Implementation of reduced shear in source selection for clusters by eduardojsbarroso · Pull Request #398 · LSSTDESC/TXPipe

eduardojsbarroso · 2025-08-01T11:38:49Z

So far, the cluster code only measures Delta Sigma from the select source galaxies around a cluster. We need to implement the same for gamma and we are also interested in using N(z) to select the galaxies behind the cluster instead of redshift PDF or zmode or zmean.

eduardojsbarroso · 2025-08-01T11:41:36Z

In this code. We set a maximum radius of search for galaxies around a cluster. Given the cluster redshift and a fiducial cosmology, we can transform this distance into an angle that changes for each cluster given their redshift.

If I do not want to assume any fiducial cosmology and search the galaxies by angle directly, should I use the same search angle for all clusters? Or is there some other approach to reescale this distance with redshift without using cosmology? @marina-ricci @joezuntz

joezuntz · 2025-08-01T13:56:03Z

There's definitely no way to convert a distance to an angle without knowing the cosmology. So I guess you would have to use the same angle, and make sure it's large enough. Why don't you want to use a fiducial cosmology? It shouldn't have a significant impact on the results at this level.

eduardojsbarroso · 2025-08-04T11:55:14Z

There's definitely no way to convert a distance to an angle without knowing the cosmology. So I guess you would have to use the same angle, and make sure it's large enough. Why don't you want to use a fiducial cosmology? It shouldn't have a significant impact on the results at this level.

Thank you for the reply Joe. I just wanted to know if there was any other way around it since we do not need the cosmology for the tangential signal. But ok, then I think I will leave this untouched and maybe add an option to use a constant angle if wanted

marina-ricci · 2025-09-22T14:45:51Z

@eduardojsbarroso for this special case I don't think it is a problem. Indeed we can use a fiducial cosmology just to define theta_max, it doesn't (should't) interfer with the "scientific" measurements. And even having radial binning in Mpc is not as much of a problem as having e.g. sigma_crit.

marina-ricci · 2025-09-22T14:58:16Z

Also, note for ourselves : we should through an error message when our binning is larger than the search radius (in case it's not hapening already)

marina-ricci · 2025-09-23T13:09:46Z

We decided it makes no sense to implement the N(z) method in source_compute_select, as it will be done more cleanly with the NG correlation.

eduardojsbarroso · 2025-09-26T14:33:27Z

@marina-ricci I have cleaned the option of Nz, as as discussed with @combet, we can implement this in another pull request, and a third one for the treecor option. Could you take a look at the code and check what you think?

eduardojsbarroso · 2025-09-26T14:33:46Z

Also, note for ourselves : we should through an error message when our binning is larger than the search radius (in case it's not hapening already)

I think this is already done in the make_ensemble_profile.py

            cat_max_distance = (clmm.utils.convert_units(np.max(bg_cat["distance_arcmin"]), "arcmin", "Mpc", z_cl, self.clmm_cosmo) if shear_units == "mpc" else np.max(bg_cat["distance_arcmin"]))
            if cat_max_distance < self.distance_bins[-1]:
                print ("!!! maximum distance of source smaller than distance_bins")

eduardojsbarroso · 2025-10-02T15:08:40Z

@marina-ricci I added a pipeline file to test this new functionalities and a notebook if you want to explore the results. Everything seems into place for me

joezuntz

In general I don't know enough to comment on the clusters code, so I will just review to make sure that things look structurally correct and don't cause any issues with the wider code.

This looks fine from that perspective, though I've added a note about a change you'll need to make because of other changes in the rest of the imports. I'm happy for you to merge it whenever the CL people are happy!

joezuntz · 2025-10-09T09:10:52Z

+    name: mini
+    interval: 3.0
+modules: >
+    txpipe


You will need to add txpipe.extensions here because of changes in the structuring of the code.

combet · 2026-02-06T13:57:09Z

    nzbins: 301
-    columns_file: ./data/bpz_riz.columns
-    data_path: ./data/example/rail-bpz-inputs
+    columns_file: /sps/lsst/groups/clusters/cl_pipeline_project/TXPipe_data/bpz_riz.columns


Maybe avoid hard-coded paths (this is for cc-in2p3 only) so users may more easily use the examples. Or at least add a comment to explain how they should change it to match their setup.

combet · 2026-02-06T13:57:58Z


-output_dir: ./data/cosmodc2/outputs-20deg2-CL
-config: ./examples/cosmodc2/Cluster_pipelines/config-20deg2-CL.yml
+output_dir: /sps/lsst/users/ebarroso/TXPipe/data/cosmodc2/outputs-20deg2-CL


Same as above regarding path

combet · 2026-02-06T14:01:58Z

-        "r_min" : 0.2, #in Mpc
-        "r_max" : 5.0, #in Mpc
+        #distance bin definition
+        "d_min" : 0.5, #in Mpc or acrmin


To me, r_min and r_max is more intuitive than d_min, d_max. And 'radial bin' is more used I think, even when thinking of angular separation. Just a suggestion.

combet · 2026-02-06T14:04:34Z

-        radius_edges = np.logspace(np.log10(rmin), np.log10(rmax), len(radius_centers) + 1)
-        for i in range(len(radius_edges) - 1):
-            bin_radius_dict[f'radius_{i}'] = (radius_edges[i], radius_edges[i+1], radius_centers[i])
+        distance_centers = np.array(data['bin_zbin_0_richbin_0']['clmm_cluster_ensemble'].stacked_data['radius'])


Similar comment as above. distance_centers is a bit cryptic. Maybe radial_bin_centers instead, or something along those lines?

