v0.3.0 release

[zsusswein]

This release focuses on cleanup and testing of known bugs. It has room to add additional documentation, bugs, or feature requests as they become apparent.

The goal for this release is the end of March.

CI/CD

• Move actions to cfa-actions and implement here?

Model

• Clean up env vars #165
• Edit Horizon for sample_fit.RDS in tests #162
• Clean up metadata handling #122
• Add test for diagnostically failing fit #66
• Should diagnostic thresholds be exposed in the config #53
• Suppress only some warnings in PMF formatting #31

Vignette

• Vignette #38
• If needed, another vignette with a different focus?

[zsusswein]

There's an open question for @kgostic about the urgency of thinking about a random walk and whether we want to track it in this release or one after? See #37 for more.

[kgostic]

Thanks for calling me in.

I do think there's a lot of interest from JP and JA in being able to compare the daily random walk to the approximate GP. There have been a few weeks where the GP may have amplified non-neutral end behavior and affected the trend. We've been asked to test this several times by running an rw side by side and doing the comparison, but in the absence of an easy way to deploy the RW, I keep kicking this down the road.

I agree with @zsusswein that if we do this the priority should be simplicity and readability. We're building this code for ourselves, not for some general user, and we want our partners to be able to read what we're doing easily. Can we just add an if/else statement and ignore the gp-specific parameters if the rw is specified? I know this is bad design because in theory because of the potential for explosive complexity, but we've been running this pipeline for almost two years now and we know that this is the one thing that we actually do want to switch between sometimes.

[kgostic]

I'll add that if this is going to be really challenging to implement, we could save it for the pyrenew testing, or for after we build a better comparison tool. Let's talk f2f about difficulty @zsusswein