RDF.rb: Linked Data for Ruby

This is a pure-Ruby library for working with Resource Description Framework (RDF) data.

Gem Version Build Status Coverage Status Gitter chat

Table of contents

  1. Features
  2. Differences between RDF 1.0 and RDF 1.1
  3. Differences between RDF 1.1 and RDF 1.2
  4. Tutorials
  5. Command Line
  6. Examples
  7. Reader/Writer convenience methods
  8. RDF 1.2
  9. Documentation
  10. Dependencies
  11. Installation
  12. Download
  13. Resources
  14. Mailing List
  15. Authors
  16. Contributors
  17. Contributing
  18. License

Features

  • 100% pure Ruby with minimal dependencies and no bloat.
  • Fully compatible with RDF 1.1 specifications.
  • Provisional support for RDF 1.2 specifications.
  • 100% free and unencumbered public domain software.
  • Provides a clean, well-designed RDF object model and related APIs.
  • Supports parsing and serializing N-Triples and N-Quads out of the box, with more serialization format support available through add-on extensions.
  • Includes in-memory graph and repository implementations, with more storage adapter support available through add-on extensions.
  • Implements basic graph pattern (BGP) query evaluation.
  • Plays nice with others: entirely contained in the RDF module, and does not modify any of Ruby's core classes or standard library.
  • Based entirely on Ruby's autoloading, meaning that you can generally make use of any one part of the library without needing to load up the rest.
  • Compatible with Ruby Ruby >= 3.0, Rubinius and JRuby 9.0+.
    • Note, changes in mapping hashes to keyword arguments for Ruby 3+ may require that arguments be passed more explicitly, especially when the first argument is a Hash and there are optional keyword arguments. In this case, Hash argument may need to be explicitly included within {} and the optional keyword arguments may need to be specified using **{} if there are no keyword arguments.
  • Performs auto-detection of input to select appropriate Reader class if one cannot be determined from file characteristics.

HTTP requests

RDF.rb uses Net::HTTP for retrieving HTTP and HTTPS resources. If the RestClient gem is included, that will be used instead to retrieve remote resources. Clients may also consider using RestClient Components to enable client-side caching of HTTP results using Rack::Cache or other Rack middleware.

See RDF::Util::File for configuring other mechanisms for retrieving resources.

Term caching and configuration

RDF.rb uses a weak-reference cache for storing internalized versions of URIs and Nodes. This is particularly useful for Nodes as two nodes are equivalent only if they're the same node.

By default, each cache can grow to an unlimited size, but this can be configured using RDF.config, for general limits, along with URI- or Node-specific limits.

For example, to limit the size of the URI intern cache only:

RDF.config.uri_cache_size = 10_000

The default for creating new caches without a specific initialization size can be set using:

RDF.config.cache_size = 100_000

Differences between RDF 1.0 and RDF 1.1

This version of RDF.rb is fully compatible with RDF 1.1, but it creates some marginal incompatibilities with RDF 1.0, as implemented in versions prior to the 1.1 release of RDF.rb:

