Overview

The GraphDB Connectors provide extremely fast normal and facet (aggregation) searches that are typically implemented by an external component or service such as Elasticsearch, but have the additional benefit to stay automatically up-to-date with the GraphDB repository data.

The Connectors provide synchronisation at the entity level, where an entity is defined as having a unique identifier (a URI) and a set of properties and property values. In terms of RDF, this corresponds to a set of triples that have the same subject. In addition to simple properties (defined by a single triple), the Connectors support property chains. A property chain is defined as a sequence of triples where each triple's object is the subject of the following triple.

Features

The main features of the GraphDB Connectors are:

• maintaining an index that is always in sync with the data stored in GraphDB;
• multiple independent instances per repository;
• the entities for synchronisation are defined by:
  • a list of fields (on the Elasticsearch side) and property chains (on the GraphDB side) whose values will be synchronised;
  • a list of rdf:type's of the entities for synchronisation;
  • a list of languages for synchronisation (the default is all languages);
  • additional filtering by property and value.
• full-text search using native Elasticsearch queries;
• snippet extraction: highlighting of search terms in the search result;
• faceted search;
• sorting by any preconfigured field;
• paging of results using offset and limit;
• custom mapping of RDF types to Elasticsearch types;

Each feature is described in detail below.

Sample data

All examples use the following sample data, which describes five fictitious wines: Yoyowine, Franvino, Noirette, Blanquito and Rozova, as well as the grape varieties needed to make these wines. The minimum required ruleset level in GraphDB is RDFS.

Usage

All interactions with the Elasticsearch GraphDB Connector shall be done through SPARQL queries.

There are three types of SPARQL queries:

• INSERT for creating and deleting connectors;
• SELECT for listing connectors and querying connector configuration parameters;
• INSERT/SELECT for storing and querying data as part of the normal GraphDB data workflow.

In general this corresponds to INSERT adds or modifies data and SELECT queries existing data.

Each connector implementation defines its own URI prefix to distinguish it from other connectors. For the Elasticsearch GraphDB Connector, this is *http://www.ontotext.com/connectors/elasticsearch#*. Each command or predicate executed by the connector uses this prefix, e.g. <http://www.ontotext.com/connectors/elasticsearch##createConnector> to create a connector for Elasticsearch.

Individual instances of a connector are distinguished by unique names that are also URIs. They have their own prefix to avoid clashing with any of the command predicates. For Elasticsearch, the instance prefix is http://www.ontotext.com/connectors/elasticsearch/instance#.

Using a connector with a GraphDB cluster

This release introduces support for Elasticsearch connectors in a GraphDB cluster. The connectors require a transactional entity pool, which is off by default. Please, refer to [GraphDB Entity Pool] to enable the transactional entity pool.

Creating a connector

Creating a connector is done by sending a SPARQL query with the following configuration data:

• the name of the connector (e.g. my_index);
• classes to synchronise;
• properties to synchronise.

The configuration data has to be provided as a JSON string representation and passed together with the create command.

The create command is triggered by a SPARQL INSERT with the createConnector predicate, e.g. this will create a connector called my_index that will synchronise the wines from the sample data above:

The above command will create a new Elasticsearch connector that will connect to the Elasticsearch instance accessible at port 9300 on the localhost as specified by the "elasticsearchUrl" key.

The "types" key defines the RDF type of the entities to synchronise and in the example it is only entities of the type <http://www.ontotext.com/example/wine#Wine> (and its subtypes). The "fields" key defines the mapping from RDF to Elasticsearch. The basic building block is the property chain, i.e. a sequence of RDF properties where the object of each property is the subject of the following property. In the example we map three bits of information - the wine's grape, sugar content, and year. Each chain is assigned a short and convenient field name: "grape", "sugar", and "year". The field names are later used in the queries.

Grape is an example of a property chain composed of more than one property. First we take the wine's madeFromGrape property, the object of which is an instance of type Grape, and then we take the rdfs:label of this instance. Sugar and year are both composed of a single property that links the value directly to the wine.

Schema and core management

By default GraphDB will manage (create, delete or update if needed) the Solr core and the Solr schema. This makes it easier to use Solr as everything will be done automatically. This behaviour can be changed by the following options:

• manageIndex: if true, GraphDB will manage the index. True by default.
• manageSchema: if true, GraphDB will manage the schema. True by default.

