Add load/save functionality

[cfregly]

required for sharing the model between training and serving layers.

[cfregly]

refactored the code in this repo - and combined code from this pull request - and got this all working in a more production-like env with Spark Streaming.

here are the relevant links:

save/load code:
https://github.com/fluxcapacitor/pipeline/blob/master/myapps/spark/streaming/src/main/scala/com/advancedspark/streaming/ml/incremental/LatentMatrixFactorization.scala

use of the above code in a Spark Streaming app pulling from Kafka:
https://github.com/fluxcapacitor/pipeline/blob/master/myapps/spark/streaming/src/main/scala/com/advancedspark/streaming/rating/ml/TrainMFIncremental.scala

the save/load code needs to be updated to use parquet similar to the rest of the Spark ML models, but this is a huge step forward.

thanks @brkyvz (this repo) and @jjyycchh (save/load pull request).

closing this and moving future development to the new repo @ https://github.com/fluxcapacitor/pipeline