autopilot: shuffle set of candidates using Fisher–Yates before selecting

In this commit we modify the ConstrainedPrefAttachment.Select method to
first shuffle the set of potential candidates before selecting them.
This serves to remove the existing grouping between candidates which
may have influenced the selection.

autopilot/prefattach.go

@@ -107,6 +107,21 @@ func NewNodeID(pub *btcec.PublicKey) NodeID {
 	return n
 }
 
+// shuffleCandidates shuffles the set of candidate nodes for preferential
+// attachment in order to break any ordering already enforced by the sorted
+// order of the public key for each node. To shuffle the set of candidates, we
+// use a version of the Fisher–Yates shuffle algorithm.
+func shuffleCandidates(candidates []Node) []Node {
+	shuffledNodes := make([]Node, len(candidates))
+	perm := prand.Perm(len(candidates))
+
+	for i, v := range perm {
+		shuffledNodes[v] = candidates[i]
+	}
+
+	return shuffledNodes
+}
+
 // Select returns a candidate set of attachment directives that should be
 // executed based on the current internal state, the state of the channel
 // graph, the set of nodes we should exclude, and the amount of funds
@@ -208,8 +223,9 @@ func (p *ConstrainedPrefAttachment) Select(self *btcec.PublicKey, g ChannelGraph
 		// Given our selection slice, we'll now generate a random index
 		// into this slice. The node we select will be recommended by
 		// us to create a channel to.
-		selectedIndex := prand.Int31n(int32(len(selectionSlice)))
-		selectedNode := selectionSlice[selectedIndex]
+		candidates := shuffleCandidates(selectionSlice)
+		selectedIndex := prand.Int31n(int32(len(candidates)))
+		selectedNode := candidates[selectedIndex]
 
 		// TODO(roasbeef): cap on num channels to same participant?