Pocket-dimension

Added the data quality parameters, min_n_features & min_n_observations, to the initialization of TFIDFVectorizer class. This should have been done in v0.1.3, but was overlooked. Also added a test for TFIDFVectorizer when using min_n_features and/or min_n_observations since this was missed in the previous release.

Added new data quality checks, min_n_features and min_n_observations. If, after filtering of the features, there aren't at least min_n_features features that have a total of min_n_observations counts, then no vector is created.

For the vectorizers, if a record has all its features filtered out, then it would simply return a vector of all 0.0. But when we normalized, it would turn that into a vector of np.nan. Now we just don't return the vector at all and its corresponding id.

Added a section in the test_tf_filtered() to check that this is the expected behavior.

Also fixed a bug with the distributional J-L optimal delta. If you gave a sparse_dim with 2**31 -1 it returned an np.ndarray , whose shape was (1,), and not a float.

Expecting this to be reasonably stable. The code has been split into two modules

* random_projection - This extends sklearn.random_projection.BaseRandomProjection class so it works like sklearn with fit, transform
* vectorizer - This holds the TFVectorizer and TFIDFVectorizer that use the JustInTimeRandomProjection

Initial release