Additional tests for get_regional_data

[RichardMN]

Working through #119 we've hit some odd problems with the data coming out of the end of our pipe and I think @seabbs asked how we could have detected them earlier.

Here are some possible tests, likely to be applied to get_regional_data output, not on CRAN, but possibly as part of our regular nightly dataset sanity checking.

These should probably be run with localisation of regions off to make it easier to write generically. They will also rely in some places on comparing output at level 1 with output at level 2.

• Is the number of level 1 regions < the number of level 2 regions (when running with level 2 data) (perhaps this should be <=) (@RichardMN More additional tests on data in regional data #312)
• Is the number of level 1 regions (running level 1 data) = the number of level 1 regions (running level 2 data) (@RichardMN More additional tests on data in regional data #312)
• Is the number of level 1 region codes in [the number of level 1 regions - 2, the number of level 1 regions] (slightly amended, I think, in @RichardMN More additional tests on data in regional data #312)
• Is the number of level 1 region codes corresponding to NA in level 1 region codes in [0,1] (@RichardMN More additional tests on data in regional data #312)
• Is the number of level 2 region codes corresponding to NA in level 2 region codes in [0,1] (if there are level 2 region codes) (@RichardMN More additional tests on data in regional data #312)
• Is the level 1 region all NA? (@joseph-palmer Additional tests #307)
• Is the level 2 region all NA? (@joseph-palmer Additional tests #307)
• Is the total number of cases > the total number of deaths for each region and nationally? (@joseph-palmer Cases greater than deaths #309)

[seabbs]

This all looks pretty spot on. I can look at this this weekend/next week.

[joseph-palmer]

I've made a start with a couple of them, I will open an additional tests PR to load these into and if anyone else does any just merge into that PR :)

[joseph-palmer]

I think it's a good idea to run these nightly as many will need to work on the full downloaded dataset rather than the snapshot. Can anyone tell me how to ensure a test only runs on this?

[seabbs]

In my mind the get_ functions and friends should be tested and documented in a dataset independent way with dataset tests happening elsewhere. I've updated the title of this issue to reflect this.

[github-actions]

This issue has been flagged as stale due to lack of activity

[seabbs]

This has got a bit stuck. What do we think still needs doing vs nice to have here?

[RichardMN]

It's stuck. I might be able to take over @joseph-palmer's PR #307 or just embed it into #312 and try to bring it up to speed with the rest of the package. I don't think I'd do it better but I know this is something which I might be able to do and free up efforts for elsewhere.

[github-actions]

This issue has been flagged as stale due to lack of activity

[github-actions]

This issue has been flagged as stale due to lack of activity