Core

setup_param(C::Dict{String,Any}, reset::Bool = true)

Setup the configuration dictionaries via function call. Here C contains parameters for general setup. If reset is true, then the configuration dictionaries will be reset to their default values at first. Later, C will be used to customized the dictionaries further.

Arguments

See above explanations.

Returns

N/A

See also: read_param.

read_param()

Setup the configuration dictionaries via an external file. The valid format of a configuration file is toml.

Arguments

N/A

Returns

N/A

See also: setup_param.

make_data_std()

Try to build a standard dataset (ACT100), which contains 100 typical spectral functions and the corresponding correlation functions. These spectral functions are defined in src/dataset.jl. Please use the script util/acstd.jl to do this job.

Arguments

N/A

Returns

N/A

make_data()

Try to generate spectral functions and the corresponding Green's functions.

Arguments

N/A

Returns

N/A

make_peak(rng::AbstractRNG)

Generate peak to build the final spectral function.

Arguments

• rng -> Random number generator.

Returns

• 𝑝 -> A Peak struct (subtype of AbstractPeak).

See also: AbstractPeak.

make_spectrum(rng::AbstractRNG, mesh::AbstractMesh)

Generate a spectral function randomly at given mesh.

Arguments

• rng -> Random number generator.
• mesh -> Real frequency mesh, ω.

Returns

• sf -> A SpectralFunction struct.

See also: make_data.

make_spectrum(
    mesh::AbstractMesh,
    pv::Vector{<:AbstractPeak},
    sv::Vector{F64}
)

Generate a spectral function with given peaks and signs at given mesh.

Arguments

• mesh -> Real frequency mesh, ω.
• pv -> List of peaks.
• sv -> List of signs.

Returns

• sf -> A SpectralFunction struct.

See also: make_data_std.

make_green(
    rng::AbstractRNG,
    sf::SpectralFunction,
    kernel::Matrix{F64},
    grid::AbstractGrid
)

For given spectral function A and kernel matrix K, try to generate the corresponding correlation function G (note that G ≡ KA).

Arguments

• rng -> Random number generator.
• sf -> A SpectralFunction struct, A(ω).
• kernel -> Kernel matrix.
• grid -> Grid for correlation function.

Returns

• gf -> A GreenFunction struct.

make_grid(_grid = nothing)

To generate imaginary time grid or Masubara grid for many-body correlator. It will return a sub-type of the AbstractGrid struct.

Arguments

• _grid -> If it is present, we should use it to setup grid.

Returns

• grid -> Imaginary time or imaginary frequency grid.

See also: AbstractGrid.

make_mesh(_ktype = nothing, _mesh = nothing)

Try to generate an uniform (linear) or non-uniform (non-linear) mesh for the spectral function in real axis.

Arguments

• _ktype -> If it is present, we should use it to setup ktype.
• _mesh -> If it is present, we should use it to setup mesh.

Returns

• mesh -> Real frequency mesh. It should be a subtype of AbstractMesh.

See also: LinearMesh, TangentMesh, LorentzMesh.

make_kernel(am::AbstractMesh, ag::AbstractGrid)

Try to generate various kernel functions.

Arguments

• am -> Real frequency mesh.
• ag -> Imaginary axis grid.

Returns

• kernel -> Kernel function, a 2D array, (ntime,nmesh) or (nfreq,nmesh).

See also: AbstractMesh, AbstractGrid.

reprod(am::AbstractMesh, kernel::Matrix{F64}, A::Vector{F64})

Try to reproduce the input data, which can be compared with the raw data to see whether the analytic continuation is reasonable.

Arguments

• am -> Real frequency mesh.
• kernel -> The kernel function.
• A -> The calculated spectral function, A(ω).

Returns

• G -> Reconstructed correlators, G(τ) or G(iωₙ), Vector{F64}.

See also: AbstractMesh.