FPFS: Fourier Power Function Shaplets (A fast, accurate 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.

Installation

For stable (old) version, which have not been updated:

pip install fpfs

Or clone the repository:

git clone https://github.com/mr-superonion/FPFS.git
cd FPFS
pip install -e . --user

Before using the code, please setup the jax environment

source fpfs_config

Reference

The 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.

Development

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