Precalculate DTM or word frequency in the backend as the corpus loaded

[milysun]

DTM or frequency list does not change for a certain corpus, unless the vectoriser is defined differently.
Pre-calculating this and associate with the corpus DF will significantly improve performance.

[sebhaan]

PR Australian-Text-Analytics-Platform/ldaca-wordflow#3 makes DTM pre-calculation computationally viable. However, it doesn't implement the full caching/pre-calculation system yet.