  • Introduces RDF::IRI, as a synonym for RDF::URI either RDF::IRI or RDF::URI can be used interchangeably. Versions of RDF.rb prior to the 1.1 release were already compatible with IRIs. Internationalized Resource Identifiers (see [RFC3987][]) are a super-set of URIs (see [RFC3986][]) which allow for characters other than standard US-ASCII.
  • RDF::URI no longer uses the Addressable gem. As URIs typically don't need to be parsed, this provides a substantial performance improvement when enumerating or querying graphs and repositories.
  • RDF::List no longer emits a rdf:List type. However, it will now recognize any subjects that are RDF::Node instances as being list elements, as long as they have both rdf:first and rdf:rest predicates.
  • RDF::Graph adding a graph_name to a graph may only be done when the underlying storage model supports graph_names (the default RDF::Repository does). The notion of graph_name in RDF.rb is treated equivalently to Named Graphs within an RDF Dataset, and graphs on their own are not named.
  • RDF::Graph, RDF::Statement and RDF::List now include RDF::Value, and not RDF::Resource. Made it clear that using RDF::Graph does not mean that it may be used within an RDF::Statement, for this see RDF::Term.
  • RDF::Statement now is stricter about checking that all elements are valid when validating.
  • RDF::NTriples::Writer and RDF::NQuads::Writer now default to validate output, only allowing valid statements to be emitted. This may disabled by setting the :validate option to false.
  • RDF::Dataset is introduced as a class alias of RDF::Repository. This allows closer alignment to the RDF concept of Dataset.
  • The graph_name of a graph within a Dataset or Repository may be either an RDF::IRI or RDF::Node. Implementations of repositories may restrict this to being only RDF::IRI.
  • There are substantial and somewhat incompatible changes to RDF::Literal. In RDF 1.1, all literals are typed, including plain literals and language tagged literals. Internally, plain literals are given the xsd:string datatype and language tagged literals are given the rdf:langString datatype. Creating a plain literal, without a datatype or language, will automatically provide the xsd:string datatype; similar for language tagged literals. Note that most serialization formats will remove this datatype. Code which depends on a literal having the xsd:string datatype being different from a plain literal (formally, without a datatype) may break. However note that the #has\_datatype? will continue to return false for plain or language-tagged literals.
  • RDF::Query#execute now accepts a block and returns RDF::Query::Solutions. This allows enumerable.query(query) to behave like query.execute(enumerable) and either return an enumerable or yield each solution.
  • RDF::Queryable#query now returns RDF::Query::Solutions instead of an Enumerator if it's argument is an RDF::Query.
  • RDF::Util::File.open_file now performs redirects and manages base_uri based on W3C recommendations:
    • base_uri is set to the original URI if a status 303 is provided, otherwise any other redirect will set base_uri to the redirected location.
    • base_uri is set to the content of the Location header if status is success.
  • Additionally, RDF::Util::File.open_file sets the result encoding from charset if provided, defaulting to UTF-8. Other access methods include last_modified and content_type,
  • RDF::StrictVocabulary added with an easy way to keep vocabulary definitions up to date based on their OWL or RDFS definitions. Most vocabularies are now StrictVocabularies meaning that an attempt to resolve a particular term in that vocabulary will error if the term is not defined in the vocabulary.
  • New vocabulary definitions have been added for ICal, Media Annotations (MA), Facebook OpenGraph (OG), PROV, SKOS-XL (SKOSXL), Data Vocabulary (V), VCard, VOID, Powder-S (WDRS), and XHV.

Notably, RDF::Queryable#query and RDF::Query#execute are now completely symmetric; this allows an implementation of RDF::Queryable to optimize queries using implementation-specific logic, allowing for substantial performance improvements when executing BGP queries.

Differences between RDF 1.1 and RDF 1.2

  • RDF::Literal has an optional direction property for directional language-tagged strings.
  • Removes support for legacy text/plain (as an alias for application/n-triples) and text/x-nquads (as an alias for application/n-quads)

Tutorials

Command Line

When installed, RDF.rb includes a rdf shell script which acts as a wrapper to perform a number of different operations on RDF files using available readers and writers.

  • count: Parse and RDF input and count the number of statements.
  • predicates: Returns unique objects from parsed input.
  • objects: Returns unique objects from parsed input.
  • serialize: Parse an RDF input and re-serializing to N-Triples or another available format using --output-format option.
  • subjects: Returns unique subjects from parsed input.

The serialize command can also be used to serialize as a vocabulary.

Different RDF gems will augment the rdf script with more capabilities, which may require specifying the appropriate --input-format option to revel.

Examples

require 'rdf'
include RDF

Writing RDF data using the N-Triples format

require 'rdf/ntriples'
graph = RDF::Graph.new << [:hello, RDF::RDFS.label, "Hello, world!"]
graph.dump(:ntriples)

or

RDF::Writer.open("hello.nt") { |writer| writer << graph }

Reading RDF data in the N-Triples format

require 'rdf/ntriples'
graph = RDF::Graph.load("https://ruby-rdf.github.io/rdf/etc/doap.nt")

or

RDF::Reader.open("https://ruby-rdf.github.io/rdf/etc/doap.nt") do |reader|
  reader.each_statement do |statement|
    puts statement.inspect
  end
end

