@ParametersAreNonnullByDefault

Package eu.fbk.knowledgestore

KnowledgeStore core API (ks-core).

See: Description

Interface Summary

Interface	Description
KnowledgeStore	A KnowledgeStore instance.
Session	A KnowledgeStore user session.

Class Summary

Class	Description
AbstractKnowledgeStore
AbstractSession
ForwardingKnowledgeStore
ForwardingSession
Operation
Operation.Count
Operation.Create
Operation.Delete
Operation.Download	Download operation.
Operation.Match
Operation.Merge
Operation.Retrieve
Operation.Sparql	SPARQL query operation.
Operation.SparqlDelete
Operation.SparqlUpdate
Operation.Update
Operation.Upload	Upload operation.
Outcome	The outcome of the invocation of a KnowledgeStore operation.

Enum Summary

Enum	Description
Outcome.Status	Enumeration of Outcome status codes.

Exception Summary

Exception	Description
OperationException	Signals the failure of a KnowledgeStore Operation invocation.

Package eu.fbk.knowledgestore Description

KnowledgeStore core API (ks-core).

Data model

A total ordering relation is defined on nodes, allowing to compare any node against any other node. This relation generalizes the single ordering relation that can be naturally defined for specific types of nodes, and is provided for ease of use (e.g., to provide a canonical way to order nodes) and for simplifying the definition of conditions on nodes. More in details, scalar nodes are different and are sorted before record nodes, while specific rules are applied for ordering among scalars and among records (see, respectively, the documentation for Record and Scalar).

Resources. A resource is a self-contained piece of unstructured content, such as a news article or a multimedia object, having some descriptive metadata (e.g., from DCTERMS vocabulary), a binary representation ( KS.STORED_AS) and zero or more mentions contained within it (KS.HAS_MENTION ). Note that (1) representation metadata is generated automatically by the system based on the uploaded content, so there is no need to manually set it; and (2) the actual binary data is not available through the resource record object, but must be instead accessed using other API methods of the KnowledgeStore.

Representations. A digital representation is possibly stored for each resource in the KnowledgeStore. A Representation record contains the metadata of such a representation, while its actual binary content can be accessed via the Representation class using specific methods of the KnowledgeStore API. The representation metadata stored by this kind of record comprise the filename (NFO.FILE_NAME), the size in bytes ( NFO.FILE_SIZE ), the date and time the representation was created (NFO.FILE_CREATED) and the MIME type (see NIE.MIME_TYPE). Representation records are automatically created by the system for stored resources.

Mentions. A mention is a snippet of a resource ( KS.MENTION_OF), such as some characters in a text document or some pixels in an image, that may refer to some entity of interest ( KS.REFERS_TO). Mentions can be automatically extracted by natural language and multimedia processing tools, that can enrich them with additional attributes about how they denote their referent (e.g., name, qualifiers, 'sentiment'). Therefore, mentions present both unstructured and structured facets not available in resources and entities layers alone, and are thus a valuable source of information on their own. A number of axioms ( KS.EXPRESSES) may be expressed by the mention, i.e., can be extracted from its snippet and attributes.

Entities. An entity is described by a (possibly empty) set of axioms ( KS.DESCRIBED_BY) and is referred by a (possibly empty) set of mentions ( KS.REFERRED_BY).

Axioms. Axioms describe entities of the KnowledgeStore. An axiom is encoded by a set of RDF statements (KS.ENCODED_BY) that relate one or more entities and its validity may be restricted to a specific context ( KS.HOLDS_IN); an axiom may be annotated with additional metadata and can be expressed by a number of mentions ( KS.EXPRESSED_BY). Note that a null context is used in case of axioms holding universally and not in a specific context. An axiom may be simple or complex, with the first consisting of exactly one RDF statement (possibly contextualized and associated to additional metadata); simple axioms are generally used for ABox assertions, while complex axioms are associated mainly to TBox declarations. Axioms IDs should be assigned based on RDF statements and context.

Contexts A context defines the circumstances under which a certain axiom holds; each axiom is associated exactly to a context in the KnowledgeStore (which may be shared by multiple axioms) via the KS.HOLDS_IN property (if null, the axiom holds universally). A context is identified by a set of attributes or contextual dimensions, which are the properties of the context record. IDs of context records should be assigned based on all the values of all the contextual dimensions.