SWI-Prolog Semantic Web Library
Jan Wielemaker
University of Amsterdam/VU University Amsterdam
The Netherlands
E-mail: J.Wielemaker@cs.vu.nl
This document describes a library for dealing with standards from the W3C standard for the Semantic Web. Like the standards themselves (RDF, RDFS and OWL) this infrastructure is modular. It consists of Prolog packages for reading, querying and storing semantic web documents as well as XPCE libraries that provide visualisation and editing. The Prolog libraries can be used without the XPCE GUI modules. The library has been actively used with upto 10 million triples, using approximately 1GB of memory. Its scalability is limited by memory only. The library can be used both on 32-bit and 64-bit platforms.

Table of Contents

1 Introduction
2 Provided libraries
3 library(semweb/rdf_db): The RDF database
3.1 Query the RDF database
3.1.1 Literal matching and indexing
3.2 Predicate properties
3.3 Modifying the database
3.3.1 Modifying predicates
3.3.2 Transactions
3.4 Loading and saving to file
3.4.1 library(semweb/rdf_cache): Caching triples
3.4.2 Partial save
3.4.3 Fast loading and saving
3.4.4 MD5 digests
3.5 Namespace Handling
3.5.1 Namespace handling for custom predicates
3.6 Monitoring the database
3.7 Miscellaneous predicates
3.8 Issues with rdf_db
4 Plugin modules for rdf_db
4.1 Hooks into the RDF library
4.2 library(semweb/rdf_zlib_plugin): Reading compressed RDF
4.3 library(semweb/rdf_http_plugin): Reading RDF from a HTTP server
4.4 library(semweb/rdf_litindex): Indexing words in literals
4.4.1 Literal maps: Creating additional indices on literals
4.5 library(semweb/rdf_persistency): Providing persistent storage
4.5.1 Enriching the journals
5 library(semweb/rdf_turtle): Turtle: Terse RDF Triple Language
6 library(semweb/rdf_turtle_write): Turtle - Terse RDF Triple Language writer
7 library(semweb/rdfs): RDFS related queries
7.1 Hierarchy and class-individual relations
7.2 Collections and Containers
7.3 Labels and textual search
8 Managing RDF input files
8.1 The Manifest file
8.1.1 Support for the VoID and VANN vocabularies
8.1.2 Finding manifest files
8.2 Usage scenarios
8.2.1 Referencing resources
8.3 Putting it all together
8.4 Example: A metadata file for W3C WordNet
9 library(semweb/sparql_client): SPARQL client library
10 library(semweb/rdf_compare): Compare RDF graphs
11 library(semweb/rdf_portray): Portray RDF resources
12 library(semweb/rdf_edit): Keep track of edits (deprecated)
12.1 Transaction management
12.2 File management
12.3 Encapsulated predicates
12.4 High-level modification predicates
12.5 Undo
12.6 Journalling
12.7 Broadcasting change events
13 Related packages and issues
14 OWL

1 Introduction

SWI-Prolog has started support for web-documents with the development of a small and fast SGML/XML parser, followed by an RDF parser (early 2000). With the semweb library we provide more high level support for manipulating semantic web documents. The semantic web is the likely point of orientation for knowledge representation in the future, making a library designed in its spirit promising.

2 Provided libraries

Central to this library is the module library(semweb/rdf_db.pl), providing storage and basic querying for RDF triples. This triple store is filled using the RDF parser realised by library(rdf.pl). The storage module can quickly save and load (partial) databases. The modules library(semweb/rdfs.pl) and library(semweb/owl.pl) add querying in terms of the more powerful RDFS and OWL languages. Module library(semweb/rdf_edit.pl) adds editing, undo, journaling and change-forwarding. Finally, a variety of XPCE modules visualise and edit the database. Figure figure 1 summarised the modular design.

Figure 1 : Modules for the Semantic Web library

3 library(semweb/rdf_db): The RDF database

The central module is called rdf_db. It provides storage and indexed querying of RDF triples. Triples are stored as a quintuple. The first three elements denote the RDF triple. File and Line provide information about the origin of the triple.

{Subject Predicate Object File Line}

The actual storage is provided by the foreign language (C) module rdf_db.c. Using a dedicated C-based implementation we can reduced memory usage and improve indexing capabilities.1The orginal implementation was in Prolog. This version was implemented in 3 hours, where the C-based implementation costed a full week. The C-based implementation requires about half the memory and provides about twice the performance. Currently the following indexing is provided.

3.1 Query the RDF database

rdf(?Subject, ?Predicate, ?Object)
Elementary query for triples. Subject and Predicate are atoms representing the fully qualified URL of the resource. Object is either an atom representing a resource or literal(Value) if the object is a literal value. If a value of the form NameSpaceID : LocalName is provided it is expanded to a ground atom using expand_goal/2. This implies you can use this construct in compiled code without paying a performance penalty. See also section 3.5. Literal values take one of the following forms:
If the value is a simple atom it is the textual representation of a string literal without explicit type or language (xml:lang) qualifier.
lang(LangID, Atom)
Atom represents the text of a string literal qualified with the given language.
type(TypeID, Value)
Used for attributes qualified using the rdf:datatype TypeID. The Value is either the textual representation or a natural Prolog representation. See the option convert_typed_literal(:Convertor) of the parser. The storage layer provides efficient handling of atoms, integers (64-bit) and floats (native C-doubles). All other data is represented as a Prolog record.

For string querying purposes, Object can be of the form literal(+Query, -Value), where Query is one of the terms below. Details of literal matching and indexing are described in section 3.1.1.

Perform exact match and demand the language or type qualifiers to match. This query is fully indexed.2This should have been the default when using literal with one argument because it is logically consisent (i.e., (rdf(S,P,literal(X)), X == hello) would have been the same as rdf(S,P,literal(hello). In addition, this is consistent with SPARQL literal identity definition.
Perform exact, but case-insensitive match. This query is fully indexed.
Match any literal that contains Text as a case-insensitive substring. The query is not indexed on Object.
Match any literal that contains Text delimited by a non alpha-numeric character, the start or end of the string. The query is not indexed on Object.
Match any literal that starts with Text. This call is intended for completion. The query is indexed using the binary tree of literals. See section 3.1.1 for details.
Match any literal that is equal or larger then Literal in the ordered set of literals.
Match any literal that is equal or smaller then Literal in the ordered set of literals.
between(+Literal1, +Literal2)
Match any literal that is between Literal1 and Literal2 in the ordered set of literals. This may include both Literal1 and Literal2.
Match any literal that matches Pattern case insensitively, where the `*' character in Pattern matches zero or more characters.

Backtracking never returns duplicate triples. Duplicates can be retrieved using rdf/4. The predicate rdf/3 raises a type-error if called with improper arguments. If rdf/3 is called with a term literal(_) as Subject or Predicate object it fails silently. This allows for graph matching goals like rdf(S,P,O),rdf(O,P2,O2) to proceed without errors.3Discussion in the SPARQL community votes for allowing literal values as subject. Although we have no principal objections, we fear such an extension will promote poor modelling practice.

rdf(?Subject, ?Predicate, ?Object, ?Source)
As rdf/3 but in addition return the source-location of the triple. The source is either a plain atom or a term of the format Atom : Integer where Atom is intended to be used as filename or URL and Integer for representing the line-number. Unlike rdf/3, this predicate does not remove duplicates from the result set.
rdf_has(?Subject, ?Predicate, ?Object, -TriplePred)
This query exploits the RDFS subPropertyOf relation. It returns any triple whose stored predicate equals Predicate or can reach this by following the recursive subPropertyOf relation. The actual stored predicate is returned in TriplePred. The example below gets all subclasses of an RDFS (or OWL) class, even if the relation used is not rdfs:subClassOf, but a user-defined sub-property thereof.4This predicate realises semantics defined in RDF-Schema rather than RDF. It is part of the library(rdf_db) module because the indexing of this module incorporates the rdfs:subClassOf predicate.
subclasses(Class, SubClasses) :-
        findall(S, rdf_has(S, rdfs:subClassOf, Class), SubClasses).

In addition, rdf_has/4 handles the predicate property (see rdf_predicate_property/2) symetric(true) and inverse_of(P2).

Note that rdf_has/4 and rdf_has/3 can return duplicate answers if they use a different TriplePred.

rdf_has(?Subject, ?Predicate, ?Object)
Same as rdf_has(Subject, Predicate, Object, _).
rdf_reachable(?Subject, +Predicate, ?Object)
Is true if Object can be reached from Subject following the transitive predicate Predicate or a sub-property thereof, while repecting the symetric(true) or inverse_of(P2) properties.

If used with either Subject or Object unbound, it first returns the origin, followed by the reachable nodes in breath-first search-order. The implementation internally looks one solution ahead and succeeds deterministically on the last solution. This predicate never generates the same node twice and is robust against cycles in the transitive relation.

With all arguments instantiated, it succeeds deterministically if a path can be found from Subject to Object. Searching starts at Subject, assuming the branching factor is normally lower. A call with both Subject and Object unbound raises an instantiation error.5Should be using bi-directional search. The following example generates all subclasses of rdfs:Resource:

?- rdf_reachable(X, rdfs:subClassOf, rdfs:'Resource').
X = 'http://www.w3.org/2000/01/rdf-schema#Resource' ;
X = 'http://www.w3.org/2000/01/rdf-schema#Class' ;
X = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#Property' ;
rdf_reachable(?Subject, +Predicate, ?Object, +MaxD, -D)
Same as rdf_reachable/3, but in addition, MaxD limits the number of relations expanded and D is unified with the `distance' between Subject and Object. Distance 0 means Subject and Object are the same resource. MaxD can be the constant infinite to impose no distance-limit.
Enumerate resources appearing as a subject in a triple. The main reason for this predicate is to generate the known subjects without duplicates as one gets using rdf(Subject, _, _).
Enumerate all known literals. Like rdf_subject/1, the motivation is to provide access to literals without generation duplicates. Otherwise the call is the same as rdf(_,_,literal(Literal)).

3.1.1 Literal matching and indexing

Starting with version 2.5.0 of this library, literal values are ordered and indexed using a balanced binary tree (AVL tree). The aim of this index is threefold.

As string literal matching is most frequently used for searching purposes, the match is executed case-insensitive and after removal of diacritics. Case matching and diacritics removal is based on Unicode character properties and independent from the current locale. Case conversion is based on the `simple uppercase mapping' defined by Unicode and diacritic removal on the `decomposition type'. The approach is lightweight, but somewhat simpleminded for some languages. The tables are generated for Unicode characters upto 0x7fff. For more information, please check the source-code of the mapping-table generator unicode_map.pl available in the sources of this package.

Currently the total order of literals is first based on the type of literal using the ordering

numeric < string < term
Numeric values (integer and float) are ordered by value, integers preceed floats if they represent the same value. strings are sorted alphabetically after case-mapping and diacritic removal as described above. If they match equal, uppercase preceeds lowercase and diacritics are ordered on their unicode value. If they still compare equal literals without any qualifier preceeds literals with a type qualifier which preceeds literals with a language qualifier. Same qualifiers (both type or both language) are sorted alphabetically.7The ordering defined above may change in future versions to deal with new queries for literals.