Changed to radial_bins_centers, Also changed in the file dmin to rmin, and so on.

combet · 2026-02-06T14:06:01Z

-    def add_deltasigma_data(self, sacc_obj, data: dict, survey_name: str):
+    def add_shear_data(self, sacc_obj, data: dict, survey_name: str, profile_type: str):
        """
        Adds cluster shear (delta sigma) data to the SACC object.


Adds cluster reduced shear or delta sigma (depending on profile_type) profile to the SACC object.

combet · 2026-02-06T14:32:37Z

-        self.radial_bins = clmm.dataops.make_bins(self.config["r_min"], self.config["r_max"], nbins=self.config["nbins"], method="evenlog10width")
-        print (self.radial_bins) 
+        import clmm.cosmology.ccl 
+        if self.config["units"].lower() == "arcmin":


I find it a bit counter-intuitive to have a units keywords decide the type of radial binning (angular or in physical distance). Maybe there should be another keyword radial_bin_type (e.g. 'radius' or 'angle' ) or something like this to control it?

Good to keep the units information though, maybe to save it later as metadata in the sacc file?

I checked this part of the code. I decided that r_min is for radial bin, thus I dont need to add radial_bin_type, because I use the units on the ClusterEnsemble only and this gives me the profile.

The profile type information is needed only in the individual computation. Now we just need:
radial bin max, radial bin min, and units. And clmm does everything automatic. It is just taking the min, max and divide in n bins. And then when stacking, giving these bins + units measure.

I think the radial bin info is not needed in the ensemble, but on the individual only.

combet · 2026-02-06T14:35:14Z

-
-            if (clmm.utils.convert_units(np.max(bg_cat["distance_arcmin"]), 'arcmin', 'Mpc', z_cl, self.clmm_cosmo)< self.radial_bins[-1]):
-                print ("!!! maximum radial distance of source smaller than radial_bins")
+            cat_max_distance = (clmm.utils.convert_units(np.max(bg_cat["distance_arcmin"]), "arcmin", "Mpc", z_cl, self.clmm_cosmo) if shear_units == "mpc" else np.max(bg_cat["distance_arcmin"]))


Need to find another name for shear_units. "shear" in this code is either dimensionless or surface mass density if working with DeltaSigma.

I will change this to bin_units, like done in clmm and do the same in the configs.

combet · 2026-02-06T14:35:55Z

            gc_object.make_radial_profile(
-                    "Mpc", 
-                    bins=self.radial_bins,
+                    shear_units, 


Same as above

Initial implementation of gamma and nz source selectio

ab5fb6c

eduardojsbarroso mentioned this pull request Sep 16, 2025

Tangential shear implementation in TXPipe data vector and N(z) source selection LSSTDESC/CLPipeline#12

Open

5 tasks

Continued make cluster ensemble implementation

0671897

eduardojsbarroso added 3 commits September 25, 2025 04:46

Initial clean of Nz choice

0737dd0

Cleaned and tested reduced shear

9468ec6

Cleaned Nz option and tested notebook

64de747

eduardojsbarroso changed the title ~~Implementation of reduced shear and N(z) in source selection for clusters~~ Implementation of reduced shear in source selection for clusters Sep 26, 2025

eduardojsbarroso requested a review from marina-ricci September 26, 2025 14:30

eduardojsbarroso closed this Sep 26, 2025

eduardojsbarroso reopened this Sep 26, 2025

eduardojsbarroso marked this pull request as ready for review September 26, 2025 14:35

eduardojsbarroso marked this pull request as draft September 29, 2025 09:10

Cleaned Notebook and added 20deg example

f9df36f

eduardojsbarroso marked this pull request as ready for review October 2, 2025 15:07

eduardojsbarroso requested a review from joezuntz October 3, 2025 08:06

joezuntz approved these changes Oct 9, 2025

View reviewed changes

eduardojsbarroso and others added 4 commits January 22, 2026 10:58

Merge branch 'master' into clusters_source_selection

dcd80ef

Fixed binning for angular binning

e6ff3aa

Fix typo

d2f07cf

Fix angle choice

2a67cb9

eduardojsbarroso requested a review from combet February 6, 2026 10:53

Updated notebook

a8cb73a

combet reviewed Feb 6, 2026

View reviewed changes

eduardojsbarroso added 3 commits February 17, 2026 13:46

Changed covariance computation and save in sacc

8fdb79f

Changed mock ingestion to add shape noise as true

29cf107

Applied review comments

3d06d20

eduardojsbarroso force-pushed the clusters_source_selection branch from 72f8348 to 3d06d20 Compare June 4, 2026 14:04

eduardojsbarroso merged commit e93ea91 into master Jun 5, 2026
26 checks passed

Uh oh!

Conversation

eduardojsbarroso commented Aug 1, 2025

Uh oh!

eduardojsbarroso commented Aug 1, 2025

Uh oh!

joezuntz commented Aug 1, 2025

Uh oh!

eduardojsbarroso commented Aug 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

marina-ricci commented Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

marina-ricci commented Sep 22, 2025

Uh oh!

marina-ricci commented Sep 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eduardojsbarroso commented Sep 26, 2025

Uh oh!

eduardojsbarroso commented Sep 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

eduardojsbarroso commented Oct 2, 2025

Uh oh!

joezuntz left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

combet Feb 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

eduardojsbarroso commented Aug 4, 2025 •

edited

Loading

marina-ricci commented Sep 22, 2025 •

edited

Loading

marina-ricci commented Sep 23, 2025 •

edited

Loading

eduardojsbarroso commented Sep 26, 2025 •

edited

Loading

combet Feb 6, 2026 •

edited

Loading