0.7.5 - optimisation and parametric simulation

accim/parametric_and_optimisation/main.py

@@ -555,15 +555,17 @@ def sampling_full_set(self):
         Combines all values from all parameters and saves it into a pandas DataFrame, stored in an internal variable
         named parameters_values_df.
         """
+        from accim.parametric_and_optimisation.utils import make_all_combinations
         if self.descriptors_has_options:
             num_samples = 1
             parameters_values = {}
             for p in self.parameters_list:
                 num_samples = num_samples * len(p.value_descriptors[0].options)
                 parameters_values.update({p.value_descriptors[0].name: p.value_descriptors[0].options})
-            from itertools import product
-            combinations = list(product(*parameters_values.values()))
-            parameters_values_df = pd.DataFrame(combinations, columns=parameters_values.keys())
+            # from itertools import product
+            # combinations = list(product(*parameters_values.values()))
+            # parameters_values_df = pd.DataFrame(combinations, columns=parameters_values.keys())
+            parameters_values_df = make_all_combinations(parameters_values)
         else:
             raise KeyError('sampling_full_set method can only be used with option (i.e. category) descriptors.')
 
@@ -765,18 +767,27 @@ def run_optimisation(
         self.outputs_optimisation = outputs_optimisation
         # return outputs_optimisation
 
-    def get_hourly_df(self):
+    def get_hourly_df(self, start_date: str = '2024-01-01 01'):
         """
-        Transforms the hourly values of outputs_param_simulation to a new pandas DataFrame
-        named outputs_param_simulation_hourly.
+        Transforms the hourly values of outputs_param_simulation to a new pandas DataFrame, saved in the
+         internal variable named outputs_param_simulation_hourly.
 
+        :param start_date: the start date for the simulation results, in format 'YYY-MM-DD HH'
         """
         parameter_columns = [i.name for i in self.parameters_list]
         parameter_columns.append('epw')
-        self.outputs_hourly_columns = identify_hourly_columns(self.outputs_param_simulation)
-        self.outputs_param_simulation_hourly = expand_to_hourly_dataframe(self.outputs_param_simulation, parameter_columns)
-
+        self.outputs_param_simulation_hourly = expand_to_hourly_dataframe(
+            df=self.outputs_param_simulation,
+            parameter_columns=parameter_columns,
+            start_date=start_date
+        )
 
+    def get_hourly_df_columns(self):
+        """
+        Identifies the columns which contain hourly values, and save the names in a list, saved in the
+        internal variable named outputs_hourly_columns
+        """
+        self.outputs_hourly_columns = identify_hourly_columns(self.outputs_param_simulation)
 
 class AccimPredefModelsParamSim(OptimParamSimulation):
     def __init__(

accim/parametric_and_optimisation/utils.py

@@ -76,3 +76,15 @@ def identify_hourly_columns(df):
         df[col].apply(lambda x: isinstance(x, str) and x.startswith('[') and x.endswith(']')).all()
     ]
     return hourly_columns
+
+def make_all_combinations(parameters_values_dict: dict) -> pd.DataFrame:
+    """
+    Takes all values from all the parameters and return a pandas DataFrame with all possible combinations.
+
+    :param parameters_values_dict: a dictionary in the format {'parameter name': list_of_values}
+    :return: a pandas DataFrame with all possible combinations
+    """
+    from itertools import product
+    combinations = list(product(*parameters_values_dict.values()))
+    parameters_values_df = pd.DataFrame(combinations, columns=parameters_values_dict.keys())
+    return parameters_values_df

parametric_simulation_usage_v00_accim_custom.py

@@ -40,7 +40,7 @@
 
 accim.utils.set_occupancy_to_always(idf_object=building)
 
-test_class_instance = OptimParamSimulation(
+parametric = OptimParamSimulation(
     building=building,
     parameters_type='accim custom model'
     # output_keep_existing=False,
@@ -50,7 +50,7 @@
 
 # Setting the Output:Variable and Output:Meter objects in the idf
 #todo do not print on screen the process of accis, only the first time
-df_output_variables_idf = test_class_instance.get_output_var_df_from_idf()
+df_output_variables_idf = parametric.get_output_var_df_from_idf()
 
 df_output_variables_idf_mod = df_output_variables_idf.copy()
 
@@ -68,7 +68,7 @@
 
 [i for i in building.idfobjects['energymanagementsystem:program'] if i.Name.lower() == 'setinputdata']
 
-test_class_instance.set_output_var_df_to_idf(outputs_df=df_output_variables_idf_mod)
+parametric.set_output_var_df_to_idf(outputs_df=df_output_variables_idf_mod)
 
 output_meters = [
     # 'HeatingCoils:EnergyTransfer',
@@ -77,10 +77,10 @@
     'Cooling:Electricity',
     'Electricity:HVAC',
 ]
-test_class_instance.set_output_met_objects_to_idf(output_meters=output_meters)
+parametric.set_output_met_objects_to_idf(output_meters=output_meters)
 
 # Checking the Output:Meter and Output:Variable objects in the simulation
-df_outputmeters_2, df_outputvariables_2 = test_class_instance.get_outputs_df_from_testsim()
+df_outputmeters_2, df_outputvariables_2 = parametric.get_outputs_df_from_testsim()
 
 #Other variables could be reported. These can be read in the rdd, mdd and mtd files
 df_rdd = get_rdd_file_as_df()
@@ -105,7 +105,7 @@ def return_time_series(result):
 df_outputvariables_3 = df_outputvariables_3.drop(index=[2, 4])
 df_outputvariables_3['name'] = df_outputvariables_3['variable_name'] + '_time series'
 