Note that if either of the options is set to false you will be responsible for creating, updating or removing the core/schema and the connector will not function correctly if you misconfigured Elasticsearch.

Using a non-managed schema

The present version of the connectors provides no support for changing some advanced options, such as stopwords, on a per field basis. The recommended way to do that for now is to manage the mapping yourself and tell the connector to just sync the object values in the appropriate fields. Here is an example:

this will create the same connector as above but it expects fields with the specified fieldnames to be already present in the index mapping, as well as some internal GraphDB fields. For the example you must have the following fields:

_graphdb_id and _chains are used internally by GraphDB and are always required.

Dropping a connector

Dropping a connector removes all references to its external store from GraphDB as well as the Sorl core associated with it. Dropping a connector is achieved through a SPARQL INSERT query with the following parameter:

• Name of the connector

The drop command is triggered by a SPARQL INSERT with the dropConnector predicate, e.g. this will remove the connector :my_index:

Listing available connectors

Listing connectors returns all previously created connectors. It is a SELECT query with the listConnectors predicate:

?cntUri will be bound to the prefixed URI of the connector that was used during creation, e.g. <http://www.ontotext.com/connectors/elasticsearch/instance#my_index>, while ?cntStr will be bound to a string, representing the part after the prefix, e.g. "my_index".

Status check

The internal state of each connector can be queried using a SELECT query and the connectorStatus predicate:

?cntUri will be bound to the connector prefixed URI, while ?cntStatus will be bound to a string representation of the status of the connector represented by this URI. The status is key-value based.

Adding, updating and deleting data

From the user's point of view all synchronisation will happen transparently without using any additional predicates or naming a specific store explicitly, i.e. the user should simply execute standard SPARQL INSERT/DELETE queries. This is achieved by intercepting all changes in the plugin and determining which abstract documents need to be updated.

Querying data

Once a connector has been created it will be possible to query data from it through SPARQL. For each matching abstract document, the connector returns the document's subject. In its simplest form querying is achieved by using a SELECT and providing the Elasticsearch query as the object of the :query predicate:

The result will bind ?entity to the two wines made from grapes that have "cabernet" in their name, namely :Yoyowine and :Franvino.

Note that you must use the field names you chose when you created the connector. It is perfectly valid to have field names identical to the property URIs but then you responsible for escaping any special characters according to what Elasticsearch expects.

First we get an instance of the requested connector by using the RDF notation "X a Y" (= X rdf:type Y), where X is a variable and Y is a connector. X will be bound to an instance of this connector. Then we assign a query to that instance by using the system predicate :query. Finally we request the matching entities through the :entities predicate.

It is also possible to provide per query search options by using one or more option predicates. The option predicates are described in details below.

Raw queries

If you want to access a Elasticsearch query parameter that is not exposed through a special predicate you can do it with a raw query. Instead of providing a full text query in the :query part you specify raw Elasticsearch parameters. For example, if you want to boost some parts of your full text query as described here you can use the following query:

Combining Elasticsearch results with GraphDB data

The bound ?entity can be used in other SPARQL triples in order to build complex queries that fetch additional data from GraphDB. For example to see the actual grapes in the matching wines as well as the year they were made:

The result will look like this:

Note that :Franvino is returned twice because it is made from two different grapes, both of which are returned.

Entity match score

It is possible to access the match score returned by Elasticsearch with the :score predicate. As each entity has its own score, the predicate must come at the entity level. For example:

The result will look like this but the actual score might be different as it depends on the specific Elasticsearch version:

Basic faceting

Consider the sample wine data and the my_index connector described previously. We can use the same connector to query facets too:

It is important to specify the fields we want to facet by using the facetFields predicate. Its value must be a simple comma-delimited list of field names. In order to get the faceted results, we have to use the facets predicate and as each facet has three components (name, value and count), the facets predicate binds a blank node, which in turn can be used to access the individual values for each component through the predicates facetName, facetValue, and facetCount.

The resulting bindings will look like in the table below:

We can easily see that there are three wines produced in 2012 and two in 2013. We also see that three of the wines are dry, while two are medium. However, it is not necessarily true that the three wines produced in 2012 are the same as the three dry wines as each facet is computed independently.

