Connected Components: replace random tip in Entrypoint with random tip from largest component

[alon-e]

Currently, Entrypoint is chosen by first uniformly selecting a random tip from all available tips and then walking back until X cumulative weight is reached.

In #86 we explored a few issues with this approach and offered solutions. solution 2.ii. (Connected components) was chosen, as a first step, as a way to differentiate tips.

To enable this feature we tie #88 #89 #90 together: instead of starting with https://github.com/iotaledger/cliri/blob/5640411b17825ad4ebdef7427ec713d6649ca1d6/src/main/java/com/iota/iri/service/tipselection/impl/EntryPointSelectorCumulativeWeightThreshold.java#L43 we uniformly select a random tip from the largest connected component of the recent transactions (the latest X transactions).

pseudo code:

(in EntryPoint)
latestTips = tipsViewModel.getLatestSolidTips(M)
tip = ConnectedComponentsCalculator.randomlySelectTip(latestTips)


func ConnectedComponentsCalculator.randomlySelectTip(tips):
  latestTransactions = findNMostRecentTransactions(tips)
  connectedComponents = getConnectedComponents(latestTransactions)
  return randomlySelectTipFromLargestConnectedComponent(connectedComponents, tips)