The ordered tree is used for indexed execution of literal(prefix(Prefix), Literal) as well as literal(like(Like), Literal) if Like does not start with a `*'. Note that results of queries that use the tree index are returned in alphabetical order.

3.2 Predicate properties

The predicates below form an experimental interface to provide more reasoning inside the kernel of the rdb_db engine. Note that symetric, inverse_of and transitive are not yet supported by the rest of the engine.

Enumerate all predicates that are used in at least one triple. Behaves as the code below, but much more efficient.
rdf_current_predicate(Predicate) :-
        findall(P, rdf(_,P,_), Ps),
        sort(Ps, S),
        member(Predicate, S).

Note that there is no relation to defined RDF properties. Properties that have no triples are not reported by this predicate, while predicates that are involved in triples do not need to be defined as an instance of rdf:Property.

rdf_set_predicate(+Predicate, +Property)
Define a property of the predicate. This predicate currently supports the properties symmetric, inverse_of and transitive as defined with rdf_predicate_property/2. Adding an A inverse_of B also adds B inverse_of A. An inverse relation is deleted using inverse_of([]). `
rdf_predicate_property(?Predicate, -Property)
Query properties of a defined predicate. Currently defined properties are given below.
True if the predicate is defined to be symetric. I.e., {A} P {B} implies {B} P {A}. Setting symmetric is equivalent to inverse_of(Self).
True if this predicate is the inverse of Inverse. This property is used by rdf_has/3, rdf_has/4, rdf_reachable/3 and rdf_reachable/5.
True if this predicate is transitive. This predicate is currently not used. It might be used to make rdf_has/3 imply rdf_reachable/3 for transitive predicates.
Unify Triples with the number of existing triples using this predicate as second argument. Reporting the number of triples is intended to support query optimization.
Unify Float with the average number of triples associated with each unique value for the subject-side of this relation. If there are no triples the value 0.0 is returned. This value is cached with the predicate and recomputed only after substantial changes to the triple set associated to this relation. This property is indented for path optimalisation when solving conjunctions of rdf/3 goals.
Unify Float with the average number of triples associated with each unique value for the object-side of this relation. In addition to the comments with the subject_branch_factor property, uniqueness of the object value is computed from the hash key rather than the actual values.
Same as rdf_subject_branch_factor/1, but also considering triples of `subPropertyOf' this relation. See also rdf_has/3.
Same as rdf_object_branch_factor/1, but also considering triples of `subPropertyOf' this relation. See also rdf_has/3.

3.3 Modifying the database

As depicted in figure 1, there are two levels of modification. The rdf_db module simply modifies, where the rdf_edit library provides transactions and undo on top of this. Applications that wish to use the rdf_edit layer must never use the predicates from this section directly.

3.3.1 Modifying predicates

rdf_assert(+Subject, +Predicate, +Object)
Assert a new triple into the database. This is equivalent to rdf_assert/4 using SourceRef user. Subject and Predicate are resources. Object is either a resource or a term literal(Value). See rdf/3 for an explanation of Value for typed and language qualified literals. All arguments are subject to name-space expansion (see section 3.5).
rdf_assert(+Subject, +Predicate, +Object, +SourceRef)
As rdf_assert/3, adding SourceRef to specify the orgin of the triple. SourceRef is either an atom or a term of the format Atom:Int where Atom normally refers to a filename and Int to the line-number where the description starts.
rdf_retractall(?Subject, ?Predicate, ?Object)
Remove all matching triples from the database. As rdf_retractall/4 using an unbound SourceRef.
rdf_retractall(?Subject, ?Predicate, ?Object, ?SourceRef)
As rdf_retractall/3, also matching SourceRef. This is particulary useful to remove all triples coming from a loaded file. See also rdf_unload/1.
rdf_update(+Subject, +Predicate, +Object, +Action)
Replaces one of the three fields on the matching triples depending on Action:
Changes the first field of the triple.
Changes the second field of the triple.
Changes the last field of the triple to the given resource or literal(Value).
Moves the triple from its current named graph to Graph. Note that updating the source has no consequences for the semantics and therefore the generation (see rdf_generation/1) is not updated.
rdf_update(+Subject, +Predicate, +Object, +Graph, +Action)
As rdf_update/4 but allows for specifying the graph.

3.3.2 Transactions

The predicates from section 3.3.1 perform immediate and atomic modifications to the database. There are two cases where this is not desirable:

  1. If the database is modified using information based on reading the same database. A typical case is a forward reasoner examining the database and asserting new triples that can be deduced from the already existing ones. For example, if length(X) > 2 then size(X) is large:
            (   rdf(X, length, literal(L)),
                atom_number(L, IL),
                IL > 2,
                rdf_assert(X, size, large),
            ;   true

    Running this code without precautions causes an error because rdf_assert/3 tries to get a write lock on the database which has an a read operation (rdf/3 has choicepoints) in progress.

  2. Multi-threaded access making multiple changes to the database that must be handled as a unit.

Where the second case is probably obvious, the first case is less so. The storage layer may require reindexing after adding or deleting triples. Such reindexing operatations however are not possible while there are active read operations in other threads or from choicepoints that can be in the same thread. For this reason we added rdf_transaction/2. Note that, like the predicates from section 3.3.1, rdf_transaction/2 raises a permission error exception if the calling thread has active choicepoints on the database. The problem is illustrated below. The rdf/3 call leaves a choicepoint and as the read lock originates from the calling thread itself the system will deadlock if it would not generate an exception.

1 ?- rdf_assert(a,b,c).

2 ?- rdf_assert(a,b,d).

3 ?- rdf(a,b,X), rdf_transaction(rdf_assert(a,b,e)).
ERROR: No permission to write rdf_db `default' (Operation would deadlock)
^  Exception: (8) rdf_db:rdf_transaction(rdf_assert(a, b, e)) ? no debug
4 ?-
Same as rdf_transaction(Goal, user).
rdf_transaction(:Goal, +Id)
After starting a transaction, all predicates from section 3.3.1 append their operation to the transaction instead of modifying the database. If Goal succeeds rdf_transaction cuts all choicepoints in Goal and executes all recorded operations. If Goal fails or throws an exception, all recorded operations are discarded and rdf_transaction/1 fails or re-throws the exception.

On entry, rdf_transaction/1 gains exclusive access to the database, but does allow readers to come in from all threads. After the successful completion of Goal rdf_transaction/1 gains completely exclusive access while performing the database updates.

Transactions may be nested. Committing a nested transactions merges its change records into the outer transaction, while discarding a nested transaction simply destroys the change records belonging to the nested transaction.

The Id argument may be used to identify the transaction. It is passed to the begin/end events posted to hooks registered with rdf_monitor/2. The Id log(Term) can be used to enrich the journal files with additional history context. See section 4.5.1.

True if Id is the identifier of a currently active transaction (i.e. rdf_active_transaction/1 is called from rdf_transaction/2 with matching Id). Note that transaction identifier is not copied and therefore need not be ground and can be further instantiated during the transaction. Id is first unified with the innermost transaction and backtracking with the identifier of other active transaction. Fails if there is no matching transaction active, which includes the case where there is no transaction in progress.

3.4 Loading and saving to file

The rdf_db module can read and write RDF-XML for import and export as well as a binary format built for quick load and save described in section 3.4.3. Here are the predicates for portable RDF load and save.

Load triples from In, which is either a stream opened for reading, an atom specifying a filename, a URL or a list of valid inputs. This predicate calls process_rdf/3 to read the source one description at a time, avoiding limits to the size of the input. By default, this predicate provides for caching the results for quick-load using rdf_load_db/1 described below. Caching strategy and options are description in section 3.4.1.
rdf_load(+FileOrList, +Options)
As rdf_load/1, providing additional options. The options are handed to the RDF parser and implemented by process_rdf/3. In addition, the following options are provided:
If true (default), try to use cached data or create a cache file. Otherwise load the source.
Deprecated. New code should use the graph(+Graph) option.
Specify the source format explicitly. Normally this is deduced from the filename extension or the mime-type. The core library understands the formats xml (RDF/XML) and triples (internal quick load and cache format).
Load the data in the given named graph. The default is the URL of the source.
Condition under which to load the source. Condition is the same as for the Prolog load_files/2 predicate: changed (default) load the source if it was not loaded before or has changed; true (re-)loads the source unconditionally and not_loaded loads the source if it was not loaded, but does not check for modifications.
If Bool is true, the message reporting completion is printed using level silent. Otherwise the level is informational. See also print_message/2.
If true (default false), register xmlns:ns=url namespace declarations as rdf_db:ns(ns,url) namespaces if there is no conflict.
Remove all triples loaded from Spec. Spec is either a graph name or a source specificatipn. If Spec does not refer to a loaded database the predicate succeeds silently.
Save all known triples to the given File. Same as rdf_save(File,[]).
rdf_save(+File, +Options)
Save with options. Provided options are:
Save all triples that belong to the named-graph URI. Saving arbitrary selections is possible using predicates from section 3.4.2.
Deprecated synonym for graph(URI).
if anon(false) is provided anonymous resources are only saved if the resource appears in the object field of another triple that is saved.
If provided, emit xml:base="BaseURI" in the header and emit all URIs that are relative to the base-uri. The xml:base declaration can be suppressed using the option write_xml_base(false)
If false (default true), do not emit the xml:base declaration from the given base_uri option. The idea behind this option is to be able to create documents with URIs relative to the document itself:
                 [ base_uri(BaseURI),
If present, raw literal values are first passed to Converter to apply the reverse of the convert_typed_literal option of the RDF parser. The Converter is called with the same arguments as in the RDF parser, but now with the last argument instantiated and the first two unbound. A proper convertor that can be used for both loading and saving must be a logical predicate.
Define the XML encoding used for the file. Defined values are utf8 (default), iso_latin_1 and ascii. Using iso_latin_1 or ascii, characters not covered by the encoding are emitted as XML character entities (&#...;).
The value XMLLang is used for the xml:lang attribute in the outermost rdf:RDF element. This language acts as a default, which implies that the xml:lang tag is only used for literals with a different language identifier. Please note that this option will cause all literals without language tag to be interpreted using XMLLang.
Explicitely specify saved namespace declarations. See rdf_save_header/2 option namespaces for details.
True if DB is the name of a graph with at least one triple.
rdf_graph_property(?Graph, ?Property)
rue when Property is a property of Graph. Defined properties are:
Hash is the (MD5-)hash for the content of Graph. See section 3.4.4 for details.
The graph is loaded from the Source.
Time is the last-modified timestamp of Source at the moment that the graph was loaded from Source.
True when Count is the number of triples in Graph.
Deprecated. Use rdf_graph/1 or rdf_source/2 in new code.
rdf_source(?DB, ?SourceURL)
True if the named graph DB was loaded from the source SourceURL. A named graph is associated with a SourceURL by rdf_load/2. The association is stored in the internal binary format, which ensures proper maintenance of the original source through caching and the persistency layer.
Re-load all RDF sourcefiles (see rdf_source/1) that have changed since they were loaded the last time. This implies all triples that originate from the file are removed and the file is re-loaded. If the file is cached a new cache-file is written. Please note that the new triples are added at the end of the database, possibly changing the order of (conflicting) triples.

