Skip to content

Conversation

@sethaxen
Copy link
Member

This PR uses Preferences.jl to set defaults for:

  • ci_kind
  • ci_prob
  • weights_method
  • pretty-printing settings (internal)

The goal is to mirror Python ArviZ's use of RcParams for deeply customizing defaults used in stats functions and plots, with the key difference that in Python one can use a context manager to modify these preferences interactively, whereas in Julia these should be set for a project using LocalPreferences.toml before launching an interactive session for analysis. As a result, this is primarily useful for project-wide settings (e.g. a user prefers HDI over ETI and wants a different CI probability everywhere). A secondary goal is that dependent packages (e.g. MCMCChains/InferenceObjects) can then access these preferences when setting keyword arguments a little more safely than just copying over our defaults.

At the moment this is a proof-of-concept I'll keep as a draft until I receive user/dependant-dev feedback on whether this is sufficiently useful to merge.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant