Resolve #8: Prevent ride clustering k > node count crash

Author: MathBunny

unsupervised/kmeans_rides.py

@@ -4,6 +4,11 @@
 # Data is expected as a,b|c,d|e,f|
 def clusterActivities(data, num_clusters):
     input = data.split('|')
+
+    if data == '':
+        return ''
+    num_clusters = min(len(input), num_clusters)
+
     arr = []
     for point in input:
         pointArr = point.split(',')
@@ -13,4 +18,5 @@ def clusterActivities(data, num_clusters):
     kmeans = KMeans(n_clusters=num_clusters, random_state=0).fit(X)
     return '|'.join(str(v) for v in np.array(kmeans.labels_).tolist())
 
+
 # print clusterActivities('1,1|2,2|4,4|6,6|32132,13231|33443,3432|-322,-32123|-321,-43211', 3)