3.4.1 library(semweb/rdf_cache): Caching triples

The library library(semweb/rdf_cache) defines the caching strategy for triples sources. When using large RDF sources, caching triples greatly speedup loading RDF documents. The cache library implements two caching strategies that are controlled by rdf_set_cache_options/1.

Local caching This approach applies to files only. Triples are cached in a sub-directory of the directory holding the source. This directory is called .cache (_cache on Windows). If the cache option create_local_directory is true, a cache directory is created if posible.

Global caching This approach applies to all sources, except for unnamed streams. Triples are cached in directory defined by the cache option global_directory.

When loading an RDF file, the system scans the configured cache files unless cache(false) is specified as option to rdf_load/2 or caching is disabled. If caching is enabled but no cache exists, the system will try to create a cache file. First it will try to do this locally. On failure it will try to configured global cache.

Set cache options. Defined options are:
If true (default), caching is enabled.
Local directory to use for caching. Default .cache (Windows: _cache).
If true (default false), create a local cache directory if none exists and the directory can be created.
Global directory to use for caching. The directory is created if the option create_global_directory is also given and set to true. Sub-directories are created to speedup indexing on filesystems that perform poorly on directories with large numbers of files. Initially not defined.
If true (default false), create a global cache directory if none exists.

3.4.2 Partial save

Sometimes it is necessary to make more arbitrary selections of material to be saved or exchange RDF descriptions over an open network link. The predicates in this section provide for this. Character encoding issues are derived from the encoding of the Stream, providing support for utf8, iso_latin_1 and ascii.

rdf_save_header(+Stream, +Options)
Save an RDF header, with the XML header, DOCTYPE, ENTITY and opening the rdf:RDF element with appropriate namespace declarations. It uses the primitives from section 3.5 to generate the required namespaces and desired short-name. Options is one of:
Only search for namespaces used in triples that belong to the given named graph.
Deprecated synonym for graph(FileRef).
Where List is a list of namespace abbreviations (see section 3.5). With this option, the expensive search for all namespaces that may be used by your data is omitted. The namespaces rdf and rdfs are added to the provided List. If a namespace is not declared, the resource is emitted in non-abreviated form.
Close the work opened with rdf_save_header/2.
rdf_save_subject(+Stream, +Subject, :Options)
Save everything known about Subject. Options:
Only save properties from Graph.
See rdf_save/2 for a description of these options.
rdf_quote_uri(+URI, -Quoted)
Quote a UNICODE URI. First the Unicode is represented as UTF-8 and then the unsafe characters are mapped to be represented as US-ASCII.

3.4.3 Fast loading and saving

Loading and saving RDF format is relatively slow. For this reason we designed a binary format that is more compact, avoids the complications of the RDF parser and avoids repetitive lookup of (URL) identifiers. Especially the speed improvement of about 25 times is worth-while when loading large databases. These predicates are used for caching by rdf_load/[1,2] under certain conditions.

Save all known triples into File. The saved version includes the SourceRef information.
rdf_save_db(+File, +FileRef)
Save all triples with SourceRef FileRef, regardless of the line-number. For example, using user all information added using rdf_assert/3 is stored in the database.
Load triples from File.

3.4.4 MD5 digests

The rdf_db library provides for MD5 digests. An MD5 digest is a 128 bit long hash key computed from the triples based on the RFC-1321 standard. MD5 keys are computed for each individual triple and added together to compute the final key, resulting in a key that describes the triple-set but is independant from the order in which the triples appear. It is claimed that it is practically impossible for two different datasets to generate the same MD5 key. The Triple20 editor uses the MD5 key for detecting whether the triples associated to a file have changed as well as to maintain a directory with snapshots of versioned ontology files.

rdf_md5(+Graph, -MD5)
Return the MD5 digest for all triples in the database associated to Graph. The MD5 digest itself is represented as an atom holding a 32-character hexadecimal string. The library maintains the digest incrementally on rdf_load/[1,2], rdf_load_db/1, rdf_assert/[3,4] and rdf_retractall/[3,4]. Checking whether the digest has changed since the last rdf_load/[1,2] call provides a practical means for checking whether the file needs to be saved.

Deprecated. New code should use rdf_graph_property(Graph, hash(Hash)).

rdf_atom_md5(+Text, +Times, -MD5)
Computes the MD5 hash from Text, which is an atom, string or list of character codes. Times is an integer >= 1. When > 0, the MD5 algorithm is repeated Times times on the generated hash. This can be used for password encryption algorithms to make generate-and-test loops slow.

Deprecated. New code should use the library(crypt) library provided by the clib package.

3.5 Namespace Handling

Prolog code often contains references to constant resources in a known XML namespace. For example, http://www.w3.org/2000/01/rdf-schema#Class refers to the most general notion of a class. Readability and maintability concerns require for abstraction here. The dynamic and multifile predicate rdf_db:ns/2 maintains a mapping between short meaningful names and namespace locations very much like the XML xmlns construct. The initial mapping contains the namespaces required for the semantic web languages themselves:

ns(rdf,     'http://www.w3.org/1999/02/22-rdf-syntax-ns#').
ns(rdfs,    'http://www.w3.org/2000/01/rdf-schema#').
ns(owl,     'http://www.w3.org/2002/7/owl#').
ns(xsd,     'http://www.w3.org/2000/10/XMLSchema#').
ns(dc,      'http://purl.org/dc/elements/1.1/').
ns(dcterms, 'http://purl.org/dc/terms/').
ns(skos,    'http://www.w3.org/2004/02/skos/core#').
ns(eor,     'http://dublincore.org/2000/03/13/eor#').

All predicates for the semweb libraries use goal_expansion/2 rules to make the SWI-Prolog compiler rewrite terms of the form Id : Local into the fully qualified URL. In addition, the following predicates are supplied:

rdf_equal(Resource1, Resource2)
Defined as Resource1 = Resource2. As this predicate is subject to goal-expansion it can be used to obtain or test global URL values to readable values. The following goal unifies X with http://www.w3.org/2000/01/rdf-schema#Class without more runtime overhead than normal Prolog unification.
        rdf_equal(rdfs:'Class', X)
[nondet]rdf_current_ns(?Alias, ?URI)
Query defined namespace aliases (prefixes).8Older versions of this library did not export the table rdf_db:ns/2. Please use this new public interface.
rdf_register_ns(+Alias, +URL)
Same as rdf_register_ns(Alias, URL,[]).
rdf_register_ns(+Alias, +URL, +Options)
Register Alias as a shorthand for URL. Note that the registration must be done before loading any files using them as namespace aliases are handled at compiletime through goal_expansion/2. If Alias already exists the default is to raise a permission error. If the option force(true) is provided, the alias is silently modified. Rebinding an alias must be done before any code is compiled that relies on the alias. If the option keep(true) is provided the new registration is silently ignored.
rdf_global_id(?Alias:Local, ?Global)
Runtime translation between Alias and Local and a Global URL. Expansion is normally done at compiletime. This predicate is often used to turn a global URL into a more readable term.
rdf_global_object(?Object, ?NameExpandedObject)
As rdf_global_id/2, but also expands the type field if the object is of the form literal(type(Type, Value)). This predicate is used for goal expansion of the object fields in rdf/3 and similar goals.
rdf_global_term(+Term0, -Term)
Expands all Alias:Local in Term0 and return the result in Term. Use infrequently for runtime expansion of namespace identifiers.

3.5.1 Namespace handling for custom predicates

If we implement a new predicate based on one of the predicates of the semweb libraries that expands namespaces, namespace expansion is not automatically available to it. Consider the following code computing the number of distinct objects for a certain property on a certain object.

cardinality(S, P, C) :-
        (   setof(O, rdf_has(S, P, O), Os)
        ->  length(Os, C)
        ;   C = 0

Now assume we want to write labels/2 that returns the number of distict labels of a resource:

labels(S, C) :-
        cardinality(S, rdfs:label, C).

This code will not work as rdfs:label is not expanded at compile time. To make this work, we need to add an rdf_meta/1 declaration.

:- rdf_meta
This predicate defines the argument types of the named predicates, which will force compile time namespace expansion for these predicates. Heads is a coma-separated list of callable terms. Defined argument properties are:
Argument is a goal. The goal is processed using expand_goal/2, recursively applying goal transformation on the argument.
The argument is instantiated at entry. Nothing is changed.
The argument is not instantiated at entry. Nothing is changed.
The argument is unbound or instantiated at entry. Nothing is changed.
The argument is not changed.
The argument must be a resource. If it is a term <namespace>:<local> it is translated.
The argument is an object or resource.
The argument is a term that must be translated. Expansion will translate all occurences of <namespace>:<local> appearing anywhere in the term.

As it is subject to term_expansion/2, the rdf_meta/1 declaration can only be used as a directive. The directive must be processed before the definition of the predicates as well as before compiling code that uses the rdf meta-predicates. The atom rdf_meta is declared as an operator exported from library rdf_db.pl. Files using rdf_meta/1 must explicitely load rdf_db.pl. The example below defines the rule concept/1.

:- use_module(library(semweb/rdf_db)).  % for rdf_meta
:- use_module(library(semweb/rdfs)).    % for rdfs_individual_of

:- rdf_meta

%%      concept(?C) is nondet.
%       True if C is a concept.

concept(C) :-
        rdfs_individual_of(C, skos:'Concept').

In addition to expanding calls, rdf_meta/1 also causes expansion of clause-heads for predicates that match a declaration. This is typically used write Prolog statements about resources. The following example produces three clauses with expanded (single-atom) arguments:

:- use_module(library(semweb/rdf_db)).

:- rdf_meta


3.6 Monitoring the database

Considering performance and modularity, we are working on a replacement of the rdf_edit (see section 12) layered design to deal with updates, journalling, transactions, etc. Where the rdf_edit approach creates a single layer on top of rdf_db and code using the RDF database must select whether to use rdf_db.pl or rdf_edit.pl, the new approach allows to register monitors. This allows multiple modules to provide additional services, while these services will be used regardless of how the database is modified.

