save

Project.toml

@@ -4,6 +4,7 @@ authors = ["Zhongyi Ni"]
 version = "1.0.0-DEV"
 
 [deps]
+Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
 LuxorGraphPlot = "1f49bdf2-22a7-4bc4-978b-948dc219fbbc"
 TensorInference = "c2297e78-99bd-40ad-871d-f50e56b81012"

src/TensorQEC.jl

@@ -24,10 +24,11 @@ export PauliString, PauliGroup, isanticommute
 export syndrome_inference, measure_syndrome!
 
 # encoder
-export Bimatrix, toric_code, stabilizers, syndrome_transform
+export Bimatrix, toric_code, stabilizers, syndrome_transform, encode_stabilizers
 export ToricCode, SurfaceCode
 
 #error correction
+export correction_pauli_string
 
 
 

src/encoder.jl

@@ -155,3 +155,11 @@ function stabilizers(::SurfaceCode{3})
     return pauli_string
 end
 
+
+function encode_stabilizers(stabilizers::AbstractVector{PauliString{N}})where N
+	bimat = stabilizers2bimatrix(stabilizers)
+    gaussian_elimination!(bimat)
+    qc = encode_circuit(bimat)
+	data_qubits = bimat.ordering[size(bimat.matrix,1)+1:end]
+	return qc, data_qubits, bimat
+end

src/errorcorrect.jl

@@ -1,13 +1,14 @@
 function correction_pauli_string(qubit_num::Int, syn::Dict{Int,Bool}, prob::Dict{Int,Vector{Float64}})
     ps = ones(Int, qubit_num)
     for (k, v) in prob
+        @show syn[k]
         if syn[k] #syn[k] is true, measure outcome is -1, use X or Y
-            ps[k]=(findmax(v)[2]) == 1 ? 2 : 3
+            ps[k] = (findmax(v)[2]) == 1 ? 2 : 3
         elseif findmax(v)[2] == 2
-            ps[k]=2
+            ps[k] = 4
         end
     end
-    return PauliString(ps...)
+    return PauliString(ps[1:end]...)
 end
 
 

src/inferences.jl

@@ -14,13 +14,11 @@ end
 function syndrome_inference(cl::CliffordNetwork{T}, syn::Dict{Int,Bool}, p::Vector{Vector{Float64}})where T
 	n = length(p)
 	ps = Dict([i=>BoundarySpec((p[i]...,), false) for i in 1:n])
-	qs = Dict([i=>BoundarySpec((ones(T,4)...,), false) for i in 1:n])
+	qs = Dict([i=>BoundarySpec((ones(T,4)...,), true) for i in 1:n])
 	for (k, v) in syn
-		@show k,v
 		qs[k] = BoundarySpec((v ?  (0.0,1.0,1.0,0.0) : (1.0,0.0,0.0,1.0) ),true)
 	end
-	#@show qs
 	tn = generate_tensor_network(cl, ps, qs)
-	mp=marginals(tn)
+	mp = marginals(tn)
 	return result = Dict([k => mp[[cl.mapped_qubits[k]]] for k in keys(syn)])
 end

src/paulibasis.jl

@@ -8,7 +8,7 @@
 # 4. pauli_basis(2)[5]
 # 5. LinearAlgebra.kron(X,I2)
 # 6. IX shown in the print
-
+# 7. Apply I on the first qubit and X on the second qubit. This will convert the state |00> to |10>.
 
 function pauli_basis(nqubits::Int)
 	paulis = [I2, X, Y, Z]
@@ -23,15 +23,15 @@ function pauli_decomposition(m::AbstractMatrix)
 end
 
 # defined the linear mapping in the pauli basis
-# from the coding space to the physical qubits, i.e. (physical..., coding...)
+# from the coding space to the physical qubits, i.e. (physical..., coding...) or (output..., input...)
 function pauli_mapping(m::AbstractMatrix)
 	nqubits = Int(log2(size(m, 1)))
 	paulis = pauli_basis(nqubits)
 	return [real(tr(pi * m * pj * m')/size(m, 1)) for pi in paulis, pj in paulis]
 end
 
 function pauli_string_map(ps::PauliString{N}, paulimapping::Array, qubits::Vector{Int}) where N
-    c=findall(!iszero, paulimapping[ps.ids[qubits]...,:,:])[1]
+    c=findall(!iszero, paulimapping[fill(:,length(size(paulimapping)) ÷ 2)...,ps.ids[qubits]...])[1]
     return PauliString(([k ∈ qubits ? c[findfirst(==(k),qubits)] : ps.ids[k] for k in 1:N]...,))
 end
 

test/encoder.jl

@@ -100,4 +100,15 @@ end
 	for i in 1:size(code.matrix, 1)
 		@test u * mat(ComplexF64, put(9, code.ordering[i] => Z)) * u' ≈ mat(ComplexF64, stabilizers[i])
 	end
+end
+
+
+@testset "encode_stabilizers" begin
+	result = TensorQEC.stabilizers(SurfaceCode{3}())
+	qc, data_qubits, bimat = TensorQEC.encode_stabilizers(result)
+	u = mat(ComplexF64, qc)
+	stabilizers = TensorQEC.bimatrix2stabilizers(bimat)
+	for i in 1:size(bimat.matrix, 1)
+		@test u * mat(ComplexF64, put(9, bimat.ordering[i] => Z)) * u' ≈ mat(ComplexF64, stabilizers[i])
+	end
 end

test/errorcrrect.jl

@@ -1,2 +1,7 @@
 using Test, TensorQEC, TensorQEC.Yao
+using Distributions
 
+function random_pauli_string(N::Int,p::Vector)
+    values = [1, 2, 3, 4] # 1: I, 2: X, 3: Y, 4: Z
+    return PauliString(([values[rand(Categorical(p[i]))] for i in 1:N]...,)) 
+end

test/inferences.jl

@@ -3,7 +3,6 @@ using Test, TensorQEC, TensorQEC.Yao
     result=stabilizers(SurfaceCode{3}())
     code = TensorQEC.stabilizers2bimatrix(result)
     TensorQEC.gaussian_elimination!(code)
-    st = TensorQEC.bimatrix2stabilizers(code)
     qc = TensorQEC.encode_circuit(code)
     #display(vizcircuit(qc))
 
