[distsim] feature request - scope limitation by top N terms

[gilnoh]

(It is possible that the feature is already there but unknown to me. But just in case)

Lexical resource - For evaluation and efficiency purpose, it is often convenient to limit the scope of knowledge resource. Something like top-10k terms from the corpus, etc.

However, currently, the way to limit the terms (elements) are to set min-count. This works the same way, but it would be nice if we can limit the terms by giving "top 10k frequent terms" that appeared in the corpus, or something like that. This will produce (a roughly) predictable size of lexical resource, with known set of lexical terms.