Python and C++ split step method simulation library for nonlinear optical processes and Gaussian quantum optics.
The code currently supports ꭓ⁽²⁾ and ꭓ⁽³⁾ processes, with 1-dimensional propagation. This includes dispersion/diffraction and nonlinear effects, for arbitrary pump/signal shapes. It is straightforward to add differential equations with an arbitrary number of signal modes.
The nonlinearmedium
module contains the classes for simulating optical propagation based on given material and optical-field parameters.
The simulations solve the dimensionless propagation equations.
Green's functions can be calculated with linear equations.
nonlinearmedium
is a compiled C++ library meant to be imported and used in Python.
To compile, it requires Pybind11 for Python binding and Eigen for vectorized operations.
It will also compile with fftw if it is found.
The program is written using the curiously recurring template pattern (CRTP) to efficiently implement as many differential equation solvers as one can dream of.
The solvers are implemented and described in the solver/
directory, and registered in src/nlmModulePy.cpp
.
This repository also contains a collection of Jupyter notebooks that test for correct behavior or reproduce some published results.
These are found in the tests/
directory, and may serve as useful examples.
They are saved with Jupytext.
The classical
, poling
and multimode
Python modules provide functions for configuring simulations and analysis routines, for example:
- Converting to dimensionless quantities,
- Generating poling patterns,
- Calculating covariance matrices and related quantities.
The decompositions
module is borrowed from Strawberry Fields with minor improvements and provides good implementations of matrix decompositions.
For example, Bloch-Messiah, which is useful in Gaussian quantum optics.
The materials
library is built using Symbolic Python (Symengine or SymPy), and can evaluate Sellmeier equations as a function of wavelength and temperature.
This allows easy calculation of the index, group index, group velocity dispersion, or wavenumber β coefficients up to 3rd order
(ie k = β₀ + β₁ Δω + β₂ Δω² + β₃ Δω³ …).
It is straightforward to add materials: only a Sellmeier equation is required, Python decorators take care of the rest.
To compile nonlinearmedium
, create a build directory and, from inside, run cmake .. -DCMAKE_BUILD_TYPE=Release
and make -jX
(replacing X
with some number of compilation cores, eg -j4
).
The binary will be created in the main directory, and can be imported like a regular Python module.