Reading RDF data in other formats

RDF::Reader.open and RDF::Repository.load use a number of mechanisms to determine the appropriate reader to use when loading a file. The specific format to use can be forced using, e.g. format: :ntriples option where the specific format symbol is determined by the available readers. Both also use MimeType or file extension, where available.

require 'rdf/nquads'

graph = RDF::Graph.load("https://ruby-rdf.github.io/rdf/etc/doap.nq", format: :nquads)

A specific sub-type of Reader can also be invoked directly:

require 'rdf/nquads'

RDF::NQuads::Reader.open("https://ruby-rdf.github.io/rdf/etc/doap.nq") do |reader|
  reader.each_statement do |statement|
    puts statement.inspect
  end
end

Reader/Writer implementations may override RDF::Format.detect, which takes a small sample if input and return a boolean indicating if it matches that specific format. In the case that a format cannot be detected from filename or other options, or that more than one format is identified, RDF::Format.for will query each loaded format by invoking it's detect method, and the first successful match will be used to read the input.

Writing RDF data using other formats

RDF::Writer.open, RDF::Enumerable#dump, RDF::Writer.dump take similar options to RDF::Reader.open to determine the appropriate writer to use.

require 'linkeddata'

RDF::Writer.open("hello.nq", format: :nquads) do |writer|
  writer << RDF::Repository.new do |repo|
    repo << RDF::Statement.new(:hello, RDF::RDFS.label, "Hello, world!", graph_name: RDF::URI("http://example/graph_name"))
  end
end

A specific sub-type of Writer can also be invoked directly:

require 'rdf/nquads'

repo = RDF::Repository.new << RDF::Statement.new(:hello, RDF::RDFS.label, "Hello, world!", graph_name: RDF::URI("http://example/graph_name"))
File.open("hello.nq", "w") {|f| f << repo.dump(:nquads)}

Reader/Writer convenience methods

RDF::Enumerable implements to_{format} for each available instance of RDF::Reader. For example, if rdf/turtle is loaded, this allows the following:

graph = RDF::Graph.new << [:hello, RDF::RDFS.label, "Hello, world!"]
graph.to_ttl

Similarly, RDF::Mutable implements from_{format} for each available instance of RDF::Writer. For example:

graph = RDF::Graph.new
graph.from_ttl("[ a <http://www.w3.org/1999/02/22-rdf-syntax-ns#Resource>]")

Note that no prefixes are loaded automatically, however they can be provided as arguments:

graph.from_ttl("[ a rdf:Resource]", prefixes: {rdf: RDF.to_uri})

Querying RDF data using basic graph patterns (BGPs)

require 'rdf/ntriples'

graph = RDF::Graph.load("https://ruby-rdf.github.io/rdf/etc/doap.nt")
query = RDF::Query.new({
  person: {
    RDF.type  => FOAF.Person,
    FOAF.name => :name,
    FOAF.mbox => :email,
  }
}, **{})

query.execute(graph) do |solution|
  puts "name=#{solution.name} email=#{solution.email}"
end

The same query may also be run from the graph:

graph.query(query) do |solution|
  puts "name=#{solution.name} email=#{solution.email}"
end

In general, querying from using the queryable instance allows a specific implementation of queryable to perform query optimizations specific to the datastore on which it is based.

A separate SPARQL gem builds on basic BGP support to provide full support for SPARQL 1.1 queries.

Using pre-defined RDF vocabularies

DC.title      #=> RDF::URI("http://purl.org/dc/terms/title")
FOAF.knows    #=> RDF::URI("http://xmlns.com/foaf/0.1/knows")
RDF.type      #=> RDF::URI("http://www.w3.org/1999/02/22-rdf-syntax-ns#type")
RDFS.seeAlso  #=> RDF::URI("http://www.w3.org/2000/01/rdf-schema#seeAlso")
RSS.title     #=> RDF::URI("http://purl.org/rss/1.0/title")
OWL.sameAs    #=> RDF::URI("http://www.w3.org/2002/07/owl#sameAs")
XSD.dateTime  #=> RDF::URI("http://www.w3.org/2001/XMLSchema#dateTime")

Using ad-hoc RDF vocabularies

foaf = RDF::Vocabulary.new("http://xmlns.com/foaf/0.1/")
foaf.knows    #=> RDF::URI("http://xmlns.com/foaf/0.1/knows")
foaf[:name]   #=> RDF::URI("http://xmlns.com/foaf/0.1/name")
foaf['mbox']  #=> RDF::URI("http://xmlns.com/foaf/0.1/mbox")

RDF 1.2

RDF.rb includes provisional support for RDF 1.2 with an N-Triples/N-Quads syntax for quoted triples in the subject or object position. RDF.rb includes provisional support for RDF 1.2 directional language-tagged strings, which are literals of type rdf:dirLangString having both a language and direction.

