Setting up the scda.2022 repo

[shajoezhu]

Hi @pawelru and @nikolas-burkoff , just did a clean sheet update for the repo. Hope it's ok now.

[pawelru]

great thanks 👍 now let's open a branch and PR and figure out what to change:

• first point is of course version and news - i suggest to start now with 0.0.0.9000 and we would bump it up to 0.1.0 during next release
• then it is a real data snapshots - are we expecting a new RData file now (i.e. 202101??) or we are waiting for the next release?
• dependening on the previous one - stage_deps file - i can see that here we have scda.2022 <- tern and we need to have tern -> scda.2022. Moreover, I guess DESCRIPTION file needs to be updated. That needs to be done for every repo - that's quite a lot of PRs. Am I right @nikolas-burkoff ?

[nikolas-burkoff]

So for the DESC files - what we really want is in places which only require "latest" data for their tests and examples is

"Suggests scda and at least one scda.xxxx" we probably don't need to change those if we're happy to use scda.2021 for now and at some point in the future we can safely replace scda.2021 with scda.2022 when we archive scda.2021.

For packages which rely on a specific version of data from scda.2021 for their tests then they need to stay like that until the tests are updated to point to new data in scda.2022 (possibly updating test cases) - at that point we can remove the suggests 2021 and add 2022

For staged.deps - any package adding scda.2022 into its suggests also needs to add it as an upstream dep into the staged.deps file (with scda.2022 adding that package as a downstream package in its yaml file)

It's a shame we called it scda.year as that forces us to update each year, we could have just had scda for now and if that package got too big then either had scda_2 or removed some of the older to-be deprecated datasets.

[shajoezhu]

Thanks guys!

@pawelru

first point is of course version and news - i suggest to start now with 0.0.0.9000 and we would bump it up to 0.1.0 during next release Done then it is a real data snapshots - are we expecting a new RData file now (i.e. 202101??) or we are waiting for the next release? We are fixing upstream and refactoring in random.cdisc.data, it will come in the release in march

@nikolas-burkoff and @pawelru I am thinking out loud here, by any chance we can abstract the stage.dependencies to scda.202* ? We are versioning these data's by years. In examples, in theory, it should be able to scda and choose the year version, and release version. right?

What is the rational here to scda by years anyway? I am just wondering @gmbecker @waddella

Thanks guys!

[pawelru]

let me take it over

[shajoezhu]

Thanks so much Pawel!