type distribution vs parameter distribution format

[till-tietz]

@macartan

I just noticed that we export the transpose of posterior / prior distributions over causal types from get_type_prob_multiple (i.e. causal types ares rows and draws are columns) while the causal type posterior in stan_objects as well as distributions over parameters have the draws in rows. Is there any particular reason for this? I didn't want to change anything to avoid introducing breaking changes in downstream functions.

[macartan]

Yes this has bothered me

the natural shape would be as a data.frame with rows for draws

I think there's an internal use where we use the transpose

On Fri 22. Mar 2024 at 20:37, Till Tietz @.***> wrote:

@macartan https://github.com/macartan

I just noticed that we export the transpose of posterior / prior distributions over causal types from get_type_prob_multiple (i.e. causal types ares rows and draws are columns) while the causal type posterior in stan_objects as well as distributions over parameters have the draws in rows. Is there any particular reason for this? I didn't want to change anything to avoid introducing breaking changes in downstream functions.

— Reply to this email directly, view it on GitHub https://github.com/integrated-inferences/CausalQueries/issues/333, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADBE57IH7A35IA5DGHAXB53YZSCALAVCNFSM6AAAAABFD4ICLKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGIYDGMJYHA2DANQ . You are receiving this because you were mentioned.Message ID: @.***>