-test_class_instance.set_outputs_for_simulation(
+parametric.set_outputs_for_simulation(
     df_output_meter=df_outputmeters_3,
     # df_output_variable=df_outputvariables_3,
     df_output_variable=df_outputvariables_3,
@@ -145,33 +145,64 @@ def return_time_series(result):
 # from besos.parameters import wwr, RangeParameter
 # other_parameters = [wwr(RangeParameter(0.1, 0.9))]
 
-test_class_instance.set_parameters(
+parametric.set_parameters(
     accis_params_dict=accis_parameters,
     # additional_params=other_parameters
 )
 
+##
+param_dict = {
+    'CustAST_m': [0.1, 0.6],
+    'CustAST_n': [22, 8],
+    'CustAST_ASToffset': [2.5, 4],
+    'CustAST_ASTall': [10, 10],
+    'CustAST_ASTaul': [35, 35],
+}
+from accim.parametric_and_optimisation.utils import make_all_combinations
+all_combinations = make_all_combinations(param_dict)
 
+##
 
 [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setinputdata']
 [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setvofinputdata']
 [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'applycat']
 
+##
+set_input_data = [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setinputdata'][0]
+set_vof_input_data = [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setvofinputdata'][0]
+applycat = [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'applycat'][0]
+setast = [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setast'][0]
+setapplimits = [i for i in building.idfobjects['EnergyManagementSystem:Program'] if i.Name.lower() == 'setapplimits'][0]
+other_args = {'SetpointAcc': setast.Program_Line_1}
+cust_ast_args = {
+    'ACSToffset': applycat.Program_Line_4,
+    'AHSToffset': applycat.Program_Line_5,
+    'm': setast.Program_Line_2,
+    'n': setast.Program_Line_3,
+    'ACSTaul': setapplimits.Program_Line_2,
+    'ACSTall': setapplimits.Program_Line_3,
+    'AHSTaul': setapplimits.Program_Line_4,
+    'AHSTall': setapplimits.Program_Line_5,
+}
+
+# building.savecopy('TestModel_mod.idf')
 
+##
 # Let's set the problem
-test_class_instance.set_problem()
+parametric.set_problem()
 
 # Let's generate a sampling dataframe
-test_class_instance.sampling_full_factorial(level=5)
-temp_full_fac = test_class_instance.parameters_values_df
+parametric.sampling_full_factorial(level=5)
+temp_full_fac = parametric.parameters_values_df
 
-test_class_instance.sampling_lhs(num_samples=3)
-temp_lhs = test_class_instance.parameters_values_df
+parametric.sampling_lhs(num_samples=3)
+temp_lhs = parametric.parameters_values_df
 
-# test_class_instance.sampling_full_set()
-# temp_full_set = test_class_instance.parameters_values_df
+# parametric.sampling_full_set()
+# temp_full_set = parametric.parameters_values_df
 
 #todo try to return series of pmot, acst, ahst and optemp and plot them in facetgrid
-outputs = test_class_instance.run_parametric_simulation(
+outputs = parametric.run_parametric_simulation(
     epws=[
         'Sydney.epw',
         'Seville.epw'

parametric_simulation_usage_v00_accim_predef.py

@@ -125,7 +125,7 @@
     accis_params_dict=accis_parameters,
     # additional_params=other_parameters
 )
-##
+
 # Let's set the problem
 parametric.set_problem()
 
@@ -139,7 +139,7 @@
 parametric.sampling_full_set()
 temp_full_set = parametric.parameters_values_df
 
-outputs = parametric.run_parametric_simulation(
+parametric.run_parametric_simulation(
     epws=[
         'Sydney.epw',
         'Seville.epw'
@@ -154,6 +154,7 @@
 
 # outputs.to_excel('WIP_outputs.xlsx')
 
+[i.name for i in parametric.sim_outputs]
 ##
 
 
@@ -189,17 +190,21 @@
 
 ##
 
+#Let's make a copy of the dataframe to not to modify the original one
 df = parametric.outputs_param_simulation_hourly.copy()
-set(df['epw'])
+
+# The name of the column for the Running mean outdoor temperature is very long, so let's save it in the variable rmot:
 rmot = [i for i in df.columns if 'Running Average' in i][0]
-optemp = [i for i in df.columns if 'Zone Operative Temperature' in i][0]
 
+#Let's remove the columns where value is the same for all rows
 for c in df.columns:
     if len(set(df[c])) == 1:
         df = df.drop(columns=[c])
-df = df.drop(columns=['hour', 'datetime'])
+#Now let's remove the hour and datetime columns, since will
+df = df.drop(columns=['hour'])
 
-df = df.melt(id_vars=['CAT', 'epw', rmot])
+# Now let's reshape the df for plotting purposes
+df = df.melt(id_vars=['datetime', 'CAT', 'epw', rmot])
 
 
 ##
@@ -219,7 +224,7 @@
     alpha=0.5
 )
 g.set_axis_labels('RMOT (°C)', 'Indoor Operative Temperature (°C)')
-g.add_legend()
+g.add_legend(loc='lower center')
 
 for lh in g._legend.legend_handles:
     lh.set_markersize(5)
@@ -234,7 +239,30 @@
 #     markerscale=2
 # )
 
+plt.tight_layout()
+
+
+##
+import seaborn as sns
+g = sns.FacetGrid(
+    data=df,
+    row='CAT',
+    col='epw'
+)
+g.map_dataframe(
+    sns.lineplot,
+    x='datetime',
+    y='value',
+    hue='variable',
+    # s=1,
+    # alpha=0.5
+)
+g.set_axis_labels('RMOT (°C)', 'Indoor Operative Temperature (°C)')
+g.add_legend(loc='lower center')
+plt.tight_layout()
 
+# for lh in g._legend.legend_handles:
+#     lh.set_markersize(5)
 
 
 ##