main_regression_model_class.py

""" Main file to run seasonal and monthly forecast"""
# Current scripts directory

from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import LinearRegression
from class_regression_v2 import RegressionModel


def run_model(month_initialisation, cod_cuenca):
        """
        @Parameters
        X : predictors. e.g. pp_mm cumsum, tem_C avg, soil_moisture_mm cumsum, snow_swe avg, etc
        y : cumulative volume in a period. e.g. vol_mm sep-mar, jan-mar, etc
        q : average monthly streamflows in a period (same as volume). e.g q_mm sep, q_mm oct... q_mm mar
        """
        ## Step 1: import and split data (predictors, volume)
        model = RegressionModel(month_initialisation = month_initialisation,
                                window_strategy = 'dynamic',
                                catchment_code= cod_cuenca,
                                predictor_list=["pr_sum_-1months","tem_mean_3months"],
                                wy_holdout=1981)
        
        ## Step 2: compute determinist and ensemble forecast (regression model)
        model.set_model_parameters(model_scaler = StandardScaler(),
                                   model_regressor= LinearRegression())
        
        model.forecast_vol_ensemble(n_ens_members=1000)
        # ## Step 3: plot ensemble forecast
        # model.subplot_boxplot_chronological_median_order(export=False, subplot=False)
        # ## Step 4: compute monthly streamflows based on closest predictors from previous years
        # model.forecast_q_ensemble(n_neighbors=10,
        #                           weights='distance',
        #                           p=2)
        
        # ## Step 5: plot ensemble flow forecast
        # model.plot_knn_flow(export=False)
        
        #model.save_modelinfo()
        
        return model

#if __name__ == '__main__':

for month in ["may","jul","sep","nov"]:
    model_info=run_model(month,'5410002')