furciferi / deswl

A collection of scripts and software related to DES weak lensing analysis.

This repository serves several purposes.

1. First, the issues page on GitHub can serve as a general clearing house for any issue related to the WL pipeline, tests, the catalogs, etc. If you have any issue that you would like to bring to the attention of the WL pipeline testing subgroup, you may post it here.

   We may repost the issue in a different location if there is a more appropriate place, or we may deal with it here. But as a user, you do not need to worry about finding the right place to post any problem you are having. Posting it here is perfectly fine.

   If there is any sensitive information or plots that you need to reference, we ask that you post them to redmine wiki

   https://cdcvs.fnal.gov/redmine/projects/deswlwg/wiki/Shear_Pipeline_Development_and_Testing

   and link to them from here. But since none of us find the redmide issues page terribly useful, we are encouraging the use of this GitHub page for all issues reporting related the WL pipeline.

2. Second, this repository does host some code that is related to the DES WL pipeline or testing thereof that does not have some other, more appropriate home.

   Items in this repository:

   • e2e

     This directory contains the code to make the so-called "end-to-end" test images. The script generates low noise images corresponding to a single coadd tile, including all the corresponding single-epoch images. It is intended to be used to test that the weak lensing pipeline gets all the various sign conventions correct for the wcs, the PSFEx interpolation, etc.

     • The starting point is an existing MEDS file from actual data.
     • It uses the existing data images for the WCS, although there are plans to modify this to use either more extreme or simpler WCS functions to make the tests either.
     • It uses an elliptical Gaussian PSF. The variation across each chip uses second order polynomials for the size, e1, and e2.
     • It adds a small amount of noise, set to be the minimum flux / 1000.

   • psfex

     This directory contains the code to rerun PSFEx on the red images using custom parameters and star selection appropriate for the requirements of the weak lensing pipeline. In particular:

     • It uses our own star-galaxy separation to select more reliable stars.
     • It removes bright stars that suffer from the bright-fatter problem.
     • It removes stars that fall in or next to the tape bumps to avoid these stars from contaminating the fits.*
     • It flags chips that are considered bad for any of the following reasons:*
       • Too few stars (<50)
       • Too many stars (>500)
       • Too large seeing (>1.8 arcsec FWHM)
       • Bad rho statistics (TBD)

     Note: Items above with the * are planned. Not yet done.

If you are a DESWL developer and would like to add code to this repo and/or edit existing code, please email Mike Jarvis and he will make you a collaborator, which will give you push access.

Other repos of relevance:

• The im3shape shear estimation code:

  • https://bitbucket.org/joezuntz/im3shape

• The ngmix shear estimation code:

  • https://github.com/esheldon/ngmix

  • https://github.com/esheldon/gmix_meds

• Code for making the meds files (also includes the shapelet PSF estimation code as well as a shapelet shear estimation code that we used to use but has too large noise biases):

  • https://github.com/rmjarvis/deswl_shapelets

• Code for computing shear correlation functions (i.e. the corr2 executable and also the treecorr python module):

  • https://github.com/rmjarvis/TreeCorr

• GalSim

  • https://github.com/GalSim-developers/GalSim