This repo is no longer maintained. Please see AnaCal for analytical shear estimator
Fourier Power Function Shapelets (FPFS) is an innovative estimator for the
shear responses of galaxy shape, flux, and detection. Utilizing leading-order
perturbations of shear (a vector perturbation) and image noise (a tensor
perturbation), FPFS determines shear and noise responses for both
measurements and detections. Unlike traditional methods that distort each
observed galaxy repeatedly, FPFS employs analytical shear responses of select
basis functions, including Shapelets basis and peak basis. Remarkably
efficient, FPFS can process approximately 1,000 galaxies within a single CPU
second. Testing under simple simulations has proven its capability to maintain
a multiplicative shear estimation bias below 0.5%, even amidst blending
challenges. For further details, refer to the FPFS module documentation
here.
For stable (old) version, which have not been updated:
pip install fpfsOr clone the repository:
git clone https://github.com/mr-superonion/FPFS.git
cd FPFS
pip install -e . --userBefore using the code, please setup the jax environment
source fpfs_configThe following papers are ready to be cited if you find any of these papers interesting or use the pipeline. Comments are welcome.
-
version 3: Li & Mandelbaum (2022) correct for detection bias from pixel level by interpreting smoothed pixel values as a projection of signal onto a set of basis functions.
-
version 2: Li , Li & Massey (2022) derive the covariance matrix of FPFS measurements and corrects for noise bias to second-order. In addition, it derives the correction for selection bias.
-
version 1: Li et. al (2018) build up the FPFS formalism based on Fourier_Quad and polar shapelets.
Before sending pull request, please make sure that the modified code passed the pytest and flake8 tests. Run the following commands under the root directory for the tests:
flake8
pytest -vv