After PLANNER-1924, add 1 new constraint:

Proposal A)

1) For each employee, pair every 2 sequential shifts.
For example: given shifts

Monday06-14, tuesday06-14, tuesday14-22, wednesday14-22

that gives

{Monday06-14,, tuesday06-14},
{tuesday06-14, tuesday14-22},
{tuesday14-22, wednesday14-22}

Now, for each pair: if the two shift's wards (= spots) have a different covidWard boolean, then penalize for each pair.
It doesnt matter if the migrate between wards that are both covid or both non-covid.

Proposal A2)
Same as A) but only when going from covid to non-covid. The reverse doesn't matter.

Proposal B)
Time matters. If you've just had a long weekend off, then it matters less that your last shift was in a covid ward and your new one isn't.
However if you have 2 sequential shifts, and the first in a covid ward, then you might still have covid particles sticking to you when walking into the other ward...

On the other hand, covid's incubation is 4-6 days on average (up to 12 days) for symptoms. But infection starts 2 days before that, last I heard.

B not specced out. I think it's not needed.

Besides implementing A) as a 10 soft penalty, we should add a hard constraint C that penalizes starting a non-covid shift after finishing a covid shift less than 8 hours earlier.