Internally, an RDF::Statement is treated as another resource, along with RDF::URI and RDF::Node, which allows an RDF::Statement to have a #subject or #object which is also an RDF::Statement.

Note: This feature is subject to change or elimination as the standards process progresses.

Serializing a Graph containing quoted triples

require 'rdf/ntriples'
statement = RDF::Statement(RDF::URI('bob'), RDF::Vocab::FOAF.age, RDF::Literal(23))
graph = RDF::Graph.new << [statement, RDF::URI("ex:certainty"), RDF::Literal(0.9)]
graph.dump(:ntriples, validate: false)
# => '<<<bob> <http://xmlns.com/foaf/0.1/age> "23"^^<http://www.w3.org/2001/XMLSchema#integer>>> <ex:certainty> "0.9"^^<http://www.w3.org/2001/XMLSchema#double> .'

Reading a Graph containing quoted triples

By default, the N-Triples reader will reject a document containing a subject resource.

nt = '<<<bob> <http://xmlns.com/foaf/0.1/age> "23"^^<http://www.w3.org/2001/XMLSchema#integer>>> <ex:certainty> "0.9"^^<http://www.w3.org/2001/XMLSchema#double> .'
graph = RDF::Graph.new do |graph|
  RDF::NTriples::Reader.new(nt) {|reader| graph << reader}
end
# => RDF::ReaderError

Documentation

https://ruby-rdf.github.io/rdf

RDF Object Model

RDF Serialization

RDF Serialization Formats

The following is a partial list of RDF formats implemented either natively, or through the inclusion of other gems:

The meta-gem LinkedData includes many of these gems.

RDF Datatypes

RDF.rb only implements core datatypes from the RDF Datatype Map. Most other XSD and RDF datatype implementations can be find in the following:

Graph Isomorphism

Two graphs may be compared with each other to determine if they are isomorphic. As BNodes within two different graphs are no equal, graphs may not be directly compared. The RDF::Isomorphic gem may be used to determine if they make the same statements, aside from BNode identity (i.e., they each entail the other)

  • RDF::Isomorphic

RDF Storage

RDF Querying

RDF Vocabularies

Dependencies

Installation

The recommended installation method is via RubyGems. To install the latest official release of RDF.rb, do:

% [sudo] gem install rdf             # Ruby 3+

Download

To get a local working copy of the development repository, do:

% git clone git://github.com/ruby-rdf/rdf.git

Alternatively, download the latest development version as a tarball as follows:

% wget https://github.com/ruby-rdf/rdf/tarball/master

Resources

Mailing List

Authors

Contributors

Contributing

This repository uses Git Flow to mange development and release activity. All submissions must be on a feature branch based on the develop branch to ease staging and integration.

  • Do your best to adhere to the existing coding conventions and idioms.
  • Don't use hard tabs, and don't leave trailing whitespace on any line. Before committing, run git diff --check to make sure of this.
  • Do document every method you add using YARD annotations. Read the tutorial or just look at the existing code for examples.
  • Don't touch the .gemspec or VERSION files. If you need to change them, do so on your private branch only.
  • Do feel free to add yourself to the CREDITS file and the corresponding list in the the README. Alphabetical order applies.
  • Don't touch the AUTHORS file. If your contributions are significant enough, be assured we will eventually add you in there.
  • Do note that in order for us to merge any non-trivial changes (as a rule of thumb, additions larger than about 15 lines of code), we need an explicit public domain dedication on record from you, which you will be asked to agree to on the first commit to a repo within the organization. Note that the agreement applies to all repos in the Ruby RDF organization.

License

This is free and unencumbered public domain software. For more information, see https://unlicense.org/ or the accompanying UNLICENSE file.