ml.ALSModel function parity: ALSModel should support recommendforAll

For function parity with MatrixFactorizationModel, ALS model should support API:
recommendUsersForProducts
recommendProductsForUsers

There're another two APIs missing (lower priority):
recommendProducts:
recommendUsers:

The function requirement comes from mailing-list: http://apache-spark-developers-list.1001551.n3.nabble.com/Question-about-using-collaborative-filtering-in-MLlib-td19677.html