search

CosmoStat / unions-shear-ustc-cea

Projects of the USTC - CEA collaboration with UNIONS weak-lensing data
MIT License
0 stars 3 forks source link

GGL physical coordinates #19

Closed martinkilbinger closed 1 year ago

martinkilbinger commented 1 year ago

Stacking of GGL in physical instead of angular coordinates.

martinkilbinger commented 1 year ago

Basic directory for tests is candide:/home/mkilbing/astro/Runs/UNIONS/ggl-agn/jobs.

martinkilbinger commented 1 year ago

Test 1: Comparison of three different ways to compute bg - fg correlations. Runs of compute_ng_binned_samples.py. Results in Shen22_and_Liu19_logM_min_7_nz_0_z_0.1_0.8_test_nfg.

  1. treecorr.process (stack=auto)
  2. treecorr.process_cross + treecorr.finalize (stack=cross)
  3. treecorr.process + unions_wl.post-process stacking (stack=post)
martinkilbinger commented 1 year ago

Test 2: For narrow redshift range, compare angular and physical stacking. Runs in Shen22_and_Liu19_logM_min_7_nz_0_z_0.5_0.8_test_physical. Setting:

image

martinkilbinger commented 1 year ago

Test 3: Compare measured g_t to theoretical prediction. Run in Shen22_and_Liu19_logM_min_7_nz_0_z_0.5_0.8_test_physical.

Averaging still to check.

image `

martinkilbinger commented 1 year ago

Updated code (commit 9e7d03558abcfeb7ff163ee17512e2885dd45888, https://github.com/CosmoStat/unions-shear-ustc-cea/pull/27/commits/9e7d03558abcfeb7ff163ee17512e2885dd45888) seems to work now.

image

martinkilbinger commented 1 year ago

Averaging in theory.py:gamma_t_theo_phys now makes sense.

Range estimation of angular from physical coordinates (compute_ng_binned_samples.py:get_theta_min_max) seems to work now.

Jackknife mean and std agree with sample mean and Poisson.

To create last plot above, re-run ggl_compare_data_theory.py with --scales=angular, it will use the meanr column instead of r_nom (which in this case are the physical coordinates in Mpc)