-
Notifications
You must be signed in to change notification settings - Fork 4
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
User interaction with Stan code #78
Comments
As an example of bad thing due to current implementation:
I think if you just did:
then it would put in the default as |
For this specific example can't you pass family in the same way you can pass data |
I.e epidist_standata(data, family) |
Issue here about how close to put the family related code together. Prefer to have it all in function but then that means having a combination of the family object and the Stan object together. Might be preferable (if) when we also have the family as a class |
I propose to:
Thoughts on this proposal @seabbs @kgostic @parksw3? (I will bring this up next meeting [today] and record here anyway.) My reasons for preferring this:
Uncertainties:
|
Third option:
Obvious option to change:
|
I have started working on this in #115. So far what I have done is move
This matches the specification under "third option" in comment #78 (comment) but I don't think is the complete solution we had in mind. It's not clear to me the best next step now. Options as I see them:
Tagging @seabbs to give input on this before I move forward with implementing. |
Nice issue.
Could you flesh this out a bit please? |
Yep, as to why I had that impression:
If these do not seem important issues I would be happy to move forward with the solution as written in the PR currently, and then we can revisit it when we add more models. I would also open an issue for "add custom Stan code prior" would would entail:
|
@seabbs ping here -- could you give thoughts before I either look to get a "leave as is" solution merged, or try implementing the "move the family into |
I think leave as is (the 2. Point is handled by S3 dispatch) And make issues for your point about custom priors and reorg the family function I think? |
I have made an issue for the priors point (#123). I don't fully understand what making an issue for the family function reorganisation would entail but I am going to try. |
The #70 PR raised some issues about how we want the user to be able to interact with the Stan code passed in for specific models (in this case
ltcad
). This will be an ongoing issue as more models are added.Currently:
epidist_stancode
method available and exposedQuestions:
Moving forward:
epidist_stancode.epidist_ltcad
what the different objects are and how they change the model. This would then allow the user to (perhaps) alter the output of that Stan codeThe text was updated successfully, but these errors were encountered: