first implementation of the normalization min_max and z_score

imputegap/imputation/_imputation.py

@@ -1,3 +1,5 @@
+import os
+
 import toml
 from imputegap.algorithms.cdrec import native_cdrec_param
 from imputegap.evaluation._evaluation import EvaluationGAP
@@ -13,12 +15,15 @@ def __init__(self):
         """
         self.config = self.load_toml()
 
-    def load_toml(self):
+    def load_toml(self, filepath = "../env/default_values.toml"):
         """
         Load default values of algorithms
         :return: the config of default values
         """
-        with open("../env/default_values.toml", "r") as file:
+        if os.path.exists(filepath) == False:
+            filepath = "./env/default_values.toml"
+
+        with open(filepath, "r") as file:
             config = toml.load(file)
         return config
 

imputegap/manager/_manager.py

@@ -99,7 +99,13 @@ def normalization_min_max(self):
         """
         print("Normalization of the original time series dataset with min/max...")
 
-        data_normalized = (self.ts - self.ts.min()) / (self.ts.max() - self.ts.min())
+        ts_min = self.ts.min(axis=0)  # Min for each series
+        ts_max = self.ts.max(axis=0)  # Max for each series
+
+        range_ts = ts_max - ts_min
+        range_ts[range_ts == 0] = 1  # To avoid division by zero for constant series
+
+        data_normalized = (self.ts - ts_min) / range_ts
         self.normalized_ts = data_normalized
 
         return self.normalized_ts