Agnostic for input type

Now works for scalars, arrays and dataframes

decomposition.py

@@ -1,27 +1,33 @@
 import numpy as np
 import itertools
 
-def decomposition(n, values_start, values_end):
-
-    # Make sure lists are numpy arrays
-    values_start = np.array(values_start)
-    values_end = np.array(values_end)
+def prod(lst):
+    value = 1.0
+    for x in lst:
+        value *= x
+    return value
 
+def decomposition(n, values_start, values_end):
+
     if len(values_start) != n or len(values_end) != n:
         raise Exception("Input lists are not of length {}".format(n))
-
-    diffs = values_end - values_start
-
+
+    try:
+        diffs = [v_end - v_start for v_start, v_end in zip(values_start, values_end)]
+    except:
+        raise Exception("Cannot substract values_start and values_end.")
+
     all_indices = set(range(n))
-
+
+    # Initialise with array of zeros
     effect = [0]*n
-
+    
     # For each factor i:
     for i in range(n):
-
+        
         # Remaining indices
         indices = all_indices - set([i])
-
+        
         # For each level 1 to n (incl)
         for level in range(1,n+1):
             num_start = n - level
@@ -34,22 +40,18 @@ def decomposition(n, values_start, values_end):
 
                 indices_start = list(indices_start) # Transform tuple to list
                 indices_diff = list(all_indices - set(indices_start))
-
-                contribution_level = (1.0/level) * np.prod(values_start[indices_start]) * np.prod(diffs[indices_diff])
+
+                selection_start = [values_start[i] for i in indices_start]
+                selection_diff = [diffs[i] for i in indices_diff]
+                contribution_level = (1.0/level) * prod(selection_start) * prod(selection_diff)
 
                 effect[i] += contribution_level
-
-    value_start = np.prod(values_start)
-    value_end = np.prod(values_end)
 
-    np.testing.assert_almost_equal(np.sum(effect), value_end - value_start)
+    value_start = prod(values_start)
+    value_end = prod(values_end)
 
     return {
-        'value_start': np.prod(values_start),
-        'value_end': np.prod(values_end),
+        'value_start': prod(values_start),
+        'value_end': prod(values_end),
         'individual_effects': effect
     }
-
-
-
-