Monitors are used by the persistency library (section 4.5) and the literal indexing library (section 4.4).

rdf_monitor(:Goal, +Mask)
Goal is called for modifications of the database. It is called with a single argument that describes the modification. Defined events are:
assert(+S, +P, +O, +DB)
A triple has been asserted.
retract(+S, +P, +O, +DB)
A triple has been deleted.
update(+S, +P, +O, +DB, +Action)
A triple has been updated.
A new literal has been created. Literal is the argument of literal(Arg) of the triple's object. This event is introduced in version 2.5.0 of this library.
The literal Literal is no longer used by any triple.
transaction(+BeginOrEnd, +Id)
Mark begin or end of the commit of a transaction started by rdf_transaction/2. BeginOrEnd is begin(Nesting) or end(Nesting). Nesting expresses the nesting level of transactions, starting at `0' for a toplevel transaction. Id is the second argument of rdf_transaction/2. The following transaction Ids are pre-defined by the library:
A file is loaded using rdf_load/2. Id is one of file(Path) or stream(Stream).
All triples with source DB are being unloaded using rdf_unload/1.
Issued by rdf_reset_db/0.
load(+BeginOrEnd, +Spec)
Mark begin or end of rdf_load_db/1 or load through rdf_load/2 from a cached file. Spec is currently defined as file(Path).
Marks begin/end of a re-hash due to required re-indexing or garbage collection.

Mask is a list of events this monitor is interested in. Default (empty list) is to report all events. Otherwise each element is of the form +Event or -Event to include or exclude monitoring for certain events. The event-names are the functor names of the events described above. The special name all refers to all events and assert(load) to assert events originating from rdf_load_db/1. As loading triples using rdf_load_db/1 is very fast, monitoring this at the triple level may seriously harm performance.

This predicate is intended to maintain derived data, such as a journal, information for undo, additional indexing in literals, etc. There is no way to remove registered monitors. If this is required one should register a monitor that maintains a dynamic list of subscribers like the XPCE broadcast library. A second subscription of the same hook predicate only re-assignes the mask.

The monitor hooks are called in the order of registration and in the same thread that issued the database manipulation. To process all changes in one thread they should be send to a thread message queue. For all updating events, the monitor is called while the calling thread has a write lock on the RDF store. This implies that these events are processed strickly synchronous, even if modifications originate from multiple threads. In particular, the transaction begin, ... updates ... end sequence is never interleaved with other events. Same for load and parse.

3.7 Miscellaneous predicates

This section describes the remaining predicates of the rdf_db module.

Generate a unique reference. The returned atom is guaranteed not to occur in the current database in any field of any triple.
Generate a unique blank node reference. The returned atom is guaranteed not to occur in the current database in any field of any triple and starts with '__bnode'.
Succeeds if Id is a blank node identifier (also called anonymous resource). In the current implementation this implies it is an atom starting with a double underscore.
Succeeds if Id is a resource. Note that this resource need not to appear in any triple.
Succeeds if Id is an RDF literal term. Note that this literal need not to appear in any triple.
rdf_source_location(+Subject, -SourceRef)
Return the source-location as File:Line of the first triple that is about Subject.
Returns the Generation of the database. Each modification to the database increments the generation. It can be used to check the validity of cached results deduced from the database. Modifications changing multiple triples increment Generation with the number of triples modified, providing a heuristic for `how dirty' cached results may be.
rdf_estimate_complexity(?Subject, ?Predicate, ?Object, -Complexity)
Return the number of alternatives as indicated by the database internal hashed indexing. This is a rough measure for the number of alternatives we can expect for an rdf_has/3 call using the given three arguments. When called with three variables, the total number of triples is returned. This estimate is used in query optimisation. See also rdf_predicate_property/2 and rdf_statistics/1 for additional information to help optimisers.
Report statistics collected by the rdf_db module. Defined values for Statistics are:
lookup(?Index, -Count)
Number of lookups using a pattern of instantiated fields. Index is a term rdf(S,P,O), where S, P and O are either + or -. For example rdf(+,+,-) returns the lookups with subject and predicate specified and object unbound.
Number of unique values for the second field of the triple set.
Number of files loaded through rdf_load/1.
Number of unique values for the first field of the triple set.
Total number of unique literal values in the database. See also section 3.1.1.
Total number of triples in the database.
triples_by_file(?File, -Count)
Enumerate the number of triples associated to each file.
Number of nodes explored in rdf_reachable/3.
gc(-Count, -Time)
Number of garbage collections and time spent in seconds represented as a float.
rehash(-Count, -Time)
Number of times the hash-tables were enlarged and time spent in seconds represented as a float.
Core used by the triple store. This includes all memory allocated on behalf of the library, but not the memory allocated in Prolog atoms referenced (only) by the triple store.
rdf_match_label(+Method, +Search, +Atom)
True if Search matches Atom as defined by Method. All matching is performed case-insensitive. Defines methods are:
Perform exact, but case-insensitive match.
Search is a sub-string of Text.
Search appears as a whole-word in Text.
Text start with Search.
Text matches Search, case insensitively, where the `*' character in Search matches zero or more characters.
lang_matches(+Lang, +Pattern)
True if Lang matches Pattern. This implements XML language matching conform RFC 4647. Both Lang and Pattern are dash-separated strings of identifiers or (for Pattern) the wildcart *. Identifiers are matched case-insensitive and a * matches any number of identifiers. A short pattern is the same as *.
lang_equal(+Lang1, +Lang2)
True if Lang1 and Lang2 specify the same language, including regional and other modifiers. Language-specifiers are case-insensitive.
Erase all triples from the database and reset all counts and statistics information.
Unify Version with the library version number. This number is, like to the SWI-Prolog version flag, defined as 10,000 × Major + 100 × Minor + Patch.

3.8 Issues with rdf_db

This RDF low-level module has been created after two year experimenting with a plain Prolog based module and a brief evaluation of a second generation pure Prolog implementation. The aim was to be able to handle upto about 5 million triples on standard (notebook) hardware and deal efficiently with subPropertyOf which was identified as a crucial feature of RDFS to realise fusion of different data-sets.

The following issues are identified and not solved in suitable manner.

subPropertyOf of subPropertyOf
is not supported.
Similar to subPropertyOf, it is likely to be profitable to handle resource identity efficient. The current system has no support for it.

4 Plugin modules for rdf_db

The library(rdf_db) module provides several hooks for extending its functionality. Database updates can be monitored and acted upon through the features described in section 3.6. The predicate rdf_load/2 can be hooked to deal with different formats such as rdfturtle, different input sources (e.g. http) and different strategies for caching results.

4.1 Hooks into the RDF library

The hooks below are used to add new RDF file formats and sources from which to load data to the library. They are used by the modules described below and distributed with the package. Please examine the source-code if you want to add new formats or locations.

Load files in the Turtle format. See section 5.
Load gzip compressed files transparently. See section 4.2.
Load RDF documents from HTTP servers. See section 4.3.
rdf_db:rdf_open_hook(+Input, -Stream, -Format)
Open an input. Input is one of file(+Name), stream(+Stream) or url(Protocol, URL). If this hook succeeds, the RDF will be read from Stream using rdf_load_stream/3. Otherwise the default open functionality for file and stream are used.
rdf_db:rdf_load_stream(+Format, +Stream, +Options)
Actually load the RDF from Stream into the RDF database. Format describes the format and is produced either by rdf_input_info/3 or rdf_file_type/2.
rdf_db:rdf_input_info(+Input, -Modified, -Format)
Gather information on Input. Modified is the last modification time of the source as a POSIX time-stamp (see time_file/2). Format is the RDF format of the file. See rdf_file_type/2 for details. It is allowed to leave the output variables unbound. Ultimately the default modified time is `0' and the format is assumed to be xml.
rdf_db:rdf_file_type(?Extension, ?Format)
True if Format is the default RDF file format for files with the given extension. Extension is lowercase and without a '.'. E.g. owl. Format is either a built-in format (xml or triples) or a format understood by the rdf_load_stream/3 hook.
True if Protocol is a URL protocol recognised by rdf_load/2.

4.2 library(semweb/rdf_zlib_plugin): Reading compressed RDF

This module uses the library(zlib) library to load compressed files on the fly. The extension of the file must be .gz. The file format is deduced by the extension after stripping the .gz extension. E.g. rdf_load('file.rdf.gz').

4.3 library(semweb/rdf_http_plugin): Reading RDF from a HTTP server

This module allows for rdf_load('http://...'). It exploits the library library(http/http_open.pl). The format of the URL is determined from the mime-type returned by the server if this is one of text/rdf+xml, application/x-turtle or application/turtle. As RDF mime-types are not yet widely supported, the plugin uses the extension of the URL if the claimed mime-type is not one of the above. In addition, it recognises text/html and application/xhtml+xml, scanning the XML content for embedded RDF.

4.4 library(semweb/rdf_litindex): Indexing words in literals

The library library(semweb/rdf_litindex.pl) exploits the primitives of section 4.4.1 and the NLP package to provide indexing on words inside literal constants. It also allows for fuzzy matching using stemming and `sounds-like' based on the double metaphone algorithm of the NLP package.

rdf_find_literals(+Spec, -ListOfLiterals)
Find literals (without type or language specification) that satisfy Spec. The required indices are created as needed and kept up-to-date using hooks registered with rdf_monitor/2. Numerical indexing is currently limited to integers in the range ±2^30 (±2^62 on 64-bit platforms). Spec is defined as:
and(Spec1, Spec2)
Intersection of both specifications.
or(Spec1, Spec2)
Union of both specifications.
Negation of Spec. After translation of the full specification to Disjunctive Normal Form (DNF), negations are only allowed inside a conjunction with at least one positive literal.
Matches all literals containing the word Word, doing the match case insensitive and after removing diacritics.
Matches all literals containing at least one word that has the same stem as Like using the Porter stem algorithm. See NLP package for details.
Matches all literals containing at least one word that `sounds like' Like using the double metaphone algorithm. See NLP package for details.
Matches all literals containing at least one word that starts with Prefix, discarding diacritics and case.
between(Low, High)
Matches all literals containing an integer token in the range Low..High, including the boundaries.
Matches all literals containing an integer token with value Low or higher.
Matches all literals containing an integer token with value High or lower.
Matches all literals containing the given token. See tokenize_atom/2 of the NLP package for details.
rdf_token_expansions(+Spec, -Expansions)
Uses the same database as rdf_find_literals/2 to find possible expansions of Spec, i.e. which words `sound like', `have prefix', etc. Spec is a compound expression as in rdf_find_literals/2. Expansions is unified to a list of terms sounds(Like, Words), stem(Like, Words) or prefix(Prefix, Words). On compound expressions, only combinations that provide literals are returned. Below is an example after loading the ULAN9Unified List of Artist Names from the Getty Foundation. database and showing all words that sounds like `rembrandt' and appear together in a literal with the word `Rijn'. Finding this result from the 228,710 literals contained in ULAN requires 0.54 milliseconds (AMD 1600+).
?- rdf_token_expansions(and('Rijn', sounds(rembrandt)), L).

L = [sounds(rembrandt, ['Rambrandt', 'Reimbrant', 'Rembradt',
                        'Rembrand', 'Rembrandt', 'Rembrandtsz',
                        'Rembrant', 'Rembrants', 'Rijmbrand'])]

Here is another example, illustrating handling of diacritics:

?- rdf_token_expansions(case(cafe), L).

