Fiber LP modes

Compute LP modes from GRIN fibers. The modes are computed using analytical formulas, under the weak guidance assumption.

Installation

Use in virtual environment or you own environment:

$ python -m pip install -r requirements.txt
$ pip install -e .

Main features

• Fiber: GRIN fiber with parameters to compute theoretical LP modes effective refractive index.
• Grid: 2D grid supporting LP modes 2D scalar fields computation.
• Speckle: Sum of GRIN fiber supported LP modes with complex random weights (energy is conserved), composition or decomposition is available.
• Beams: Beam definition on a 2D grid support (Gaussian, Bessel, Bessel with Gaussian envelope).
• Devices: Essentially supporting macropixel phase-shifts and support constraints to the input field. Aims at mimicking an ideal deformable mirror.
• Coupling: Couples an optical field to the GRIN fiber using LP modes decomposition. Scrambles the mode through a LP modes coupling matrix.
• Datasets: Generates datasets on support 2D grid for later usage: Pure GRIN LP modes, GRIN speckles (limited to N modes), or Propagated GRIN speckles (using defined mode coupling matrix).
• Transforms: Fourier and Fresnel transforms.

Examples

• See ./notebooks/tutorial.ipynb for basic library usage.
• See ./notebooks/generate_dataset.ipynb for basic dataset generation.

Exported dataset contents

The datasets are exported as matfiles with version 6 (no compression, limited to 4Gb), pickled numpy files, or HDF5 files. The data is saved as a dictionnary with the following fields:

• phase_maps: Phase maps used to generate the corresponding fiber-output optical field.
• intens: Intensity of the fiber-output optical field (square modulus).
• degenerated_modes: Boolean indicating if the modes decomposition has been carried on fixed degenerated modes orientations.
• coupling_matrix: Fiber modes-coefficients coupling matrix, that has been used to simulated modes propagation in the fiber.
• transfer_matrix: Transfer matrix in image shape. Has dimensions Nact x Nx x Ny.
• reshaped_transfer_matrix: Reshaped transfer matrix for simple matrix products. Has dimensions Nact x (Nx x Ny).
• length: Dataset length.
• N_modes: Number of non-degerated LP modes allowed to propagate in the simulated fiber.
• macropixels_energy: Energy E on macropixels for the selected deformable mirror partitionning scheme. Use sqrt(E) weights on phase_maps to replicate output field.
• intens_transf: Optional field. Intensity (square modulus) of the transform (Fresnel or Fourier) of the fiber-output optical field.
• fields: Optional field. Fiber-output optical fields. Returned if return_output_fields is set to True.
• input_fields: Optional field. Fiber-input optical fields. Returned if return_input_fields is set to True.

Reading data saved as HDF5 file

The data saved as a HDF5 file can easily be read as a dictionnary and converted to Numpy arrays using the following code snippet:

import h5py

mdict = {}

with h5py.File('dataset.hdf5', 'r') as hf:
    for key_name in hf.keys():
        mdict[key_name] = hf[key_name][()]

It can also equivalently be read from MATLAB using the following code snippet:

dset = struct([]);
info = h5info('dataset.hdf5');

for i = 1:numel(info.Datasets)
    if strcmpi(info.Datasets(i).Dataspace.Type, 'scalar')
        dset(1).(info.Datasets(i).Name) = h5read(dset_path, "/" + info.Datasets(i).Name);
    elseif strcmpi(info.Datasets(i).Dataspace.Type, 'simple')
        read_dset = h5read(dset_path, "/" + info.Datasets(i).Name);
        if isstruct(read_dset)
            dset(1).(info.Datasets(i).Name) = read_dset.r + 1i * read_dset.i;
        else
            dset(1).(info.Datasets(i).Name) = read_dset;
        end
    else
        error("Not Implemented Dataspace type.")
    end
end

To-Do

• Refactor architecture: Speckle is a sum of scalar fields, Beam is a scalar field, MockDeformableMirror is a particular scalar field.
• Profile code and speed-up time consuming functions (particularly for heavy dataset generation).
• Add support for Step-Index fibers.

Notes

• Tested with Python 3.9.
• Matlab files are outdated and not supported anymore. They will slowly get deleted once their features are implemented on Python.