C++17 port of the MicroscPSF-Matlab project

Status

Currently, this work is a very early work in progress. Refer to https://github.com/MicroscPSF/MicroscPSF-Matlab for the reference implementation.

Quick start on Ubuntu/Linux

Install the compiler toolchain and the BLAS/LAPACK library:

sudo apt install build-essentials

# Use Openblas or Atlas or the original BLAS/LAPACK.
sudo apt install libopenblas-dev

Install Meson the build system:

cd MicroscPSF-Cpp/
python3 -m venv .venv/
.venv/bin/pip3 install meson ninja

Compile everything

cd MicroscPSF-Cpp/
meson setup build/
ninja -C build all

Test everything

cd MicroscPSF-Cpp/build/
ninja test

(Optional) Install the example app

ninja install
cd MicroscPSF-Cpp/build/
meson configure -Dinstall_examples=true
ninja all
sudo ninja install

Other operating system support

Windows

The C++ library only offers experimental support on MSVC/MinGW toolchain. Please refer to the build instruction file for details: https://github.com/MicroscPSF/MicroscPSF-Cpp/blob/main/.github/workflows/validate-build.yml

MacOSX

Mac provides a nice Accelerate LAPACK interface in the Apple Framework. The build system is capable of automatically detecting the framework as build depdendency, so the library should be able to compile and run PSF simulation natively.

The project is looking for volunteers and incoming PRs to enable Mac support.

Wishlist:

• Enable Bessel function support from 3rd party library #14 .

• Modify meson.build to detect the Accelerate LAPACK interface in Apple Framework.

• Document the build options, i.e. meson setup -Duse_boost=true -Darmadillo-code:lapack=framework build-for-macosx

Running the examples

First, follow the Quick start instructions to build the example app. Then, given the following microscope configurations in the screenshot...

... program the example C++ file at examples/generate-psf.cpp

microscope_params_t params{};
params.NA = 1.4;
params.ti0 = 150.0_um;
params.ni = 1.5;
params.ni0 = 1.5;
params.pz = 2.0_um;

precision_li2017_t precision{};
precision.num_basis = 153;
precision.rho_samples = 1000;

const auto psf =
    makePSF(params, {0.1_um, 0.25_um}, {256, 128}, 0.610_um, precision);

Next, repeat the Quick start steps to re-compile the C++ app. Run ninja test to invoke the compiled example. Locate the outputs psf_xy.pgm, psf_xz.pgm, and psf.h5. The XZ cross-section of the C++-simulated PSF should match the screenshot above.

Appendix: Bessel function support

• The original ISO C++ proposal: https://wg21.link/p0226r1
• GCC compiler support status: https://gcc.gnu.org/onlinedocs/libstdc++/manual/status.html#status.iso.2017
• Clang support status: https://libcxx.llvm.org/Status/Cxx17.html
• Clang support ticket: llvm/llvm-project#99939