Advanced faceting and aggregations

While basic faceting allows for simple counting of documents based on the discrete values of a particular field, there are more complex faceted or aggregation searches in Elasticsearch. The connector provides a mapping from Elasticsearch results to RDF results but no mechanism for specifying the queries other than executing a raw query.

Supported Elasticsearch facets and aggregations

The Elasticsearch connector supports mapping of range, interval and pivot facets. Please refer to the documentation of Solr for more information.

RDF mapping of the results

The results are accessed through the predicate :aggregations (much like the basic facets are accessed through :facets). The predicate will bind multiple blank nodes that each contain a single aggregation bucket. The individual bucket items can be accessed through these predicates:

Sorting

It is possible to sort the entities returned by a connector query according to one or more fields. In order to be able to use a certain field for sorting, you have to specify this at the time of creating the connector instance. Sorting is achieved by the orderBy predicate the value of which must be a comma-delimited list of fields. Each field may be prefixed with a minus to indicate sorting in descending order. For example:

The result will contain wines produced in 2013 sorted according to their sugar content in descending order:

By default, entities are sorted according to their matching score in descending order.

Note that if you join the entity from the connector query to other triples stored in GraphDB, GraphDB might scramble the order. To remedy this, use ORDER BY from SPARQL.

Limit and offset

Limit and offset are supported on the Elasticsearch side of the query. This is achieved through the predicates limit and offset. Consider this example in which we specify an offset of 1 and a limit of 1:

The result will contain a single wine, Franvino, as it would be second in the list if we executed the query without the limit and offset:

Note that the specific order in which GraphDB returns the results, depends on both how Elasticsearch returns the matches, unless you specified sorting.

Snippet extraction

Snippet extraction is used to extract highlighted snippets of text that match the query. The snippets are accessed through the dedicated predicate :snippets, which binds a blank node that in in turn provides the actual snippets via the predicates :snippetField and :snippetText. The predicate :snippets must be attached to the entity, as each entity has a different set of snippets. For example, in a search for Cabernet:

The query will return the two wines made from Cabernet Sauvignon or Cabernet Franc grapes as well as the respective matching fields and snippets:

Note that the actual snippets might be somewhat different as this depends on the specific Elasticsearch implementation.

It is possible to tweak how the snippets are collected/composed by using the following option predicates:

• :snippetSize sets the maximum size of the extracted text fragment, 250 by default.
• :snippetSpanOpen text to insert before the highlighted text, <em> by default.
• :snippetSpanClose text to insert after the highlighted text, </em> by default.

The option predicates are set on the connector instance, much like the :query predicate.

Total hits

You can get the total number of hits by using the :totalHits predicate, e.g. for the connector :my_index and a query that would retrieve all wines made in 2012:

As there are three wines made in 2012, the value 3 (of type xdd:long) will be bound to ?totalHits.

Creation parameters

The creation parameters define how a connector instance is created by the :createConnector predicate. There are some required parameters and some that are optional. All parameters are provided together in a JSON object, where the parameter names are the object keys. Parameter values may be simple JSON values such as a string or a boolean, or they can be lists or objects.

All of the creation parameters can also be set conveniently from the Create Connector user interface in the GraphDB Workbench without any knowledge of JSON.

Elasticsearch instance to sync to: elasticsearchNode (string)

Since Elasticsearch is a third-party service, you have to specify the node where it is running. The format of the node value is of the form hostname.domain:port. There is no default value.

Types of entities to sync: types (list of URI)

The RDF types of entities to sync are specified as a list of URIs. At least one type URI must be provided.

What exactly to sync: fields (list of field object)

The fields define exactly what parts of each entity will be synchronised as well as the specific details on the connector side. The field is the smallest synchronisation unit and it maps a property chain from GraphDB to a field in Elasticsearch. The fields are specified as a list of field objects. At least one field object must be provided. Each field object has further keys that specify details.

Name of the field: fieldName (string)

The name of the field defines the mapping on the connector side. It is specified by the key fieldName with a string value. The field name is used at query time to refer to the field. There are few restrictions on the allowed characters in a field name but to avoid unnecessary escaping (which depends on how Elasticsearch parses its queries) we recommend to keep the field names simple.

