df_obs_sub has been modified when we don't have the data for all dates, putting NA to have the same length as length_loop

slope > -20 considered for correction, otherwise NA's are set. Barnes statistics in a different script

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)