Changes to not use the model if there is any NA. UK still not finished for LOOCV if we don't have model for a certain day/hour

minor changes to adapt the code to read csv files. IMportant changes has been applied in loop_uk_cat to adapt the local stations and observations to appli OK and LOOCV

Corrected a minor error in Set Units in the variable mod_constant. Also, we added read_year in set temporal context

we added some variables in case we treat diferent type of data or scripts, like netcdf vs csv, daily or hourly data, etc.

Modifications in config file variables, and set stations, depending if the file is from ghost (netcdf) or gene (csv)

We changed the calling name of the data script, to a variable that depends on the pollutant choosed. This variable is setted at set units

We changed the set units, now is different depending on the pollutant choosed, also the name of the data script

pixels_final.R, is a script to interpolate the results in a smaller grid (or larger) and then ploted for example with mapview function

We added in the config file (so on the affected files) the option to save model, obs, ok, loocv & stations. Previuslly it was saved allways, now we have the option to save it or not (usefull when we are debugging the code).

Some corrections to filter wrong observations before applying UK. And also, if in loocv the slope is negative, we substitut the value for the raw model without applying loocv in that hour

Some corrections to filter wrong observations before applying UK. And also, if in loocv the slope is negative, we substitut the value for the raw model without applying loocv in that hour.