L = [case(cafe, [cafe, caf\'e])]
rdf_tokenize_literal(+Literal, -Tokens)
Tokenize a literal, returning a list of atoms and integers in the range -1073741824 ... 1073741823. As tokenization is in general domain and task-dependent this predicate first calls the hook rdf_litindex:tokenization(Literal, -Tokens). On failure it calls tokenize_atom/2 from the NLP package and deletes the following: atoms of length 1, floats, integers that are out of range and the english words and, an, or, of, on, in, this and the. Deletion first calls the hook rdf_litindex:exclude_from_index(token, X). This hook is called as follows:
no_index_token(X) :-
        exclude_from_index(token, X), !.
no_index_token(X) :-

4.4.1 Literal maps: Creating additional indices on literals

`Literal maps' provide a relation between literal values, intended to create additional indexes on literals. The current implementation can only deal with integers and atoms (string literals). A literal map maintains an ordered set of keys. The ordering uses the same rules as described in section 3.1.1. Each key is associated with an ordered set of values. Literal map objects can be shared between threads, using a locking strategy that allows for multiple concurrent readers.

Typically, this module is used together with rdf_monitor/2 on the channals new_literal and old_literal to maintain an index of words that appear in a literal. Further abstraction using Porter stemming or Metaphone can be used to create additional search indices. These can map either directly to the literal values, or indirectly to the plain word-map. The SWI-Prolog NLP package provides complimentary building blocks, such as a tokenizer, Porter stem and Double Metaphone.

Create a new literal map, returning an opaque handle.
Destroy a literal map. After this call, further use of the Map handle is illegal. Additional synchronisation is needed if maps that are shared between threads are destroyed to guarantee the handle is no longer used. In some scenarios rdf_reset_literal_map/1 provides a safe alternative.
Delete all content from the literal map.
rdf_insert_literal_map(+Map, +Key, +Value)
Add a relation between Key and Value to the map. If this relation already exists no action is performed.
rdf_insert_literal_map(+Map, +Key, +Value, -KeyCount)
As rdf_insert_literal_map/3. In addition, if Key is a new key in Map, unify KeyCount with the number of keys in Map. This serves two purposes. Derived maps, such as the stem and metaphone maps need to know about new keys and it avoids additional foreign calls for doing the progress in rdf_litindex.pl.
rdf_delete_literal_map(+Map, +Key)
Delete Key and all associated values from the map. Succeeds always.
rdf_delete_literal_map(+Map, +Key, +Value)
Delete the association between Key and Value from the map. Succeeds always.
[det]rdf_find_literal_map(+Map, +KeyList, -ValueList)
Unify ValueList with an ordered set of values associated to all keys from KeyList. Each key in KeyList is either an atom, an integer or a term not(Key). If not-terms are provided, there must be at least one positive keywords. The negations are tested after establishing the positive matches.
rdf_keys_in_literal_map(+Map, +Spec, -Answer)
Realises various queries on the key-set:
Unify Answer with an ordered list of all keys.
Succeeds if Key is a key in the map and unify Answer with the number of values associated with the key. This provides a fast test of existence without fetching the possibly large associated value set as with rdf_find_literal_map/3.
Unify Answer with an ordered set of all keys that have the given prefix. See section 3.1 for details on prefix matching. Prefix must be an atom. This call is intended for auto-completion in user interfaces.
Unify Answer with all keys that are larger or equal to the integer Min.
Unify Answer with all keys that are smaller or equal to the integer Max.
between(+Min, +Max)
Unify Answer with all keys between Min and Max (including).
rdf_statistics_literal_map(+Map, +Key(-Arg...))
Query some statistics of the map. Provides keys are:
size(-Keys, -Relations)
Unify Keys with the total key-count of the index and Relation with the total Key-Value count.

4.5 library(semweb/rdf_persistency): Providing persistent storage

The library(semweb/rdf_persistency) provides reliable persistent storage for the RDF data. The store uses a directory with files for each source (see rdf_source/1) present in the database. Each source is represented by two files, one in binary format (see rdf_save_db/2) representing the base state and one represented as Prolog terms representing the changes made since the base state. The latter is called the journal.

rdf_attach_db(+Directory, +Options)
Attach Directory as the persistent database. If Directory does not exist it is created. Otherwise all sources defined in the directory are loaded into the RDF database. Loading a source means loading the base state (if any) and replaying the journal (if any). The current implementation does not synchronise triples that are in the store before attaching a database. They are not removed from the database, nor added to the presistent store. Different merging options may be supported through the Options argument later. Currently defined options are:
Number of threads used to reload databased and journals from the files in Directory. Default is the number of physical CPUs determined by the Prolog flag cpu_count or 1 (one) on systems where this number is unknown. See also concurrent/3.
The library maintains a pool of open journal files. This option specifies the size of this pool. The default is 10. Raising the option can make sense if many writes occur on many different named graphs. The value can be lowered for scenarios where write operations are very infrequent.
If true, supress loading messages from rdf_attach_db/2.
If true, nested log transactions are added to the journal information. By default (false), no log-term is added for nested transactions.

The database is locked against concurrent access using a file lock in Directory. An attempt to attach to a locked database raises a permission_error exception. The error context contains a term rdf_locked(Args), where args is a list containing time(Stamp) and pid(PID). The error can be caught by the application. Otherwise it prints:

ERROR: No permission to lock rdf_db `/home/jan/src/pl/packages/semweb/DB'
ERROR: locked at Wed Jun 27 15:37:35 2007 by process id 1748
Detaches the persistent store. No triples are removed from the RDF triple store.
Unify Directory with the current database directory. Fails if no persistent database is attached.
rdf_persistency(+DB, +Bool)
Change presistency of named database (4th argument of rdf/4). By default all databases are presistent. Using false, the journal and snapshot for the database are deleted and further changes to triples associated with DB are not recorded. If Bool is true a snapshot is created for the current state and further modifications are monitored. Switching persistency does not affect the triples in the in-memory RDF database.
Flush dirty journals. With the option min_size(KB) only journals larger than KB Kbytes are merged with the base state. Flushing a journal takes the following steps, ensuring a stable state can be recovered at any moment.
  1. Save the current database in a new file using the extension .new.
  2. On success, delete the journal
  3. On success, atomically move the .new file over the base state.

Note that journals are not merged automatically for two reasons. First of all, some applications may decide never to merge as the journal contains a complete changelog of the database. Second, merging large databases can be slow and the application may wish to schedule such actions at quiet times or scheduled maintenance periods.

4.5.1 Enriching the journals

The above predicates suffice for most applications. The predicates in this section provide access to the journal files and the base state files and are intented to provide additional services, such as reasoning about the journals, loaded files, etc.10A library library(rdf_history) is under development exploiting these features supporting wiki style editing of RDF.

Using rdf_transaction(Goal, log(Message)), we can add additional records to enrich the journal of affected databases with Term and some additional bookkeeping information. Such a transaction adds a term begin(Id, Nest, Time, Message) before the change operations on each affected database and end(Id, Nest, Affected) after the change operations. Here is an example call and content of the journal file mydb.jrn. A full explanation of the terms that appear in the journal is in the description of rdf_journal_file/2.

?- rdf_transaction(rdf_assert(s,p,o,mydb), log(by(jan))).
begin(1, 0, 1183540570.36, by(jan)).
assert(s, p, o).
end(1, 0, []).

Using rdf_transaction(Goal, log(Message, DB)), where DB is an atom denoting a (possibly empty) named graph, the system guarantees that a non-empty transaction will leave a possibly empty transaction record in DB. This feature assumes named graphs are named after the user making the changes. If a user action does not affect the user's graph, such as deleting a triple from another graph, we still find record of all actions performed by some user in the journal of that user.

rdf_journal_file(?DB, ?JournalFile)
True if File is the absolute file name of an existing named graph DB. A journal file contains a sequence of Prolog terms of the following format.11Future versions of this library may use an XML based language neutral format.
Journal has been opened. Currently Attributes contains a term time(Stamp).
Journal was closed. Currently Attributes contains a term time(Stamp).
assert(Subject, Predicate, Object)
A triple {Subject, Predicate, Object} was added to the database.
assert(Subject, Predicate, Object, Line)
A triple {Subject, Predicate, Object} was added to the database with given Line context.
retract(Subject, Predicate, Object)
A triple {Subject, Predicate, Object} was deleted from the database. Note that an rdf_retractall/3 call can retract multiple triples. Each of them have a record in the journal. This allows for `undo'.
retract(Subject, Predicate, Object, Line)
Same as above, for a triple with associated line info.
update(Subject, Predicate, Object, Action)
See rdf_update/4.
begin(Id, Nest, Time, Message)
Added before the changes in each database affected by a transaction with transaction identifier log(Message). Id is an integer counting the logged transactions to this database. Numbers are increasing and designed for binary search within the journal file. Nest is the nesting level, where `0' is a toplevel transaction. Time is a time-stamp, currently using float notation with two fractional digits. Message is the term provided by the user as argument of the log(Message) transaction.
end(Id, Nest, Others)
Added after the changes in each database affected by a transaction with transaction identifier log(Message). Id and Nest match the begin-term. Others gives a list of other databases affected by this transaction and the Id of these records. The terms in this list have the format DB:Id.
rdf_db_to_file(?DB, ?FileBase)
Convert between DB (see rdf_source/1) and file base-file used for storing information on this database. The full file is located in the directory described by rdf_current_db/1 and has the extension .trp for the base state and .jrn for the journal.

5 library(semweb/rdf_turtle): Turtle: Terse RDF Triple Language

To be done
Better error handling

This module implements the Turtle language for representing the RDF triple model as defined by Dave Beckett from the Institute for Learning and Research Technology University of Bristol in the document:

This parser passes all tests, except for test-28.ttl (decial number serialization) and test-29.ttl (uri containing ...%&...). It is unclear to me whether these tests are correct. Notably, it is unclear whether we must do %-decoding. Certainly, this is expected by various real-life datasets that we came accross with.

This module acts as a plugin to rdf_load/2, for processing files with one of the extensions .ttl, .n3 or .nt.

rdf_read_turtle(+Input, -Triples, +Options)
Read a stream or file into a set of triples of the format

rdf(Subject, Predicate, Object)

The representation is consistent with the SWI-Prolog RDF/XML and ntriples parsers. Provided options are:

Initial base URI. Defaults to file://<file> for loading files.
Blank nodes are generated as <Prefix>1, <Prefix>2, etc. If Prefix is not an atom blank nodes are generated as node(1), node(2), ...
Officially, Turtle resources are IRIs. Quite a few applications however send URIs. By default we do URI->IRI mapping because this rarely causes errors. To force strictly conforming mode, pass iri.
Return encountered prefix declarations as a list of Alias-URI
Same as prefixes(Pairs). Compatibility to rdf_load/2.
Base URI used for processing the data. Unified to[]if there is no base-uri.
In warning (default), print the error and continue parsing the remainder of the file. If error, abort with an exception on the first error encountered.
If on_error(warning) is active, this option cane be used to retrieve the number of generated errors.
rdf_load_turtle(+Input, -Triples, +Options)
Use rdf_read_turtle/3
[det]rdf_process_turtle(+Input, :OnObject, +Options)
Process Turtle input from Input, calling OnObject with a list of triples. Options is the same as for rdf_load_turtle/3.

