sparse-ir (https://github.com/SpM-lab/sparse-ir) is a Python library for the intermediate representation of propagators.
With the excellent PyCall package of julia, one can use the
many features of the sparse_ir library from within a Julia session.
To use this package, both Python and a proper version of sparse-ir library must be
installed on your system.
If PyCall is installed using Conda
(which is the default behavior if no system python is found), then the
underlying sparse-ir library will be installed automatically via Conda when the
package is first loaded.
An optional library xprec, which allows to compute the IR basis functions with greater accuracy, is not installed automatically.
If needed, xprec must be installed manually:
using Pkg
Pkg.add("Conda") # if needed
using Conda
Conda.add("xprec", channel="h.shinaoka")
As of now (Feb. 15 2022), binary packages of xprec are not available on aarch64.
The underlying Python libraries can be updated as
using Pkg
Pkg.add("Conda") # if needed
using Conda
Conda.update()
If PyCall is not installed using Conda, installing both Python and the underlying libraries can be done by other means.
using SparseIR
beta = 10.0
wmax = 1.0
eps = 1e-7
basis_f = FiniteTempBasis(fermion, beta, wmax, eps)
basis_b = FiniteTempBasis(boson, beta, wmax, eps)A more detailed tutorial and sample codes are available at