We have added an API call for retrieving individual ensembles

[msouff]

-Useful when interested in a specific ensemble or in analyzing the accuracy of each individual ensemble.

Other small changes:

-A small bug fix where the graph would not display if the high res forecast was not in store.

-Darken the return periods colors.

-Changed warning naming to awareness

[pythonweb2]

@msouff How are you calculating the probabilities? From what I understand, you should be creating a binary array (1 if exceeds return period, 0 if doesn't) and then take the average of the binary array (so if there are 4 out of 5 that exceed return period, the probability is 4/5 or 80%).