Release Notes

This is the documentation for Faunus 0.4.
Faunus was merged into Titan and renamed Titan-Hadoop in version 0.5.
Documentation for the latest Titan version is available at http://s3.thinkaurelius.com/docs/titan/current.

Faunus: Graph Analytics Engine
http://faunus.thinkaurelius.com

Faunus 0.×.y

Version 0.4.4 (Apr 22, 2014)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.4.4</version>
</dependency>

Incremented Titan version

Version 0.4.3 (Apr 21, 2014)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.4.3</version>
</dependency>

• Added error handling for invalid RDF parameters
• Added support for incremental edge loading with BlueprintsGraphOutputMapReduce
• Bumped to support Titan 0.4.3

Version 0.4.2 (Jan 9, 2014)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.4.2</version>
</dependency>

• Added Titan ElasticSearch dependency so its available in the Hadoop job jar
• Bumped to support Titan 0.4.2

Version 0.4.1 (Nov 24, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.4.1</version>
</dependency>

• Fixed a severe bug in Configuration entry orderings and MapSequence
• Changed default Cassandra partitioner from Random to Murmur3
• Broke BlueprintsGraphOutputMapReduce into a MapReduce and then a Map (speeds up edge loading)

Version 0.4.0 (Oct 16, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.4.0</version>
</dependency>

• Removed support for Java 1.6 (Java 1.7+ supported)
• Bumped to support Hadoop 1.2.1
• Bumped to support Titan 0.4.0
• Bumped to support TinkerPop 2.4.0
• Renamed FaunusRexsterExtension to FaunusRexsterInputFormatExtension
• Added FaunusRexsterExecutorExtension which allows remote execution of a Faunus script and tracking of its progress
• Global GremlinFaunus variables are now available in ScriptEngine use cases
• Reduced the width of the logger output of Hadoop via a new log4j.properties file
• Simplified ResultHookClosure with new Gremlin 2.4.0 classes
• Fixed has(key,null) NullPointerException issue
• Added support for has(key) and hasNot(key)
• Migrated from Query.Compare to Compare with Blueprints 2.4.0
• The variables hdfs and local are available to gremlin.sh -e
• Removed SequenceFile migration model via Faunus (unsustainable)

---

Version 0.3.2 (August 1, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.3.2</version>
</dependency>

• Fixed a bug in SchemaInferencerMapReduce with EmptyConfiguration
• User can now provide their own mapred.jar to deploy to the Hadoop cluster
• Graph database bulk loading is now fail safe as transaction size is the size of the input split
• Incremental bulk loading is now possible w/ loading-from-scratch=false in BlueprintsGraphOutputMapReduce
• Number of reduce tasks to use during bulk and incremental loading is computed as a function of input split size
• Added more useful PathFilter instances
• Generalized HDFSTools for more generic use throughout Faunus
• Added JobConfigurationFormat which supports dynamic job manipulation at execution time
• Bumped to support Titan 0.3.2
• Fixed a bug in CountMapReduce where a value of 0 was not displayed
• Added FileSystem.cp() to support copying paths within a filesystem

---

Version 0.3.1 (May 21, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.3.1</version>
</dependency>

• Optimized the default in-memory filter for VertexQueryFilter
• Bumped to support Titan 0.3.1
• Updated all ByteBuffers to StaticByteBuffers (and associated updates)

---

Version 0.3.0 (April 20, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.3.0</version>
</dependency>

• Skipped Faunus 0.2.y line in order to synchronize with Titan 0.3.z
• Bumped to support Titan 0.3.0
• Bumped to support Hadoop 1.1.2
• Moved IdentityMap to base directory as changed in Pipes 2.3.0
• Refactored the means by which MapReduce test cases are written
• FaunusVertex property value serialization support using Kryo 2.2.1
• FaunusVertex01 provides history deserialization support for Faunus 0.1.z → Faunus 0.3.z conversion
• Moved util.MicroElement classes into respective FaunusElement inner classes
• Added numerous test cases for file-based RecordReader formats
• Added ScriptInputFormat to support the arbitrary reading of an HDFS file lines into a FaunusVertex
• Added ScriptOutputFormat to support the arbitrary writing of an HDFS file line from a FaunusVertex
• Introduced the faunus.pipeline property namespace for properties that are global to the pipeline
• Added support for vertex-centric “filters” for isolating subsets of a graph InputFormat
• Fixed a severe bug in the propagation of Configurations throughout a Faunus job sequence (EmptyConfiguration)
• Renamed GraphSONUtility to FaunusGraphSONUtility to avoid name conflicts with Blueprints
• Added EdgeCopyMapReduce to support graph transpose for bidirectional edges
• FaunusGraph now implements Configurable and thus getConf() instead of getConfiguration()
• Removed unnecessary combiner steps when the combiner only groups messages without compressing them

---

Version 0.1.1 (March 22, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.1.1</version>
</dependency>

• OrderMapReduce semantics updated for number based properties (multiply by path count for efficiency)
• Improved the memory consumption and speed of MemoryMapper
• Bumped to support Hadoop 1.1.1
• Bumped to support Titan 0.2.1
• Bumped to support OpenRDF 2.6.10
• Bumped to support TinkerPop 2.3.0
• Added script-step which allows for arbitrary Gremlin/Groovy script executions on Faunus vertices
• Removed step-step as it was not clearly defined and difficult to use (see script-step)
• Fixed job jar finding issue by introducing FAUNUS_HOME system property
• Complete refactoring of FaunusCompiler and FaunusPipeline
• Added support for user defined Input- and OutputFormats
• Added MapReduceFormat which allows an arbitrary number of MapReduce jobs after/before input/output
• Added configuration to allow Faunus to get the actual vertex count from Rexster when determining split size
• Added configuration to allow Faunus to authenticate to Rexster, if Rexster security is turned on
• Removed dependency on Tokens.F with the inclusion of ordering in TinkerPop Pipes

---

Version 0.1.0 (January 14, 2013)

<dependency>
  <groupId>com.thinkaurelius.faunus</groupId>
  <artifactId>faunus</artifactId>
  <version>0.1.0</version>
</dependency>

• First release and therefore, no changes.