No description provided.
DOC: Silence fit.
ENH: Allow user to control formula namespace depth.
DOC: Add eval_env to documentation.
DOC: Add eval_env note to release notes.
Maybe patsy_ns or patsy_env to make it clearer that this is for Patsy?
Yeah I'm not totally sold on the name. I thought about just namespace_depth but it's too long. Does every user know that patsy evaluates the formula? I was assuming not, so that might not clear things up.
Since it is only used in from_formula I don't see any problem with the name. It should be clear from the context
We don't refer to patsy by name, we usually call everything just formula, e.g. formula_env
formula_ns or formula_env sound cleaner. eval_env seems really ambiguous since lots of things might be evaluated.
I don't feel strongly about this, but the keyword is called eval_env in the patsy documentation we point to and you could end up passing a EvalEnvironment instance to to it, so it is consistent this way. It's a bit redundant I think to need to specify formula twice, so maybe even just env?
OTOH, is there a case to give the user full control over the evaluation depth in library code? We can just have use_namespace=True and for the power users who understand the implications could overload this with an EvalEnvironment instance?
When it doubt just leave it...
Heh, I can get behind that.
I'm fine either way.
It will be an exotic, unused option for almost all users, I guess