casio-ka/Precomputation-Queue

Test for Senior Java Developer Precomputation Queue

Introduction

You are going to asynchronously precompute some statistics about product sets. Whenever a product set is scheduled for precomputation, it produces a new PrecomputationTask which is inserted into the queue. There are several precomputation nodes which should endlessly poll queue for new tasks and consume them by running a given precomputation task.

Problem statement

• Finish implementation of PrecomputationNode and PrecomputationNodeManager. Precomputation should be done in the background on several PrecomputationNodes. Exact amount of nodes is defined by nodesAmount parameter in the startPrecomputationNodes method.

• Finish implementation of the PrecomputationQueue class. Precomputation tasks should be always processed in order, prioritized by cause:

  1. ProductSet which was computed at least once
  2. ProductSet which was never computed

  If both product sets have the same cause, you should break ties by:

  1. Importance of a product set - lower value means that a product set should be processed earlier.
  2. If importance is also equal, you should follow the order of scheduling (First In, First Out).

Good Luck!

Hints

• Consider thread safety.
• Producer and consumer share queue as a buffer, but you don't have to worry about the buffer size, because you can assume that consuming can be scaled-up to be fast enough, as the number of consumers is configurable.