call indicator function

Author: mmbellani

.gitignore

@@ -1,3 +1,4 @@
 Data/
-**/_pycache__
-*.pyc
+__pycache__
+**/__pycache__
+*.pyc

Code/Plotting/__pycache__/plots.cpython-37.pyc

@@ -18,6 +18,7 @@
 # custom functions
 from Configuration.config import cfg_path
 from Code.Utils.utils import Utils
+from Code.Scoring.kpi import Kpi
 
 class Plots:
 
@@ -111,7 +112,7 @@ def sliding_fcst_plot(df, predict_col, expected_values, chart_title="", kpi=True
         ## Checking KPI
         if kpi == True:
             try:
-                mape = str(round(df.loc[~df.absolute_percentage_error.isnull(), 'absolute_percentage_error'].mean()*100, 2))
+                mape = str(round(Kpi.compute_mape(df, 'fcst', y), 2)*100)
                 min_mape_date = min(df.loc[~df.absolute_percentage_error.isnull(), date]).strftime("%d-%m-%Y")
                 max_mape_date = max(df.loc[~df.absolute_percentage_error.isnull(), date]).strftime("%d-%m-%Y")
                 chart_title = chart_title + ' - MAPE: ' +  mape + "% from " + min_mape_date + ' to ' + max_mape_date

Code/Plotting/plots.py

@@ -44,7 +44,7 @@ def sddi(array, highest=0.05, lowest=0.05):
         sddi = np.std(winsorized_array)
         return sddi
 
-    def idclass3(vect, threshold, perc, quant, highest, lowest):
+    def compute_indicator_values(vect, threshold, perc, quant, highest, lowest):
         ''' Computes indicator values
         :params: vect as numpy array, threshold as numeric, perc as numeric, quant as numeric, highest and lowest as scalars 0<=x<=1 as winsorization percentages
         :return: a dictionary
@@ -97,8 +97,8 @@ def idclass3(vect, threshold, perc, quant, highest, lowest):
 
         return res
 
-    def enh_idclass5(vect, threshold, perc, quant, highest, lowest):   
-        ''' Computes indicator values
+    def enh_compute_indicator_values(vect, threshold, perc, quant, highest, lowest):   
+        ''' Computes indicator values (enhanced)
         :params: vect as numpy array, threshold as numeric, perc as numeric, quant as numeric, highest and lowest as scalars 0<=x<=1 as winsorization percentages
         :return: a dictionary
         '''    
@@ -263,4 +263,10 @@ def classify_intermittent(df, type, thres_cv2_constant, thres_cv2, thres_adi, th
         df_profiling = pd.concat([df_regular, df_constant_zero, df_constant, df_intermittent, df_lumpy, df_erratic, df_unforecastable_time, df_unforecastable_quantity], axis=0)
 
         return df_profiling
+    
+    def call_intermittent_function(func, *args):
+        from Code.Profiling.Intermittent.intermittent import Intermittent
+        func_dict = {'enh_compute_indicator_values': Intermittent.enh_compute_indicator_values, 'compute_indicator_values': Intermittent.compute_indicator_values}
+        result = func_dict.get(func)(*args)
+        return result
 

Code/Profiling/Intermittent/__pycache__/intermittent.cpython-37.pyc

@@ -34,12 +34,9 @@ def define_train_test_set_dates(df, y, train_start_date, train_end_date, test_st
         else:
             test_end_date = pd.to_datetime(test_end_date, format='%Y-%m-%d')
 
-        if (test_start_date=='') & (min_test_start_date>max_train_end_date):
-            offset_date = pd.to_datetime(test_end_date, format='%Y-%m-%d') - pd.DateOffset(n = round(len(range)*test_size, 0) )
-            test_start_date = min([offset_date, min_test_start_date])
-            #test_start_date = (pd.to_datetime(test_end_date, format='%Y-%m-%d') - offset_date).strftime('%Y-%m-%d')
-        elif (test_start_date==''):
-            test_start_date = min_train_start_date
+        if test_start_date=='':
+            offset_date = pd.to_datetime(max_train_end_date, format='%Y-%m-%d') - pd.DateOffset(n = round(len(range)*test_size, 0) )            
+            test_start_date = offset_date
         else:
             test_start_date = pd.to_datetime(test_start_date, format='%Y-%m-%d')
 
@@ -50,7 +47,7 @@ def define_train_test_set_dates(df, y, train_start_date, train_end_date, test_st
             train_start_date = pd.to_datetime(train_start_date, format='%Y-%m-%d')
 
         if train_end_date=='':
-            train_end_date = max_train_end_date
+            train_end_date = test_start_date - pd.DateOffset(n = 1) 
         else:
             train_end_date = pd.to_datetime(train_end_date, format='%Y-%m-%d')
 

Code/Profiling/Intermittent/intermittent.py

@@ -133,6 +133,15 @@ def get_project_root(Path):
         :return: Path
         """
         return Path(__file__).parent.parent
+    
+    def create_folder_tree(folder_name):                
+        try:
+            os.makedirs(os.path.join(folder_name))
+        except OSError:
+            print("Creation of the directory failed or already present", folder_name)
+        else:
+            print("Successfully created the directory", folder_name)
+        return
 
     def add_daily_date(df):
         """