Better tokening in the output json data format for bridge reachability visualisation

Currently we use the bridge fingerprint as a unique key in the bridge reachability timeline graph. This leads to different bridges categorised as a certain pluggable transport type, since it's possible for a bridge with the same fingerprint to run on multiple pluggable transports.

I see to possible ways to avoid this:

1. Use as key H(bridge_fingerprint+pluggable_transport_type)

2. Change the logic for the graphs so that they don't use as token just the fingerprint, but also the transport type that should be present in the output json document.