Property chain to map: propertyChain (list of URI)

The property chain (propertyChain) defines the mapping on the GraphDB side. A property chain is defined as a sequence of triples where the entity URI is the subject of the first triple, its object is the subject of the next triple and so on. In this model, a property chain with a single element corresponds to a direct property defined by a single triple. Property chains are specified as a list of URIs and at least one URI must be provided. If you need to store the entity URI in the connector, you may map it by defining a property chain with a single special URI: $self. Only one field per connector may use the $self notation.

The default value: defaultValue (string)

The default value (defaultValue) provides means for specifying a default value for the field when the property chain has no matching values in GraphDB. The default value can be a plain literal, a literal with a datatype (xsd: prefix supported), a literal with language or a URI. It has no default value.

Indexing the field: indexed (boolean)

Fields are indexed by default but that can be changed by using the Boolean option "indexed". True by default.

If true this option corresponds to "index" = "analyzed" or "not_analyzed". If false it corresponds to "index" = "no".

Storing the field: stored (boolean)

Fields are stored in Elasticsearch by default but that can be changed by using the Boolean option "stored". Stored fields are required for retrieving snippets. True by default.

This option corresponds to the property "store" in the Elasticsearch mapping.

Skipping the analyser: analyzed (boolean)

When literal fields are indexed in Elasticsearch, they will be analysed according to the analyser settings. Should you require that a given field is not analysed you may use "analyzed". This option has no effect for URIs (they are never analysed). True by default.

If true this option corresponds to "index" = "analyzed" in the Elasticsearch schema. If false it corresponds to "index" = "not_analyzed".

Multivalued fields: multivalued (boolean)

RDF propreties and synchronised fields may have more than one value. If "multivalued" is set to true, all values will be synchronised to Elasticsearch. If set to false, only a single value will be synchronised. True by default.

Automatic datatype mapping

The connector will map different types of RDF values to different types of Elasticsearch values according to the basic type of the RDF value (URI or literal) and the datatype of literals. The autodetection will use the following mapping:

Manual datatype mapping: datatype (string)

The mapping can be overriden through the property "datatype", which can be specified per field. The value of "datatype" may be any of the xsd: types supported by the automatic mapping or a native Elasticsearch type prefixed by native:, e.g. both xsd:long and native:long will map to the long type in Elasticsearch.

Literals in what language: languages (list of string)

RDF data is often multilingual but you may want to map only some of the languages represented in the literal values. This can be done by specifying a list of language ranges that will be matched to the language tags of literals according to RFC 4647, Section 3.3.1. Basic Filtering. In addition an empty range can be used to include literals that have no language tag. The list of language ranges will map all existing literals that have matching language tags.

Elasticsearch index extra settings: indexCreateSettings (string)

This option will be passed directly to Elasticsearch when creating the index. It can be in JSON, YAML or properties format.

Advanced filtering and fine tuning

entityFilter (string)

The entityFilter parameter is used to fine-tune the set of entities and/or individual values for the configured fields, based on the field value. Entities and field values will be synchronised to Elasticsearch if, and only if, they pass the filter. The entity filter is similar to a FILTER() inside a SPARQL query but not exactly the same. Each configured field can be referred to in the entity filter by prefixing it with a "?", much like referring to a variable in SPARQL. Several operators are supported:

In addition to the operators there are some constructions that can be used to write filters based not on the values but on values related to them:

Accessing the previous element in chain

The construction parent(?var) can be used to go to a previous level in a property chain. It can be applied recursively as many times as needed, e.g. parent(parent(parent(?var))) will go back in the chain three times. The effective value of parent(?var) can be used with the in or not in operator like this: parent(?company) in (<urn:a>, <urn:b>).

Accessing an element beyond the chain

The construction ?var -> uri (alternatively ?var o uri or just ?var uri) can be used to access additional values that are accessible through the property uri. In essence this construction corresponds to the triple pattern value uri ?effectiveValue, where ?value is a value bound by the field var. The effective value of ?var -> uri can be used with the in or not in operator like this: ?company -> rdfs:type in (<urn:c>, <urn:d>). It can be combined with parent() like this: parent(?company) -> rdf:type.

