search

aimalz / scippr

Supernova Cosmology Inference with Probabilistic Photometric Redshifts
4 stars 2 forks source link

Write a setup.py (and turn inference notebook into corresponding module) #5

Open tinapeters opened 7 years ago

tinapeters commented 7 years ago

One goal for hack week, after the breakout.

aimalz commented 7 years ago

We didn't get to it during Hack Week, but I can take care of this once we turn the not-so-skeletal-anymore notebooks into a formal package.

aimalz commented 7 years ago

I'm going to tackle this next. There's a small bug in the inference notebook (actually, it may just be a symptom of not running MCMC long enough to get out of the burn-in) that will be easier to debug if the scippr inference is a module than if it's effectively a script in the form of a Jupyter notebook.

Edit: I take that back. I think we might want to consider a different parametrization of the p(type, redshift, distance modulus) data because we're ultimately wasting a lot of computation multiplying sparse arrays. I'm going to think about this more before writing the module.