Matthew R. Becker
Matthew R. Becker
Thanks for making this one! One comment here that I did not store any tokens in plain text on the conda-forge side on purpose to increase security. The tokens are...
We'll also need corresponding changes in the ci-setup scripts repo and possibly the webserver.
Validation depends on how the tokens are stored: https://github.com/conda-forge/conda-forge-webservices/blob/main/conda_forge_webservices/feedstock_outputs.py#L35 We could/should clean this up so we don't have these coordination issues.
I wrote that since verifying via an https request is faster. We should upstream it back to smithy.
Interesting. I was thinking of something more automated where if the conda-forge.yaml has the right setting and the PR is a version bump, it downloads the artifact, uploads to someplace...
Also 10.9
How long is the running time now? Yes in general it would. Most other options I know of have similar costs or more.
You only need to do this for a representative subset fwiw. Then you can likely fix the regularization for the rest of the survey.
> We could also consider a different normalization type that does the pure psf model version, which Eli thinks should work, but Lupton thinks won't. LLOL
> normalization is the median/mean/clipped mean/favorite statistic of the the ratio between the unnormalized psf flux and reference aperture flux Is the norm supposed to be constant per exposure, or...