Load correct dates for hindcast, forecast and observations for min/max-type daily data

Hi @eduzenli @vtorralba @nmilders

I am opening this issue in relation to what was mentioned in https://earth.bsc.es/gitlab/es/requests/-/issues/2663#note_355615

Since for variables like tasmin and tasmax, the format of the dates can be different than from other variables (shows the end of the period, which corresponds to the "next" day at 00:00), the data was not being loaded correctly and there wasn't full correspondence between the hindcast and the observations. For other variables like tas or prlr, the data was already correct.

I have fixed this for the subseasonal and seasonal cases and I am going to merge it to the master branch as soon as the pipeline passes. Apologies for the inconvenience! I will also leave this issue open to double-check the decadal case.

Bear in mind that the information provided in the data summary and in data$hcst$attrs$Dates might still seem "strange" (offset by one day) for these variables, but this is due to the actual metadata in the netCDF and not a mismatch in the data or a problem coming from SUNSET.

If you have any remaining questions or problems, please let me know and I will take care of them ASAP.

Apologies for the issues this may have caused,

Victòria

FYI @nperez