Spark Lucenerdd Versions Save

Changelog:

*Update to Spark 2.1.0 *Update to Scalatest 3.x *Add toString method in SparkScoreDoc *Update minor versions on dependencies joda-time and algebird *Fix return type on scripts

Changelog:

• Update Spark version to 2.0.2
• Add moreLikeThis functionality in LuceneRDD

Changelog:

General

• Introduce Versionable trait

Linkage:

• Multithreaded search on each partition (LuceneRDD, ShapeLuceneRDD)
• Remove unnecessary topK monoid reduce
• Response is Array[List[SparkScoreDoc] not List[SparkScoreDoc]

LuceneRDD:

• Add deduplication method dedup

ShapeLuceneRDD

• Add implicitis for DataFrames

Changelog:

• Unpersist broadcast variables on all link methods
• Add .version method on all main RDDs
• Rename GridLoader to PrefixTreeLoader
• Remove multivalued fields in SparkDoc
• Configurable prefix tree in ShapeLuceneRDD (via spatial.prefixtree.name={geohash|quad}

Changelog:

• All used classes are registered under LuceneRDDKryoRegistrator.
• Bump Spark version to 2.0.1

• Fix critical bug on TopKMonoids for LuceneRDD and FacetedLuceneRDD
• Trait SparkScoreDocAggregatable is removed.
• Set TopKMonoid size to topK input parameter (instead of MaxDefaultTopK) in LuceneRDD.link

• Fixed #48