CompositeIndicator now also handles functions defined with arrays as …

…input. This allows the user to pass a prexisting numpy function such as average

README.md

@@ -163,7 +163,7 @@ Let's create an indicator that averages Innovation Potential, Mobility Inmpact,
 
 ```
 from toolbox import Handler, CompositeIndicator
-from random_indicator import RandomIndicator
+from examples import RandomIndicator
 
 def innovation_average(indicator_values):
     avg = (indicator_values['Innovation Potential']+indicator_values['Mobility Impact']+indicator_values['Economic Impact'])/3
@@ -175,14 +175,27 @@ avg_I = CompositeIndicator(innovation_average,name='Composite')
 H.add_indicators([R,avg_I])
 ```
 
+You can also pass it a pre-existing function, such as `np.mean`. 
+```
+from toolbox import Handler, CompositeIndicator
+from examples import RandomIndicator
+import numpy as np
+
+H = Handler('corktown')
+R = RandomIndicator()
+avg_I = CompositeIndicator(np.mean,selected_indicators=['Innovation Potential','Mobility Impact','Economic Impact'],name='Composite')
+H.add_indicators([R,avg_I])
+```
+
+
 ## Custom Composite indicator
 
 Let's create an indicator that averages Innovation Potential, Mobility Inmpact, and Economic Impact.
 First, we load the RandomIndicator and pass it to a Handler.
 
 ```
 from toolbox import Handler, CompositeIndicator
-from random_indicator import RandomIndicator
+from examples import RandomIndicator
 
 H = Handler('corktown')
 R = RandomIndicator()

examples.py

@@ -58,7 +58,6 @@ def return_indicator(self, geogrid_data):
 
 
 from numpy import mean
-from numpy.random import random
 class Noise(Indicator):
     '''
     Example of Noise heatmap indicator for points centered in each grid cell.
@@ -84,7 +83,7 @@ def return_indicator(self, geogrid_data):
             lat,lon = zip(*cell['geometry']['coordinates'][0])
             lat,lon = mean(lat),mean(lon)
             feature['geometry'] = {'coordinates': [lat,lon],'type': 'Point'}
-            feature['properties'] = {self.name:random()}
+            feature['properties'] = {self.name:random.random()}
             features.append(feature)
         out = {'type':'FeatureCollection','features':features}
         return out

toolbox.py

@@ -3,6 +3,7 @@
 import json
 import Geohash
 import joblib
+import numpy as np
 from warnings import warn
 from time import sleep
 from collections import defaultdict
@@ -693,10 +694,18 @@ def load_module(self):
 
 
 class CompositeIndicator(Indicator):
-    def setup(self,compose_function,*args,**kwargs):
-        self.compose_function = compose_function
-        self.is_composite = True
-
-    def return_indicator(self, indicator_values):
-        value = self.compose_function(indicator_values)
-        return [{'name': self.name, 'value': value, 'viz_type': self.viz_type}]
+	def setup(self,compose_function,selected_indicators=[],*args,**kwargs):
+		self.compose_function = compose_function
+		self.is_composite = True
+		self.selected_indicators = selected_indicators
+
+	def return_indicator(self, indicator_values):
+		if len(self.selected_indicators)!=0:
+			indicator_values = {k:indicator_values[k] for k in indicator_values if k in self.selected_indicators}
+		try:
+			value = self.compose_function(indicator_values)
+		except:
+			indicator_values = np.array([v for v in indicator_values.values()])
+			value = self.compose_function(indicator_values)
+		return [{'name': self.name, 'value': value, 'viz_type': self.viz_type}]
+