Adds support for Jaccard bag/multiset semantics

[tcsalameh]

This commit adds support for calculating Jaccard similarity using
bag/multiset semantics, as described on pgs. 76-77 in chapter 3 of
Mining of Massive Datasets (MMDS).

MMDS uses the example of movie ratings:

If ratings are 1-to-5-stars, put a movie in a customer's set n times
if they rated the movie n-stars. Then, use Jaccard similarity for bags
when measuring the similarity of customers. The Jaccard similarity for
bags B and C is defined by counting an element n times in the
intersection if n is the minimum of the number of times the element
appears in B and C. In the union, we count the element the sum of the
number of times it appears in B and C.

To fit a model using bag semantics, the user:

• Instantiates an LSH model with the variable repeatedItems set to
  true. This is an optional variable that is false by default.
• Passes their data into the model as a List or RDD of SparseVectors,
  where the indices of each SparseVector correspond to the distinct
  items in the set, and the values of each correspond to the number of
  times each corresponding item is repeated in the set.

The only difference in running the model and getting output is that the
Jaccard similarity between two sets with bag semantics has a maximum of
0.5 rather than 1.0.

[tcsalameh]

Also: I just added a couple lines in the README about the repeatedItems option, but I'm happy to also add an example if needed.