do not fail when the vertex does not contian edges
avoiding to fail with NPE in case of missing relClass(es)
refactor: easier way to build dedup rels
align with latest dnet-actionmanager-common
commented out code that sets a counter because we reach the max num of counters and the job fails
changed WRITE_TO_WAL = true for all jobs writing to HBase tables
dedup experiments
added mapper class for hdfs actions
added anchorStats map-only job
removing useless counters
using most recent dnet-pace-core features
fixed DedupDeleteRelMapper
do not export deleted entities
added utility methods to deal with strings rather than byte[]
log the documents being compared before failing
log the number of items clustered on each key
do not consider deleted entities
updated to dnet-openaire-data-protos:3.5.0-SNAPSHOT
cleanup, extended tests to include new relationships and mapping profiles
added min distance algorithm, used to identify the connected components (dedup)
fixed tests, added new dedup specific jobs
added implementors for offline dedup person workflow