RDFtk: Core

This crate provides an implementation of the RDF abstract syntax along with a Resource type that provides a builder-like experience for models.

From RDF 1.1 Concepts and Abstract Syntax;

The core structure of the abstract syntax is a set of triples, each consisting of a subject, a predicate and an object. A set of such triples is called an RDF graph. An RDF graph can be visualized as a node and directed-arc diagram, in which each triple is represented as a node-arc-node link.

There can be three kinds of nodes in an RDF graph: IRIs, literals, and blank nodes.

In this library the triple, or statement, as well as subject, predicate, and object types are in the module statement. Literal's as objects are supported in the literal module. Traits that describe graphs are provided by the graph module.

Additional features are provided such as support for data sets (module model.data_set) as well as support for extensions to the core RDF abstract model such as RDF-star.

Example

use rdftk_core::{Literal, Statement, StatementList, SubjectNode};
use rdftk_iri::IRI;
use std::rc::Rc;
use std::str::FromStr;

pub fn make_statements() -> StatementList {
    let mut statements: StatementList = Default::default();
    
    statements.push(Statement::new(
        SubjectNode::named(
            IRI::from_str("http://en.wikipedia.org/wiki/Tony_Benn").unwrap()
        ),
        IRI::from_str("http://purl.org/dc/elements/1.1/title").unwrap(),
        Literal::new("Tony Benn").into(),
    ).into());
    // ...
    statements
}

Changes

Version 0.5.6

Feature: this release adds the variant Collection to ObjectNode as well as a corresponding Collection type. This models the RDF 1.1 collection concept as per RDF 1.1 Semantics D.3 RDF collections. Currently collections are not supported for subject nodes.

• A Collection is a wrapper around a vector of object nodes.
  • Collections are immutable.
  • Collections may contain object nodes that are themselves collections.
  • Collections have a reify method to convert to a set of collection triples.
• A new feature has been added FEATURE_STMT_OBJECT_COLLECTIONS that is now supported by both ObjectNode and Statement.

Version 0.5.5

• Fix: remove spaces from xsd:hexBinary string.

Version 0.5.4

• Feature: better implementation of Display for Literal.

Version 0.5.3

• Feature: add SKOS to PrefixMapping commons.

Version 0.5.2

• Fix: bug in hex_encode for binary literals.

Version 0.5.1

• Feature: added more From implementations for statement nodes.

Version 0.5.0

This is a radical refactor as the complexity of traits was adding more cost than value.

• Removed all traits and moved implementations from the simple module into model.
• Removed all factory interfaces to simplify the user experience.
• Added From and PartialEq implementations for the supported literal value types.

Version 0.4.5

• Refactor: Remove all Ref wrapper types.
• Refactor: Make SubjectNode and ObjectNode into concrete enums.
• Refactor: Remove Equiv trait and rely on PartialEq instead.
• Refactor: Create new Implementation trait and simple::Implementation to hold the set of factories.

Version 0.4.4

• Refactor: re-wrote Error type and removed error_chain dependency.

Version 0.4.3

• Feature: added From<&BlankNode> for both Name and String.

Version 0.4.2

• Feature: changed some methods to use Name instead of strings
• Feature: added new NamedGraph trait.
• Feature: updated DataSet to use NamedGraph instead of separate name/graph pairs.
• Build: updated dependency rdftk_iri.
• Build: updated dependency rdftk_names.

Version 0.4.1

• Build: cargo audit/outdated/udeps

Version 0.4.0

• Feature: moved to new rdftk_iri package which uses the url::Url and is more efficient.
• Feature: moved to new rdftk_names package.
• Feature: replaced own implementation of LanguageTag with that from the package language_tags.
• Refactor: made the trait ObjectNode inherit from SubjectNode.
• Refactor: move SimpleSubjectNode to be a tuple type, and implement From<> for all Subject variants.
• Build: updated Rust edition from 2018 to 2021.
• Docs: added API docs as much as possible.
• Docs: run markuplint on README.md.

Version 0.3.0

• The abstract RDF API, described in module 'model', is now entirely trait based.
• An implementation, in module 'simple', provides concrete in-memory types.
• Added a new type LanguageTag to model these values more completely.
• Added notion of features which denote capabilities supported by data sets, graphs, and statements implementations.

Version 0.2.4

• Created new Featured trait implemented by a number of types that allow client query of various optional capabilities. This also subsumes the has_index/has_all_indices capability as all index queries are now feature queries.

Version 0.2.3

• Copied some errors from rdftk_io.
• Renamed the Io variant to ReadWrite.

Version 0.2.2

• Reworked APIs to take Ref types and to be consistent in use of trait objects and types throughout.
• Added factory type for data sets.
• Made PrefixMappings a concrete type in the core::graph::mapping module.
• Added InvalidMatch and Io variants to ErrorKind.
• Added mutators to Statement.

Version 0.2.1

• Changed API, removed mutable traits for Graph and DataSet, moved methods into their base traits.
• Added factory types for graphs.
• Added Skolemization function for graphs.

Version 0.2.0

• A change to the API, all Statement, and statement components are now passed as Rc references.
  • Added additional _ref constructors to allow cleaner client code.
• A change to the API, Graph and DataSet now use type parameters to describe iterators returned by query methods.
• A change to the API, QName constructors now return errors instead of panic on invalid values.
• Added more constructors for literal values.
• Added support for chrono::Duration in literals as well as the std version as chrono supports the correct output form.
• Added eq_ methods on SubjectNode and ObjectNode for simple testing of inner values.
• Added documentation and examples throughout.

Version 0.1.15

• Fixed Clippy suggestions.
• Removed Context from statements.
• Added value_factory method to Graph.
• Placed all unit tests in tests folder.

Version 0.1.14

• Removed stand-alone named graph.
• Added DataSet as a way to associate names to graphs.
• Renamed CachingGraph to ValueFactory and made stand-alone.

Version 0.1.13

• Bug: fixed Literal constructors to produce an escape-safe literal form for strings.

Version 0.1.12

• Fixed: cargo fmt error.

Version 0.1.11

• Added: public types StatementRef and StatementList rather than having Rc obviously in all APIs.

Version 0.1.10

• DEPRECATED Support for Datasets and Quads by adding a context (type ContextNode) to Statement.

Version 0.1.9

• Support for RDF* in Statement.
• Added additional constructors to Statement.
• Renamed Resource method rdf_type to instance_of for compatibility with RDF schema usage.
• Added is_valid associated function to QName.

Version 0.1.8

• Explicit exports over pub use *.

Version 0.1.7

• Split Graph into Graph and MutableGraph.
• Split NamedGraph into NamedGraph and MutableNamedGraph.
• Added get_default_namespace to the PrefixMappings trait as a helper function.
• Altered PrefixMappings::compress and PrefixMappings::expand to take references.

Version 0.1.6

• Explicit version management.

Version 0.1.5

• Updates for rdftk_memgraph to build.

Version 0.1.4

• Made all local dependencies only major/minor valued.

Version 0.1.3

• Moved all IRI to IRIRef on interfaces.
• Moved Graph and associated types into core and deprecated rdftk_graph.

Version 0.1.2

• Clean-up changes.

Version 0.1.1

• Added From to allow direct construction of a SubjectNode from an IRI.
• Fixed a bug in QName that dropped the ":" for non-prefixed values.

Version 0.1.0

• First release.