CCLX
CCLX copied to clipboard
Replace LSST SRD Redshift Distributions and Binning file (fix filename)
- Removed
LSST SRD Redshift Distributions and Binning.ipynb(file with spaces) - Added
LSST_SRD_Redshift_Distributions_and_Binning.ipynb(correct filename with underscores) - Removed
binning.pyandsrd_redshift_distributions.py - Added
lsst_galaxy_sample.py - Added
data_output/directory with all required files - Updated
parameters/directory
This update improves and condenses the SRD nz and binning procedure.
To validate:
- Clone the
niko_nz_binning_updatebranch. - Open the validation notebook
LSST_SRD_Redshift_Distributions_and_Binning_validation.ipynb. - Compare the old and new procedure.
Let me know if further validation is needed.
@damonge @elisachisari @chrgeorgiou This can be merged without any issues. I presented the updates last week in the forecasting telecon.
For anyone who wants to verify the changes, the steps are outlined in the PR description.
The main update is the binning optimization—I revised the old code (which is now a few years old), making it more logical, structured, and easier to use. I also addressed instability issues related to redshift grid resolution.
I'm happy to merge it myself, but feel free to do so if you'd prefer. It's important that we have this code available now, both for the paper and for integration with other DESC pipeline developments.
Can you please clarify how the bin centers are being defined?
There seems to be a typo here:
The docs say I can set the forecast years to 4 and 7 as well, but those aren't working for me:
Can you please clarify these are for the full redshift range without binning?
I am not sure this definition is ok: the normalization doesn't seem to match. Maybe you just want to refer to p(z)=1/N_0 dN/dz as the redshift distribution normalized to 1 and ignore number density of galaxies overall. CCL will normalize everything internally to 1.
The branch I cloned was: srd_nz_binning_pr. Is that the correct one? It's linked at the top of the issue but you mention a different one on your comment (which I think is behind this one).
When I try to run the validation notebook, I get stuck here: