Estimating mean intrinsic fractional polarization (FPOL) and intrinsic variability (modulation index) using sequence of the observed FPOL values and their errors. Implementation of the approach described in Appendix A of Blinov et al. 2016 (doi:10.1093/mnras/stw158)
First, install boost, dlib and pybind11 C++ libraries.
Clone repo:
$ git clone https://github.com/ipashchenko/polinf.git
$ cd polinf
Compile:
$ c++ -O3 -Wall -shared -std=c++17 -fPIC -ldlib -march=native -DNDEBUG -O3 -fext-numeric-literals `python3 -m pybind11 --includes` -o pypolinf`python3-config --extension-suffix` pypolinf.cpp
This creates *.so shared library file (e.g. pypolinf.cpython-38-x86_64-linux-gnu.so) that can be used to import fit function from Python:
$ python -c "from pypolinf import fit"
Imported function fit takes two positional arguments - iterables of FPOL values and their uncertainties, and returns tuple of mean intrinsic FPOL and intrinsic modulation index.
>>> FPOL_mean, mod_indx = fit(FPOL, sigma_FPOL)
Example usage shown in example.py (first, compile shared library).