dtamayo / spock

🖖 A package to determine whether planetary orbital configurations will live long and prosper
GNU General Public License v3.0
64 stars 15 forks source link

Classes #1

Closed hannorein closed 4 years ago

hannorein commented 5 years ago

Current syntax:

from spock import spockClassifier
model = spockClassifier()
model.predict(sim)

Maybe it is too early to worry about this, and we can definitely change later what the precise name of each class is, but it could make sense to define early what each object is.

1) Is model used interchangeably with classifier? If so, let's use only one of the two. If not, the difference is not obvious to me. 2) Since spockClassifier is a python class, I should start with a capital letter. 3) I would not repeat the name spock in the name of a class. It's redundant if it's already in the package name and a bit confusing. (Will the package spock offer many different Classifier classes? If so, then it should not include one called spockClassifier, something like DefaultClassifier would make it more intuitive.)

dtamayo commented 5 years ago

I agree with all of these. I did SpockClassifier in analogy to XGBClassifier, but I agree there shouldn't be any ambiguity...if there is ambiguity users can do spock.Classifier rather than from spock import Classifier.

I think what we name things in the end will depend on whether we e.g. bundle Miles timescale regressor and this classifier (to save the repeated short integration) or not. At that point we can agree on the right name.

Thanks!

dtamayo commented 4 years ago

Changed to StabilityClassifier, which is more descriptive