  movesite,
+  movesites,
+  normalize,
+  ortho_lims,
+  orthocenter,
+  promote_itensor_eltype,
+  reset_ortho_lims!,
+  set_ortho_lims!,
+  siteinds,
+  sim!,
+
+  # autompo/
+  AutoMPO,
+  add!,
+
+  # mpo.jl
+  # Types
+  MPO,
+  # Methods
+  error_contract,
+  maxlinkdim,
+  orthogonalize,
+  orthogonalize!,
+  outer,
+  projector,
+  randomMPO,
+  truncate,
+  truncate!,
+  unique_siteind,
+  unique_siteinds,
+
+  # mps.jl
+  # Types
+  MPS,
+  # Methods
+  ⋅,
+  correlation_matrix,
+  expect,
+  inner,
+  isortho,
+  linkdim,
+  linkdims,
+  linkind,
+  linkinds,
+  productMPS,
+  randomMPS,
+  replacebond,
+  replacebond!,
+  sample,
+  sample!,
+  siteind,
+  siteinds,
+  replace_siteinds!,
+  replace_siteinds,
+  swapbondsites,
+  totalqn,
+
+  # observer.jl
+  # Types
+  AbstractObserver,
+  DMRGObserver,
+  DMRGMeasurement,
+  NoObserver,
+  # Methods
+  checkdone!,
+  energies,
+  measure!,
+  measurements,
+  truncerrors,
+
+  # projmpo.jl
+  checkflux,
+  disk,
+  ProjMPO,
+  lproj,
+  product,
+  rproj,
+  noiseterm,
+  nsite,
+  position!,
+
+  # projmposum.jl
+  ProjMPOSum,
+
+  # projmpo_mps.jl
+  ProjMPO_MPS,
+
+  # sweeps.jl
+  Sweeps,
+  cutoff,
+  cutoff!,
+  get_cutoffs,
+  get_maxdims,
+  get_mindims,
+  get_noises,
+  maxdim,
+  maxdim!,
+  mindim,
+  mindim!,
+  noise,
+  noise!,
+  nsweep,
+  setmaxdim!,
+  setmindim!,
+  setcutoff!,
+  setnoise!,
+  sweepnext