Errors encountered are sent to print_message/2, after which the parser tries to recover and parse the remainder of the data.

6 library(semweb/rdf_turtle_write): Turtle - Terse RDF Triple Language writer

To be done
Low-level string output takes 28% of the time. Move to C?

This module implements the Turtle language for representing the RDF triple model as defined by Dave Beckett from the Institute for Learning and Research Technology University of Bristol in the document:

The Turtle format is designed as an RDF serialization that is easy to read and write by both machines and humans. Due to the latter property, this library goes a long way in trying to produce human-readable output.

In addition to the human-readable format, this library can write a canonical representation of RDF graphs. The canonical representation has the following properties:

[det]rdf_save_turtle(+Out, :Options)
Save an RDF graph as Turtle. Options processed are:
Nicely align the @prefix declarations
Save relative to the given Base
If true (default false), emit numeric datatypes using Prolog's write to achieve canonical output.
It true (default), write some informative comments between the output segments
Encoding used for the output stream. Default is UTF-8.
Query an alternative graph-representation. See below.
Indentation for ; -lists. `0' does not indent, but writes on the same line. Default is 8.
Save only the named graph
If true (default), using P-O and O-grouping.
Only use prefix notation for known prefixes. Without, some documents produce huge amounts of prefixes.
If true (default false), do not print the final informational message.
If true (default false), write [...] and (...) on a single line.
Extra white lines to insert between statements about a different subject. Default is 1.
Distance between tab-stops. `0' forces the library to use only spaces for layout. Default is 8.
If true (default), use prefixes from rdf_current_ns/2.

The option expand allows for serializing alternative graph representations. It is called through call/5, where the first argument is the expand-option, followed by S,P,O,G. G is the graph-option (which is by default a variable). This notably allows for writing RDF graphs represented as rdf(S,P,O) using the following code fragment:

triple_in(RDF, S,P,O,_G) :-
    member(rdf(S,P,O), RDF).

    rdf_save_turtle(Out, [ expand(triple_in(RDF)) ]),
Out is one of stream(Stream), a stream handle, a file-URL or an atom that denotes a filename.
[det]rdf_save_canonical_turtle(+Spec, +Options)
Save triples in a canonical format. This is the same as rdf_save_turtle/3, but using different defaults. In particular:

To be done
Work in progress. Notably blank-node handling is incomplete.

7 library(semweb/rdfs): RDFS related queries

The library(semweb/rdfs) library adds interpretation of the triple store in terms of concepts from RDF-Schema (RDFS). There are two ways to provide support for more high level languages in RDF. One is to view such languages as a set of entailment rules. In this model the rdfs library would provide a predicate rdfs/3 providing the same functionality as rdf/3 on union of the raw graph and triples that can be derived by applying the RDFS entailment rules.

Alternatively, RDFS provides a view on the RDF store in terms of individuals, classes, properties, etc., and we can provide predicates that query the database with this view in mind. This is the approach taken in the library(semweb/rdfs.p)l library, providing calls like rdfs_individual_of(?Resource, ?Class).12The SeRQL language is based on querying the deductive closure of the triple set. The SWI-Prolog SeRQL library provides entailment modules that take the approach outlined above.

7.1 Hierarchy and class-individual relations

The predicates in this section explore the rdfs:subPropertyOf, rdfs:subClassOf and rdf:type relations. Note that the most fundamental of these, rdfs:subPropertyOf, is also used by rdf_has/[3,4].

rdfs_subproperty_of(?SubProperty, ?Property)
True if SubProperty is equal to Property or Property can be reached from SubProperty following the rdfs:subPropertyOf relation. It can be used to test as well as generate sub-properties or super-properties. Note that the commonly used semantics of this predicate is wired into rdf_has/[3,4].bugThe current implementation cannot deal with cycles.bugThe current implementation cannot deal with predicates that are an rdfs:subPropertyOf of rdfs:subPropertyOf, such as owl:samePropertyAs.
rdfs_subclass_of(?SubClass, ?Class)
True if SubClass is equal to Class or Class can be reached from SubClass following the rdfs:subClassOf relation. It can be used to test as well as generate sub-classes or super-classes.bugThe current implementation cannot deal with cycles.
rdfs_class_property(+Class, ?Property)
True if the domain of Property includes Class. Used to generate all properties that apply to a class.
rdfs_individual_of(?Resource, ?Class)
True if Resource is an indivisual of Class. This implies Resource has an rdf:type property that refers to Class or a sub-class thereof. Can be used to test, generate classes Resource belongs to or generate individuals described by Class.

7.2 Collections and Containers

The RDF construct rdf:parseType=Collection constructs a list using the rdf:first and rdf:next relations.

rdfs_member(?Resource, +Set)
Test or generate the members of Set. Set is either an individual of rdf:List or rdf:Container.
rdfs_list_to_prolog_list(+Set, -List)
Convert Set, which must be an individual of rdf:List into a Prolog list of objects.
rdfs_assert_list(+List, -Resource)
Equivalent to rdfs_assert_list/3 using DB = user.
rdfs_assert_list(+List, -Resource, +DB)
If List is a list of resources, create an RDF list Resource that reflects these resources. Resource and the sublist resources are generated with rdf_bnode/1. The new triples are associated with the database DB.

7.3 Labels and textual search

Textual search is partly handled by the predicates from the library(rdf_db) module and its underlying C-library. For example, literal objects are hashed case-insensitive to speed up the commonly used case-insensitive search.

[multi]rdfs_label(?Resource, ?Language, ?Label)
Extract the label from Resource or generate all resources with the given Label. The label is either associated using a sub-property of rdfs:label or it is extracted from Resource by taking the part after the last # or /. If this too fails, Label is unified with Resource. Language is unified to the value of the xml:lang attribute of the label or a variable if the label has no language specified.
rdfs_label(?Resource, ?Label)
Defined as rdfs_label(Resource, _, Label).
rdfs_ns_label(?Resource, ?Language, ?Label)
Similar to rdfs_label/2, but prefixes the result using the declared namespace alias (see section 3.5) to facilitate user-friendly labels in applications using multiple namespaces that may lead to confusion.
rdfs_ns_label(?Resource, ?Label)
Defined as rdfs_ns_label(Resource, _, Label).
rdfs_find(+String, +Description, ?Properties, +Method, -Subject)
Find (on backtracking) Subjects that satisfy a search specification for textual attributes. String is the string searched for. Description is an OWL description (see section 14) specifying candidate resources. Properties is a list of properties to search for literal objects, Method defines the textual matching algorithm. All textual mapping is performed case-insensitive. The matching-methods are described with rdf_match_label/3. If Properties is unbound, the search is performed in any property and Properties is unified with a list holding the property on which the match was found.

8 Managing RDF input files

Complex projects require RDF resources from many locations and typically wish to load these in different combinations. For example loading a small subset of the data for debugging purposes or load a different set of files for experimentation. The library library(semweb/rdf_library.pl) manages sets of RDF files spread over different locations, including file and network locations. The original version of this library supported metadata about collections of RDF sources in an RDF file called Manifest. The current version supports both the VoID format and the original format. VoID files (typically named void.ttl) can use elements from the RDF Manifest vocabulary to support features that are not supported by VoID.

8.1 The Manifest file

A manifest file is an RDF file, often in Turtle format, that provides meta-data about RDF resources. Often, a manifest will describe RDF files in the current directory, but it can also describe RDF resources at arbitrary URL locations. The RDF schema for RDF library meta-data can be found in rdf_library.ttl. The namespace for the RDF library format is defined as http://www.swi-prolog.org/rdf/library/ and abbreviated as lib.

The schema defines three root classes: lib:Namespace, lib:Ontology and lib:Virtual, which we describe below.

This is a subclass of owl:Ontology. It has two subclasses, lib:Schema and lib:Instances. These three classes are currently processed equally. The following properties are recognised on lib:Ontology:
Title of the ontology. Displayed by rdf_list_library/0.
Version of the ontology. Displayed by rdf_list_library/0.
Ontologies imported. If rdf_load_library/2 is used to load this ontology, the ontologies referenced here are loaded as well. There are two subProperties: lib:schema and lib:instances with the obvious meaning.
Defines the named graph into which the resource is loaded. If this ends in a /, the basename of each loaded file is appended to the given source. Defaults to the URL the RDF is loaded from.
Defines the base for processing the RDF data. If not provided this defaults to the named graph, which in turn defaults to the URL the RDF is loaded from.
Virtual ontologies do not refer to an RDF resource themselves. They only import other resources. For example the W3C WordNet manifest defines wn-basic and wn-full as virtual resources. The lib:Virtual resource is used as a second rdf:type:
        a lib:Ontology ;
        a lib:Virtual ;
Used by ClioPatria to combine this ontology and all data it imports into a node in the automatically generated datacloud.
Defines a URL to be a namespace. The definition provides the preferred mnemonic and can be referenced in the lib:providesNamespace and lib:usesNamespace properties. The rdf_load_library/2 predicates registers encountered namespace mnemonics with rdf-db using rdf_register_ns/2. Typically namespace declarations use @prefix declarations. E.g.
@prefix     lib: <http://www.swi-prolog.org/rdf/library/> .
@prefix    rdfs: <http://www.w3.org/2000/01/rdf-schema#> .

[ a lib:Namespace ;
  lib:mnemonic "rdfs" ;
  lib:namespace rdfs:
] .

8.1.1 Support for the VoID and VANN vocabularies

The VoID aims at resolving the same problem as the Manifest files described here. In addition, the VANN vocabulary provides the information about preferred namepaces prefixes. The RDF library manager can deal with VoID files. The following relations apply:

Currently, the RDF metadata is not stored in the RDF database. It is processed by low-level primitives that do not perform RDFS reasoning. In particular, this means that rdfs:supPropertyOf and rdfs:subClassOf cannot be used to specialise the RDF meta vocabulary.

8.1.2 Finding manifest files

The initial metadata file(s) are loaded into the system using rdf_attach_library/1.

Load meta-data on RDF repositories from FileOrDirectory. If the argument is a directory, this directory is processed recursively and each for each directory, a file named void.ttl, Manifest.ttl or Manifest.rdf is loaded (in this order of preference).

Declared namespaces are added to the rdf-db namespace list. Encountered ontologies are added to a private database of rdf_list_library.pl. Each ontology is given an identifier, derived from the basename of the URL without the extension. This, using the declaration below, the identifier of the declared ontology is wn-basic.

        a void:Dataset ;
        dcterms:title "Basic WordNet" ;
List the available resources in the library. Currently only lists resources that have a dcterms:title property. See section 8.2 for an example.

