First pass of the CDC Vaccination Indicator

[Ananya-Joshi]

Description

First pass on the CDC Indicator with tests

[Ananya-Joshi]

Missing HI and TX data by county.

[krivard]

Excellent start!

[Ananya-Joshi]

The changes requested have been made. We should talk about the two open issues still - the unknown states/counties and the megacounty thread.

[dshemetov]

Btw, you might like this version of taking diffs, grouped by geos.

[dshemetov]

I think the .isna() in that code should probably be replaced with .loc[min_time_value, :] or something, but the gist is there.

[Ananya-Joshi]

I wonder if we can keep this method for now, but then later I'll look to fix the method and use it.

[dshemetov]

Sure, that's alright. A possible refactor for later.

[Ananya-Joshi]

@krivard @dshemetov , do we need to do anything else for this indicator?

[dshemetov]

@Ananya-Joshi just the Nancodes annotation in the thread above and it should be good to go 👍

[Ananya-Joshi]

@Ananya-Joshi just the Nancodes annotation in the thread above and it should be good to go 👍

As far as I can see on my local, they're changed! Is this not the case for what you can see?

[dshemetov]

Oh, now we're missing the actual "se" and "sample_size" columns. I will make a commit with what I mean.

[dshemetov]

LGTM

[krivard]

This looks like rebase leavings:

Indeed, the gitref for "Remove test for old export func" is supposed to be 5fa6314 and not ca09586. @Ananya-Joshi and @dshemetov can you clarify what was done here? I want to make sure when we merge it we won't wind up with duplicate copies of all of those dual-author commits.

[Ananya-Joshi]

Hi Katie, I did do a rebase so that the new create_export_csv function would be used. Is there something I can fix or undo?

[krivard]

was it just

$ git rebase main

or something more complicated?

[Ananya-Joshi]

It started out that way, but I also needed to manually fix some parts that were conflicting. Should I roll back?

[krivard]

None of the files added in this branch existed before, so the fact that there were conflicts is highly suspicious. I'm also looking at the merge commit you have there -- b9c6e8a -- mixing merges and rebases is probably not helping.

There are multiple options for repairing the commit history, largely because there are multiple acceptable options for where we want to end up when we're done:

• (A) a branch starting from 56991f9 that only includes commits for cdc_vaccines or merge commits+whatever comes with each merge (merge don't rebase)
• (B) a branch starting from wherever main is now (currently 5b741ec) that only includes commits for cdc_vaccines (rebase don't merge)

This PR includes 136 commits, but only ~32 of them were for cdc_vaccines:

• 89410dc^..f67925b
• maybe 90ea653?
• d3544d0^..ea6587d

Ways to get from here to one of the options above include:

• (A1) rebase -i (no other arguments) and drop all the non-cdc_vaccines commits, then git merge main. There's only one or two ranges to drop, but that would include 104 individual lines in whatever editor you use for interactive rebases.
• (B1) rebase -i (no other arguments) and drop all the non-cdc_vaccines commits, then git rebase --onto main. If you get conflicts, call for help, for the reasons explained above.
• (B2) checkout a fresh branch from main, then cherry-pick the 32 cdc_vaccines commits into this new branch (details below). There are only 2 or 3 ranges to cherry-pick, so this probably has the lowest opportunity for error so long as there are no surprises. Once you're done, discard the old branch, then rename the new branch to replace it.
• (B3) rebase version of B2 (details below). This one is better if you run into trouble, since rebase has way better conflict & recovery handling than cherry-pick. Con: I've never used this method myself; found it on the internet. Once you're done, discard the old branch, then rename the new branch to replace it.

Whatever you choose, you'll then need to git push --force to upload it into the PR.

Cherrypick method (B2) details:

git checkout main
git checkout -b indicator_cdc_vaccines_fixed
git cherry-pick 89410dc^..f67925b
# git cherry-pick 90ea653 #(might not be needed)
git cherry-pick d3544d0^..ea6587d

Rebase method (B3) details:

git branch indicator_cdc_vaccines_fixed indicator_cdc_vaccines
git branch signpost/stop f67925b
git rebase --onto main 89410dc^ signpost/stop
# git cherry-pick 90ea653 #(might not be needed)
git rebase --onto signpost/stop d3544d0^ indicator_cdc_vaccines_fixed

[Ananya-Joshi]

Thank you Katie for this super detailed comment- I learned a lot and it seems like 2b will be successful (will push shortly or will create a new PR).