@@ -42,27 +41,65 @@ end
     @test length(pinf) == 2
 end
 
+# simple example: Z1Z2 Z2Z3 stabilizers on 3 qubits
+@testset "simple_example_inference" begin
+    stabilizers = [PauliString((1,4,4)),PauliString((4,4,1))]
+    qc, data_qubits, code = TensorQEC.encode_stabilizers(stabilizers)
+    reg = join(rand_state(1), zero_state(2))
+    apply!(reg, put(3, 3=>X))
+    @show reg.state
+    regcopy = copy(reg)
+    @show qc 
+    @show data_qubits
+    @show code.matrix
+
+    apply!(reg, qc)
+    @show reg.state
+    #apply!(reg, put(3, 2=>Y))
+    apply!(reg, put(3, 3=>X))
+    measure_outcome=measure_syndrome!(reg, stabilizers)
+    @show measure_outcome
+    syn=syndrome_transform(code, measure_outcome)
+    syn_dict=TensorQEC.generate_syndrome_dict(code,syn)
+
+    cl = clifford_network(qc)
+    p = fill([0.85,0.05,0.05,0.05],3)
+    pinf = syndrome_inference(cl, syn_dict, p)
+    @show pinf
+    ps_ec = correction_pauli_string(3, syn_dict, pinf)
+    ps_ec_phy = TensorQEC.pauli_string_map_iter(ps_ec, qc)
+    @show ps_ec,ps_ec_phy
+    @show reg.state
+    apply!(reg, Yao.YaoBlocks.Optimise.to_basictypes(ps_ec_phy))
+    @show reg.state
+    @test measure_syndrome!(reg, stabilizers) == [1,1]
+    @test fidelity(density_matrix(reg, [3]), density_matrix(regcopy, [3])) ≈ 1.0
+end
+
 @testset "syndrome inference for SurfaceCode3" begin
-    result=stabilizers(SurfaceCode{3}())
-    code = TensorQEC.stabilizers2bimatrix(result)
-    TensorQEC.gaussian_elimination!(code)
-    st = TensorQEC.bimatrix2stabilizers(code)
-    qc = TensorQEC.encode_circuit(code)
+    result = stabilizers(SurfaceCode{3}())
+    qc, data_qubits, code = TensorQEC.encode_stabilizers(result)
 
-    reg = join(rand_state(1), zero_state(8))
-    regcopy=copy(reg)
+    reg = join(zero_state(8),rand_state(1))
+    regcopy = copy(reg)
     apply!(reg, qc)
-    apply!(reg, put(9, 9=>X))
+    apply!(reg, put(9, 5=>Y))
+    apply!(reg, put(9, 3=>Z))
     measure_outcome=measure_syndrome!(reg, result)
+    @show measure_outcome
     syn=syndrome_transform(code, measure_outcome)
     syn_dict=TensorQEC.generate_syndrome_dict(code,syn)
 
     cl = clifford_network(qc)
-    p=fill([0.85,0.05,0.05,0.05],9)
-    pinf=syndrome_inference(cl, syn_dict, p)
-    ps_ec=correction_pauli_string(9, syn_dict, pinf)
-    ps_ec_phy=TensorQEC.pauli_string_map_iter(ps_ec, qc)
+    p = fill([0.94,0.02,0.02,0.02],9)
+    pinf = syndrome_inference(cl, syn_dict, p)
+    @show pinf
+    ps_ec = correction_pauli_string(9, syn_dict, pinf)
+    ps_ec_phy = TensorQEC.pauli_string_map_iter(ps_ec, qc)
+    @show 
     apply!(reg, Yao.YaoBlocks.Optimise.to_basictypes(ps_ec_phy))
-    @test measure_syndrome!(reg, result)==[1,1,1,1,1,1,1,1]
-
+    @test measure_syndrome!(reg, result) == [1,1,1,1,1,1,1,1]
+    apply!(reg, qc')
+    #fidelity
+    @test fidelity(density_matrix(reg, [9]),density_matrix(regcopy, [9])) ≈ 1.0
 end

test/paulibasis.jl

@@ -17,12 +17,16 @@ end
 end
 
 @testset "pauli_string_map" begin
-	ps = PauliString(1,3,4,2,1)
-	@test TensorQEC.pauli_string_map(ps, pauli_mapping(mat(ComplexF64,cnot(2,1,2))), [4,5]).ids ==(1,3,4,2,2)
+	ps = PauliString(2,3,4,3,2,1)
+	@test TensorQEC.pauli_string_map(ps, pauli_mapping(mat(ComplexF64,cnot(2,1,2))), [5,6]).ids ==(2,3,4,3,2,2)
 end
 
 @testset "pauli_string_map_iter" begin
 	ps = PauliString(1,3,4,2,1)
 	qc = chain(cnot(5,4,5), put(5, 3=>H))
 	@test TensorQEC.pauli_string_map_iter(ps, qc).ids ==(1,3,2,2,2)
+
+	ps = PauliString(1,2,2)
+	qc = chain(cnot(3,3,2), cnot(3,3,1))
+	@test TensorQEC.pauli_string_map_iter(ps, qc).ids == (2,1,2)
 end