It is possible for the initial set of manifests to refer to RDF files that are not covered by a manifest. If such a reference is encountered while loading or listing a library, the library manager will look for a manifest file in the directory holding the referenced RDF file and load this manifest. If a manifest is found that covers the referenced file, the directives found in the manifest will be followed. Otherwise the RDF resource is simply loaded using the current defaults.

Further exploration of the library is achieved using rdf_list_library/1 or rdf_list_library/2:

Same as rdf_list_library(Id,[]).
rdf_list_library(+Id, +Options)
Lists the resources that will be loaded if Id is handed to rdf_load_library/2. See rdf_attach_library/2 for how ontology identifiers are generated. In addition it checks the existence of each resource to help debugging library dependencies. Before doing its work, rdf_list_library/2 reloads manifests that have changed since they were loaded the last time. For HTTP resources it uses the HEAD method to verify existence and last modification time of resources.
rdf_load_library(+Id, +Options)
Load the given library. First rdf_load_library/2 will establish what resources need to be loaded and whether all resources exist. Than it will load the resources.

8.2 Usage scenarios

Typically, a project will use a single file using the same format as a manifest file that defines alternative configurations that can be loaded. This file is loaded at program startup using rdf_attach_library/1. Users can now list the available libraries using rdf_list_libraries/0 and rdf_list_libraries/1:

1 ?- rdf_list_library.
ec-core-vocabularies E-Culture core vocabularies
ec-all-vocabularies All E-Culture vocabularies
ec-hacks            Specific hacks
ec-mappings         E-Culture ontology mappings
ec-core-collections E-Culture core collections
ec-all-collections  E-Culture all collections
ec-medium           E-Culture medium sized data (artchive+aria)
ec-all              E-Culture all data

Now we can list a specific category using rdf_list_library/1. Note this loads two additional manifests referenced by resources encountered in ec-mappings. If a resource does not exist is is flagged using [NOT FOUND].

2 ?- rdf_list_library('ec-mappings').
% Loaded RDF manifest /home/jan/src/eculture/vocabularies/mappings/Manifest.ttl
% Loaded RDF manifest /home/jan/src/eculture/collections/aul/Manifest.ttl
. <file:///home/jan/src/eculture/vocabularies/mappings/mappings>
. . <file:///home/jan/src/eculture/vocabularies/mappings/interface>
. . . file:///home/jan/src/eculture/vocabularies/mappings/interface_class_mapping.ttl
. . . file:///home/jan/src/eculture/vocabularies/mappings/interface_property_mapping.ttl
. . <file:///home/jan/src/eculture/vocabularies/mappings/properties>
. . . file:///home/jan/src/eculture/vocabularies/mappings/ethnographic_property_mapping.ttl
. . . file:///home/jan/src/eculture/vocabularies/mappings/eculture_properties.ttl
. . . file:///home/jan/src/eculture/vocabularies/mappings/eculture_property_semantics.ttl
. . <file:///home/jan/src/eculture/vocabularies/mappings/situations>
. . . file:///home/jan/src/eculture/vocabularies/mappings/eculture_situations.ttl
. <file:///home/jan/src/eculture/collections/aul/aul>
. . file:///home/jan/src/eculture/collections/aul/aul.rdfs
. . file:///home/jan/src/eculture/collections/aul/aul.rdf
. . file:///home/jan/src/eculture/collections/aul/aul9styles.rdf
. . file:///home/jan/src/eculture/collections/aul/extractedperiods.rdf
. . file:///home/jan/src/eculture/collections/aul/manual-periods.rdf

8.2.1 Referencing resources

Resources and manifests are located either on the local filesystem or on a network resource. The initial manifest can also be loaded from a file or a URL. This defines the initial base URL of the document. The base URL can be overruled using the Turtle @base directive. Other documents can be referenced relative to this base URL by exploiting Turtle's URI expansion rules. Turtle resources can be specified in three ways, as absolute URLs (e.g. <http://www.example.com/rdf/ontology.rdf>), as relative URL to the base (e.g. <../rdf/ontology.rdf>) or following a prefix (e.g. prefix:ontology).

The prefix notation is powerful as we can define multiple of them and define resources relative to them. Unfortunately, prefixes can only be defined as absolute URLs or URLs relative to the base URL. Notably, they cannot be defined relative to other prefixes. In addition, a prefix can only be followed by a Qname, which excludes . and /.

Easily relocatable manifests must define all resources relative to the base URL. Relocation is automatic if the manifest remains in the same hierarchy as the resources it references. If the manifest is copied elsewhere (i.e. for creating a local version) it can use @base to refer to the resource hierarchy. We can point to directories holding manifest files using @prefix declarations. There, we can reference Virtual resources using prefix:name. Here is an example, were we first give some line from the initial manifest followed by the definition of the virtual RDFS resource.

@base <http://gollem.science.uva.nl/e-culture/rdf/> .

@prefix base:           <base_ontologies/> .

        a lib:Ontology ;
        a lib:Virtual ;
        dc:title "E-Culture core vocabularies" ;
                base:rdfs ,
                base:owl ,
                base:dc ,
                base:vra ,
        a lib:Schema ;
        a lib:Virtual ;
        rdfs:comment "RDF Schema" ;
        lib:source rdfs: ;
        lib:schema <rdfs.rdfs> .

8.3 Putting it all together

In this section we provide skeleton code for filling the RDF database from a password protected HTTP repository. The first line loads the application. Next we include modules that enable us to manage the RDF library, RDF database caching and HTTP connections. Then we setup the HTTP authentication, enable caching of processed RDF files and load the initial manifest. Finally load_data/0 loads all our RDF data.

:- use_module(server).

:- use_module(library(http/http_open)).
:- use_module(library(semweb/rdf_library)).
:- use_module(library(semweb/rdf_cache)).

:- http_set_authorization('http://www.example.org/rdf',
                          basic(john, secret)).

:- rdf_set_cache_options([ global_directory('RDF-Cache'),

:- rdf_attach_library('http://www.example.org/rdf/Manifest.ttl').

%%      load_data
%       Load our RDF data

load_data :-

8.4 Example: A metadata file for W3C WordNet

The VoID metadata below allows for loading WordNet in the two predefined versions using one of

?- rdf_load_library('wn-basic', []).
?- rdf_load_library('wn-full', []).
@prefix    void: <http://rdfs.org/ns/void#> .
@prefix    vann: <http://purl.org/vocab/vann/> .
@prefix     lib: <http://www.swi-prolog.org/rdf/library/> .
@prefix     owl: <http://www.w3.org/2002/07/owl#> .
@prefix     rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
@prefix    rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
@prefix     xsd: <http://www.w3.org/2001/XMLSchema#> .
@prefix      dc: <http://purl.org/dc/terms/> .
@prefix   wn20s: <http://www.w3.org/2006/03/wn/wn20/schema/> .
@prefix   wn20i: <http://www.w3.org/2006/03/wn/wn20/instances/> .

[ vann:preferredNamespacePrefix "wn20i" ;
  vann:preferredNamespaceUri "http://www.w3.org/2006/03/wn/wn20/instances/"
] .

[ vann:preferredNamespacePrefix "wn20s" ;
  vann:preferredNamespaceUri "http://www.w3.org/2006/03/wn/wn20/schema/"
] .

        a void:Dataset ;
        dc:description "Common files between full and basic version" ;
        lib:source wn20i: ;
                <wordnet-attribute.rdf.gz> ,
                <wordnet-causes.rdf.gz> ,
                <wordnet-classifiedby.rdf.gz> ,
                <wordnet-entailment.rdf.gz> ,
                <wordnet-glossary.rdf.gz> ,
                <wordnet-hyponym.rdf.gz> ,
                <wordnet-membermeronym.rdf.gz> ,
                <wordnet-partmeronym.rdf.gz> ,
                <wordnet-sameverbgroupas.rdf.gz> ,
                <wordnet-similarity.rdf.gz> ,
                <wordnet-synset.rdf.gz> ,
                <wordnet-substancemeronym.rdf.gz> ,
                <wordnet-senselabels.rdf.gz> .

        a void:Dataset ;
        void:subset <wnskosmap> ;
        void:dataDump <wnSkosInScheme.ttl.gz> .

        a lib:Schema ;
        lib:source wn20s: ;
                <wnskosmap.rdfs> .

        a void:Dataset ;
        lib:source wn20s: ;
                <wnbasic.rdfs> .

        a void:Dataset ;
        a lib:CloudNode ;
        dc:title "Basic WordNet" ;
        dc:description "Light version of W3C WordNet" ;
        owl:versionInfo "2.0" ;
        lib:source wn20i: ;
                <wnbasic-schema> ,
                <wn20-skos> ,
                <wn20-common> .

        a void:Dataset ;
        lib:source wn20s: ;
                <wnfull.rdfs> .

        a void:Dataset ;
        a lib:CloudNode ;
        dc:title "Full WordNet" ;
        dc:description "Full version of W3C WordNet" ;
        owl:versionInfo "2.0" ;
        lib:source wn20i: ;
                <wnfull-schema> ,
                <wn20-skos> ,
                <wn20-common> ;
                <wordnet-antonym.rdf.gz> ,
                <wordnet-derivationallyrelated.rdf.gz> ,
                <wordnet-participleof.rdf.gz> ,
                <wordnet-pertainsto.rdf.gz> ,
                <wordnet-seealso.rdf.gz> ,
                <wordnet-wordsensesandwords.rdf.gz> ,
                <wordnet-frame.rdf.gz> .

9 library(semweb/sparql_client): SPARQL client library

This module provides a SPARQL client. For example:

?- sparql_query('select * where { ?x rdfs:label "Amsterdam" }', Row,
                [ host('dbpedia.org'), path('/sparql/')]).

Row = row('http://www.ontologyportal.org/WordNet#WN30-108949737') ;

Or, querying a local server using an ASK query:

?- sparql_query('ask { owl:Class rdfs:label "Class" }', Row,
                [ host('localhost'), port(3020), path('/sparql/')]).
Row = true.
[nondet]sparql_query(+Query, -Result, +Options)
Execute a SPARQL query on an HTTP SPARQL endpoint. Query is an atom that denotes the query. Result is unified to a term rdf(S,P,O) for CONSTRUCT and DESCRIBE queries, row(...) for SELECT queries and true or false for ASK queries. Options are
The above three options set the location of the server.
Provide additional query parameters, such as the graph.
Unifies ListOfNames with a list of atoms that describe the names of the variables in a SELECT query.

Remaining options are passed to http_open/3. The defaults for Host, Port and Path can be set using sparql_set_server/1. The initial default for port is 80 and path is /sparql/.

For example, the ClioPatria server understands the parameter entailment. The code below queries for all triples using _rdfs_entailment.

