Refactor.
Now difficulty and replays are calculated at startup for 3 solvers. The 4th solver is excluded due to startup time constraints.
Could we remove these calculations for all words from startup, but take it from cache. The cache is empty at startup. 
Once a word is first returned by API, the calculations for all 4 algorithms are started for the word is a separate thread, 
and then calculated results are stored to the cache. API returns replays and difficulty values from the cache. 