The URI parameter can be a full URI within < > or the special string rdfs:type (alternatively just type), which will be expanded to http://www.w3.org/1999/02/22-rdf-syntax-ns#type.

Filtering by RDF context

The construction context(?var) can be used to access the RDF context of a field's value. The typical use case is to sync only explicit values: context(?a) not in (<http://www.ontotext.com/implicit>). The construction can be combined with parent() like this: context(parent(?a)) in (<urn:a>).

Entity filters and default values

Entity filters can be combined with default values in order to get more flexible behaviour.

A typical use-case for an entity filter is having soft deletes, i.e. instead of deleting an entity it is marked as deleted by the presence of a specific value for a given property.

Basic entity filter example

For example, if we create a connector like this:

and then insert some entities:

We could create the following index to specify a default value for city:

The default value will be used for entity:b as it has no value for city in the repository. As the value is "London", the entity will be synchronised.

Advanced entity filter example

Sometimes data represented in RDF is not well suited to map directly to non-RDF. For example, if we have news articles and they can be tagged with different concepts (locations, persons, events, etc.), one possible way to model that is a single property :taggedWith. Consider the following RDF data:

Now, if we want to map this data to Elasticsearch such that the property :taggedWith x is mapped to separate fields taggedWithPerson and taggedWithLocation according to the type of x (we are not interested in events), we can map :taggedWith twice to different fields and then use an entity filter to get the desired values:

Note: type is the short way to write <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>.

The six articles in the RDF data above will be mapped as such:

This can be checked by issuing a faceted search for taggedWithLocation and taggedWithPerson:

If the filter was applied you should get only :Berlin for taggedWithLocation and only :Einstein and :Mozart for taggedWithPerson:

Overview of connector predicates

The following diagram shows a summary of all predicates that can administer (create, drop, check status) connector instances or issue queries and retrieve results. It can be used as a quick reference of what a particular predicate needs to be attached to. For example, to retrieve entities you need to use :entities on a search instance and to retrieve snippets you need to use :snippets on an entity. Variables that are bound as a result of query are shown in green, blank helper nodes are shown in blue, literals in red, and URIs in orange. The predicates are represented by labelled arrows.

Caveats

Order of control

Even though SPARQL per se is not sensitive to the order of triple patterns, the connectors expect to receive certain predicates before others so that queries can be executed properly. In particular, predicates that specify the query or query options need to come before any predicates that fetch results.

The diagram in Overview of connector predicates provides a quick overview of the predicates.

Migrating from a pre-6.2 version

GraphDB prior to 6.2 shipped with a versions of the connectors that had different options and slightly different behaviour. Most existing connector instances will be automatically migrated to the new settings but in some cases it is not possible to continue using the same queries. It is recommended to review the connector configuration after the upgrade and if necessary recreate it with adjusted parameters.

Changes in field configuration and synchronisation

Prior to 6.2 a single field in the config could produce up to three individual fields on the Elasticsearch side based on the field options. For example for the field "firstName":

The current version always produces a single Elasticsearch field per field definition in the configuration. This means you are responsible for creating all appropriate fields based on your needs. See more under Creation parameters.

The option manageExternalIndex

Prior to 6.2 the option manageExternalIndex could be used to control the management of both the schema and the index. In the current implementation there are separate options, manageSchema and manageIndex. See Schema and index management for more information.

Migrating from the Lucene4 plugin

You can easily migrate your existing lucene4 plugin setup to the new connectors interface.

Create index queries

We provide an automated migration tool for your create index queries. The tool is distributed with GraphDB 6.0 onward and can be found in the tools subdirectory. Here is how to use it:

where input-file is your old sparql file and output-file is the new sparql file

you can find possible options with

Select queries using the index

We changed the syntax for the search queries to be able to match our needs for new features and better design. Here is an example query using the lucene4 plugin:

and here is it's connectors variant:

note the following changes:

• We are using special predicates for everything - no more key value options in a string
• The query is actually an instance of the index
• snippets belong to the entity
• snippets are now first class objects - you can also get the field of the match
• indexes are now an instance of another namespace. This allows you to create indexes with the name "entities" for example.

Look at Overview of connector predicates for more info on the new syntax and how everything is linked together.