Skip to content

Releases: visallo/vertexium

v3.2.2

23 Mar 18:09
Compare
Choose a tag to compare
  • Fixed: Improved support for multithreaded clients with InMemoryGraph

v3.2.1

13 Mar 03:56
Compare
Choose a tag to compare
  • Changed: When saving an ExistingElementMutation, the Elastisearch5Index will now apply the mutations using a painless script rather than making multiple requests.
  • Changed: When using Accumulo, all threads now share a single batch writer rather than creating a new writer for every thread. This allows client programs to more effectively use multi-threading.
  • Fixed: Prevent concurrent modification exceptions when using InMemoryGraph through synchronizing the methods and also returning Collection copies.
  • Fixed: The InputStream for a StreamingPropertyValue stored in Accumulo now supports mark/reset.
  • Fixed: Elasticsearch sort throwing errors for String properties while querying across indices

v3.2.0

01 Mar 13:49
Compare
Choose a tag to compare
  • Changed: Elasticsearch _id and _type fields to be much smaller to minimize the size of the _uid field data cache
  • Changed: Elasticsearch to only refresh the indices that were changed
  • Fixed: InMemory and Accumulo fix wrong vertex being returned after re-adding the same vertex with a different visibility after soft delete
  • Fixed: Find path when edge labels are deflated
  • Fixed: Accumulo stream property value in table data length of reference
  • Fixed: Query will now return hidden elements/vertices/edges if FetchHint.INCLUDE_HIDDEN is passed
  • Fixed: Query.hasAuthorization will now match elements whose only use of an authorization is that the element is hidden
  • Fixed: Query.hasAuthorization will now match elements whose only use of an authorization is a hidden property
  • Fixed: Elasticsearch server plugin missing fields while querying across indices
  • Fixed: Elasticsearch shard exception if geo property is not defined
  • Added: Query methods elementIds/vertexIds/edgeIds are overloaded to accept IdFetchHint, which makes it possible to include ids for hidden elements
  • Added: Graph.getExtendedDataInRange in order to bulk load ranges of extended data rows
  • Added: SearchIndex.addExtendedData to allow for reindexing extended data
  • Added: Elasticsearch exception if any shard failures occur on queries
  • Added: Elasticsearch option to force disable the use of the server side plugin

v3.1.1

13 Feb 01:22
Compare
Choose a tag to compare
  • Added: Graph.getExtendedData to get a single extended data row
  • Fixed: InMemory update extended data with different value

v3.1.0

12 Feb 01:54
Compare
Choose a tag to compare
  • Changed: Set Elasticsearch scroll api query to have a size that is the same as page size
  • Changed: Properly throw NoSuchElementException from Iterables when next is called with no more elements
  • Changed: If exact match is chosen for a property, full text is automatically enabled to support aggregation results
  • Added: Accumulo in table storage of streaming property value data
  • Added: Functions to delete extended data columns
  • Added: Multi-value extended data columns
  • Added: Sorting and aggregating of row ids and table names
  • Added: Option to execute a script from the shell
  • Fixed: Streaming property value input stream from data table
  • Fixed: Elasticsearch5 retry logic to sleep per request failure instead of batch of failures
  • Fixed: Mixed case searching on exact match properties
  • Fixed: Intermittent bug while indexing streaming property values under high load
  • Removed: ElasticSearch 1.x support

v3.0.4

12 Feb 01:50
Compare
Choose a tag to compare
  • Fixed: Infinite loop in PagingIterable if Elasticsearch returned vertices that wasn't saved in the database yet
  • Fixed: Pagination returning duplicate ids on different pages by adding default sort by score and then id. If sorts are specified, sort by score and then id after specified sorts. If sorts are not specified, sort by score and then id.
  • Fixed: Elasticsearch flush overflow causing excessive exceptions
  • Fixed: Fix Elasticsearch5 missing field exception when sorting on multiple indices

v3.0.3

07 Dec 20:09
Compare
Choose a tag to compare
  • Changed: Accumulo default data storage to use in table storage and not overflow to HDFS
  • Added: Added a hasAuthorization method to the Query class to allow searches for any element that uses an authorization string or strings.
  • Added: Query extended data on an element
  • Added: EmptyResultsGraphQuery and EmptyResultsQueryResultsIterable
  • Fixed: GeoPoint.distanceBetween and GeoPoint.calculateCenter calculations
  • Fixed: Elasticsearch5 field limit of 1000
  • Fixed: Query.hasId to work with extended data rows
  • Removed: CompositeGraphQuery

v3.0.2

28 Nov 14:30
Compare
Choose a tag to compare
  • Fixed: Elasticsearch5 throwing unsupported operation exception when adding/updating a property
  • Added: Added a has method to the Query class to allow searches on all properties of a particular data type.
  • Added: Added a has method to the Query class to allow searches for any elements with/without a property of a particular data type.
  • Added: Added a has method to the Query class to allow searches for a value across multiple properties.
  • Added: Added a has method to the Query class to allow searches for a elements with a list of properties. The presence of any property in the list will cause the document to match.
  • Added: Added a has method to the Query class to allow searches for a elements without a list of properties. The absence of all properties in the list will cause the document to match.
  • Added: Implemented support all compare operators for DateOnly field types when using Elasticsearch 5.
  • Fixed: Queries with both a query string and aggregations were throwing a "not implemented" exception in the Elasticsearch 5 plugin.

v3.0.1

08 Nov 12:07
Compare
Choose a tag to compare
  • Changed: Find Path to not return a path that contains one or more vertices that is can't be retrieved because of visibility restrictions
  • Changed: Reduced DefaultIndexSelectionStrategy cache load time from 1hr to 5min
  • Added: Added a hasId method to the Query class to allow searches to be filtered by element ID.
  • Fix: Extended data element type value for edges
  • Changed: Field removal from Elasticsearch documents is now queued as a future instead of immediate
  • Fix: Marking vertices/edges as hidden will now update the document in the search index as well as the data store
  • Fix: Fixed an infinite looping problem in the PagingIterable that may have resulted from a malformed ES document
  • Fix: Elasticsearch5SearchIndex.removeFieldsFromDocument will not run the removal script in Elasticsearch if there are no properties to remove
  • Fix: Precision and error are now configurable when mapping GeoShape properties. This allows an adjustable trade off between index performance and false positive search results.

v3.0.0

03 Oct 21:18
Compare
Choose a tag to compare
  • Changed: Removed ES 2 support and replaced it with ES 5 support
  • Changed: Removed support for NameSubstitutionStrategy from the Elasticsearch modules. NameSubstitutionStrategy is still available for Accumulo modules
  • Changed: Added property checking when using the has method to query the graph for a value. An exception will now be thrown if an attempt is made to either query a property that does not exist or to text query a property that isn't full text indexed
  • Note: Upgrading to this version will require re-indexing if you use NameSubstitutionStrategy or are switching to the ES 5 module.
  • Added: Support for extended GeoShape storage and search. Currently the new shapes are only supported by the ElasticSearch 5 module.
  • Added: GraphVisitor to visit elements, properties, extended data rows using a visitor pattern