?- sparql_query('select * where { ?s ?p ?o }',
                [ search([entailment=rdfs])
Set sparql server default options. Provided defaults are: host, port and repository. For example:
    sparql_set_server([ host(localhost),

The default for port is 80 and path is /sparql/.

sparql_read_xml_result(+Input, -Result)
Specs from http://www.w3.org/TR/rdf-sparql-XMLres/. The returned Result term is of the format:
select(VarNames, Rows)
Where VarNames is a term v(Name, ...) and Rows is a list of row(....) containing the column values in the same order as the variable names.
Where Bool is either true or false

10 library(semweb/rdf_compare): Compare RDF graphs

This library provides predicates that compare RDF graphs. The current version only provides one predicate: rdf_equal_graphs/3 verifies that two graphs are identical after proper labeling of the blank nodes.

Future versions of this library may contain more advanced operations, such as diffing two graphs.

[semidet]rdf_equal_graphs(+GraphA, +GraphB, -Substition)
True if GraphA and GraphB are the same under Substition. Substition is a list of BNodeA = BNodeB, where BNodeA is a blank node that appears in GraphA and BNodeB is a blank node that appears in GraphB.
GraphA is a list of rdf(S,P,O) terms
GraphB is a list of rdf(S,P,O) terms
Substition is a list if NodeA = NodeB terms.
To be done
The current implementation is rather naive. After dealing with the subgraphs that contain no bnodes, it performs a fully non-deterministic substitution.

11 library(semweb/rdf_portray): Portray RDF resources

To be done
- Define alternate predicate to use for providing a comment
- Use rdf:type if there is no meaningful label?
- Smarter guess whether or not the local identifier might be meaningful to the user without a comment. I.e. does it look `word-like'?

This module defines rules for user:portray/1 to help tracing and debugging RDF resources by printing them in a more concise representation and optionally adding comment from the label field to help the user interpreting the URL. The main predicates are:

Set the style used to portray resources. Style is one of:
ns() : id()
Write as NS:ID, compatible with what can be handed to the rdf predicates. This is the default.
Use quoted write of the full resource.
ns() : label()
Write namespace followed by the label. This format cannot be handed to rdf/3 and friends, but can be useful if resource-names are meaningless identifiers.
ns():id() = label()
This combines ns:id with ns:label, providing both human readable output and output that can be pasted into the commandline.
If Lang is a list, set the list or preferred languages. If it is a single atom, push this language as the most preferred language.

12 library(semweb/rdf_edit): Keep track of edits (deprecated)

It is anticipated that this library will eventually be superseeded by facilities running on top of the native rdf_transaction/2 and rdf_monitor/2 facilities. See section 3.6.

The module rdf_edit.pl is a layer than encasulates the modification predicates from section 3.3 for use from a (graphical) editor of the triple store. It adds the following features:

12.1 Transaction management

Transactions group low-level modification actions together.

Run Goal, recording all modifications to the triple store made through section 12.3. Execution is performed as in once/1. If Goal succeeds the changes are committed. If Goal fails or throws an exception the changes are reverted.

Transactions may be nested. A failing nested transaction only reverts the actions performed inside the nested transaction. If the outer transaction succeeds it is committed normally. Contrary, if the outer transaction fails, comitted nested transactions are reverted as well. If any of the modifications inside the transaction modifies a protected file (see rdfe_set_file_property/2) the transaction is reverted and rdfe_transaction/1 throws a permission error.

A successful outer transaction (`level-0') may be undone using rdfe_undo/0.

rdfe_transaction(:Goal, +Name)
As rdfe_transaction/1, naming the transaction Name. Transaction naming is intended for the GUI to give the user an idea of the next undo action. See also rdfe_set_transaction_name/1 and rdfe_transaction_name/2.
Set the `name' of the current transaction to Name.
rdfe_transaction_name(?TID, ?Name)
Query assigned transaction names.
rdfe_transaction_member(+TID, -Action)
Enumerate the actions that took place inside a transaction. This can be used by a GUI to optimise the MVC (Model-View-Controller) feedback loop. Action is one of:
assert(Subject, Predicate, Object)
retract(Subject, Predicate, Object)
update(Subject, Predicate, Object, Action)

12.2 File management

Enumerate/test whether File is modified sinds it was loaded or sinds the last call to rdfe_clear_modified/1. Whether or not a file is modified is determined by the MD5 checksum of all triples belonging to the file.
Set the unmodified-MD5 to the current MD5 checksum. See also rdfe_is_modified/1.
rdfe_set_file_property(+File, +Property)
Control access right and default destination of new triples. Property is one of
Where access is one of ro or rw. Access ro is default when a file is loaded for which the user has no write access. If a transaction (see rdfe_transaction/1) modifies a file with access ro the transaction is reversed.
Set this file to be the default destination of triples. If Default is fallback it is only the default for triples that have no clear default destination. If it is all all new triples are added to this file.
rdfe_get_file_property(?File, ?Property)
Query properties set with rdfe_set_file_property/2.

12.3 Encapsulated predicates

The following predicates encapsulate predicates from the rdf_db module that modify the triple store. These predicates can only be called when inside a transaction. See rdfe_transaction/1.

rdfe_assert(+Subject, +Predicate, +Object)
Encapsulates rdf_assert/3.
rdfe_retractall(?Subject, ?Predicate, ?Object)
Encapsulates rdf_retractall/3.
rdfe_update(+Subject, +Predicate, +Object, +Action)
Encapsulates rdf_update/4.
Encapsulates rdf_load/1.
Encapsulates rdf_unload/1.

12.4 High-level modification predicates

This section describes a (yet very incomplete) set of more high-level operations one would like to be able to perform. Eventually this set may include operations based on RDFS and OWL.

Delete all traces of resource. This implies all triples where Resource appears as subject, predicate or object. This predicate starts a transation.

12.5 Undo

Undo aims at user-level undo operations from a (graphical) editor.

Revert the last outermost (`level 0') transaction (see rdfe_transaction/1). Successive calls go further back in history. Fails if there is no more undo information.
Revert the last rdfe_undo/0. Successive calls revert more rdfe_undo/0 operations. Fails if there is no more redo information.
Test if there is another transaction that can be reverted. Used for activating menus in a graphical environment. TID is unified to the transaction id of the action that will be reverted.
Test if there is another undo that can be reverted. Used for activating menus in a graphical environment. TID is unified to the transaction id of the action that will be reverted.

12.6 Journalling

Optionally, every action through this module is immediately send to a journal-file. The journal provides a full log of all actions with a time-stamp that may be used for inspection of behaviour, version management, crash-recovery or an alternative to regular save operations.

rdfe_open_journal(+File, +Mode)
Open a existing or new journal. If Mode equala append and File exists, the journal is first replayed. See rdfe_replay_journal/1. If Mode is write the journal is truncated if it exists.
Close the currently open journal.
Test whether there is a journal and to which file the actions are journalled.
Read a journal, replaying all actions in it. To do so, the system reads the journal a transaction at a time. If the transaction is closed with a commit it executes the actions inside the journal. If it is closed with a rollback or not closed at all due to a crash the actions inside the journal are discarded. Using this predicate only makes sense to inspect the state at the end of a journal without modifying the journal. Normally a journal is replayed using the append mode of rdfe_open_journal/2.

12.7 Broadcasting change events

To realise a modular graphical interface for editing the triple store, the system must use some sort of event mechanism. This is implemented by the XPCE library library(broadcast) which is described in the XPCE User Guide. In this section we describe the terms brodcasted by the library.

A `level-0' transaction has been committed. The system passes the identifier of the transaction in Id. In the current implementation there is no way to find out what happened inside the transaction. This is likely to change in time.

If a transaction is reverted due to failure or exception no event is broadcasted. The initiating GUI element is supposed to handle this possibility itself and other components are not affected as the triple store is not changed.

rdf_undo(+Type, +Id)
This event is broadcasted after an rdfe_undo/0 or rdfe_redo/0. Type is one of undo or redo and Id identifies the transaction as above.

13 Related packages and issues

The SWI-Prolog SemWeb package is designed to provide access to the Semantic Web languages from Prolog. It consists of the low level rdf_db.pl store with layers such as library(semweb/rdfs.pl) to provide more high level querying of a triple set with relations such as rdfs_individual_of/2, rdfs_subclass_of/2, etc. SeRQL is a semantic web query language taking another route. Instead of providing alternative relations SeRQL defines a graph query on de deductive closure of the triple set. For example, under assumption of RDFS entailment rules this makes the query rdf(S, rdf:type, Class) equivalent to rdfs_individual_of(S, Class).

We developed a parser for SeRQL which compiles SeRQL path expressions into Prolog conjunctions of rdf(Subject, Predicate, Object) calls. Entailment modules realise a fully logical implementation of rdf/3 including the entailment reasoning required to deal with a Semantic Web language or application specific reasoning. The infra structure is completed with a query optimiser and an HTTP server compliant to the Sesame implementation of the SeRQL language. The Sesame Java client can be used to access Prolog servers from Java, while the Prolog client can be used to access the Sesame SeRQL server. For further details, see the project home.

14 OWL

The SWI-Prolog Semantic Web library provides no direct support for OWL. OWL(-2) support is provided through Thea, an OWL library for SWI-Prolog See http://www.semanticweb.gr/TheaOWLLib/.


This research was supported by the following projects: MIA and MultimediaN project (www.multimedian.nl) funded through the BSIK programme of the Dutch Government, the FP-6 project HOPS of the European Commision.

The implementation of AVL trees is based on libavl by Brad Appleton. See the source file avl.c for details.


compressed data
3.1 3.5.1
3.5 3.5
gz, format
12 12.6
Persistent store
3.4 3.4
3 3.1 3.1 3.1 3.1 3.2 3.3.1 3.3.2 3.3.2 3.5 7 13
3.1 4.5
3.3.1 3.3.2 3.4.3 12.3
8.1.2 8.2
12.1 12.2
12.1 12.1 12.2 12.3 12.5
12.1 12.5 12.5 12.7
4.1 4.1
4.4 4.4
rdf_graph_property/?Graph, ?Property
3.1 3.2 3.2 3.2 3.2 3.7
3 3.1 3.1 3.2
7.1 7.1
8.1 8.1
8.1.2 8.2
8.1.2 8.1.2
3.4 3.7 12.3
3.4 3.4.1 3.6 3.6 4 4.1
3.4.3 3.4.4 3.4.4
3.4 3.4.4 3.6 3.6 3.6
8.1 8.1 8.1.2 8.1.2
4.1 4.1
3.5.1 3.5.1 3.5.1 3.5.1
3.1.1 3.3.2 4.4 4.4.1 12
3.1 3.2 3.7
3.1 3.2 3.2 3.7
3.3.1 4.5.1 12.3
3.4 3.4.2
3.4 4.5 4.5.1
3.3.2 3.3.2 3.3.2
3.3.2 3.3.2 3.3.2 3.6 3.6 12
3.3.1 3.6 12.3
3.3.1 4.5.1 12.3
4.4 4.4
12 12.5