[TIME001] Adjust check_time_squareness to follow sampling times guidance

[sol1105]

In the sampling times guidance (currently here, but soon properly published), monthly time points can be reported at 00:00 of the first of the month, or 00:00 of the 15th of the month. In the PR I also allowed to provide monthly time points at the exact middle of the month, as first modeling centers already produced monthly time point data like this. check_time_squareness requires one of these three options to be consistently applied throughout the file.

To be checked if other frequency/cell_method combinations require special treatment as well.

[taylor13]

I will not have time to review the code, but I agree that

• we should not QC check what day/hour of the month monthly time "point" data is reported (because this depends on variable and on quite obscure data request specifications). Scientifically, I think it makes sense to report point data at the same time of day every month, but that obviously depends on whether data undergoes a significant diurnal cycle.
• I wouldn't place high priority on checking that for every month the data are reported on the same day. I would think it would be pretty difficult to get that wrong. But if the code written to do this works, it certainly can’t hurt to apply it.

summary: I don't see a reason to prevent committing this.

[sol1105]

For climatologies this likely still needs an update, but that could be made by a separate PR.

One could have separate checks for the time-of-day and whether there are duplicate time steps, gaps, ..
Currently both is combined in this check. The latter needs high severity while the time-of-day could be of medium or low severity. But for the time being this could be merged so the monthly point data does not generally raise high severity issues any longer and the checks could then be split in a later PR.