-
Notifications
You must be signed in to change notification settings - Fork 22
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feature: Redesign interface #112
Conversation
Currently, something I am a bit stuck on is the interface for priors. At the moment this is centralised and it would be nice if it wasn't. I think for now I am going to propose sticking with the current system as I don't really have a good idea but this does mean a future release will need to also have breaking changes. Once the dust has settled a bit if there is still no good solution will open an issue. |
This now implements everything discussed in #57 (excluding the bells and whistles in the That being said as it is feature complete it would be good to get feedback at this stage. In particular on the setup of the modularity. This is all in Something that is currently less than ideal is the passing of priors, in particular, you cannot set priors for fixed effects. This makes things a bit clunky. In other work (https://github.com/seabbs/epict/blob/main/R/model.R and https://github.com/seabbs/epict/blob/a462b2482d2ddd176085e12ed824b48c42eb1a26/R/extract.R#L59) I've done a bit more on this and also got linking with model output working. I am still not happy with that implementation so planning on punting this to a later PR (which may sadly add some breaking changes but hopefully to less used functionality). Note: I have implemented interfaces for the expectation and missingness models which should make implementing them easier. The blockers for those features are updating the likelihood in the case of the missingness model (#64) and deciding what to do about the order of the expectation process (i.e is it on log cases, growth rate, or support multiple (with this being where I am landing and something that requires some small amount of thought to future proof against the addition of renewals etc + how to handle intercepts being allowed/not (as the order 2 model needs an initial intercept whereas the order 1 model doesn't)) |
Codecov Report
@@ Coverage Diff @@
## develop #112 +/- ##
============================================
+ Coverage 64.29% 85.10% +20.80%
============================================
Files 12 12
Lines 1042 1188 +146
============================================
+ Hits 670 1011 +341
+ Misses 372 177 -195
Continue to review full report at Codecov.
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM apart from linting flagged test gaps (self-review)
This PR aims to resolve #57. See the issue for interface design details and comments below for other changes.