swish-0.10.0.3: A semantic web toolkit.

Copyright(c) 2003 Graham Klyne 2009 Vasili I Galchin
2011 2012 2014 2016 Douglas Burke
LicenseGPL V2
MaintainerDouglas Burke
Stabilityexperimental
PortabilityCPP, OverloadedStrings
Safe HaskellNone
LanguageHaskell98

Swish.RDF.Ruleset

Contents

Description

This module defines some datatypes and functions that are used to define rules and rulesets over RDF graphs.

For the routines that accept a graph in N3 format, the following namespaces are pre-defined for use by the graph: rdf: and rdfs:.

Synopsis

Data types for RDF Ruleset

type RDFFormula = Formula RDFGraph Source #

A named formula expressed as a RDF Graph.

type RDFRule = Rule RDFGraph Source #

A named inference rule expressed in RDF.

type RDFRuleMap = RuleMap RDFGraph Source #

A map for RDFRule rules.

type RDFClosure = GraphClosure RDFLabel Source #

A GraphClosure for RDF statements.

nullRDFFormula :: Formula RDFGraph Source #

The null RDF formula.

data GraphClosure lb Source #

Datatype for constructing a graph closure rule

Constructors

GraphClosure 

Fields

  • nameGraphRule :: ScopedName

    Name of rule for proof display

  • ruleAnt :: ArcSet lb

    Antecedent triples pattern (may include variable nodes)

  • ruleCon :: ArcSet lb

    Consequent triples pattern (may include variable nodes)

  • ruleModify :: VarBindingModify lb lb

    Structure that defines additional constraints and/or variable bindings based on other matched query variables. Matching the antecedents. Use varBindingId if no additional variable constraints or bindings are added beyond those arising from graph queries.

Instances
Label lb => Eq (GraphClosure lb) Source #

Equality is based on the closure rule, anrecedents and consequents.

Instance details

Defined in Swish.RDF.Ruleset

Methods

(==) :: GraphClosure lb -> GraphClosure lb -> Bool #

(/=) :: GraphClosure lb -> GraphClosure lb -> Bool #

Show (GraphClosure lb) Source # 
Instance details

Defined in Swish.RDF.Ruleset

makeGraphClosureRule :: GraphClosure RDFLabel -> Rule RDFGraph Source #

Define a value of type Rule based on an RDFClosure value.

makeRDFGraphFromN3Builder :: Builder -> RDFGraph Source #

Helper function to parse a string containing Notation3 and return the corresponding RDFGraph value.

makeRDFFormula Source #

Arguments

:: Namespace

namespace to which the formula is allocated

-> LName

local name for the formula in the namespace

-> Builder

graph in Notation 3 format

-> RDFFormula 

Create an RDF formula.

makeRDFClosureRule Source #

Arguments

:: ScopedName

scoped name for the new rule

-> [RDFGraph]

RDFGraphs that are the entecedent of the rule.

(Note: bnodes and variable names are assumed to be shared by all the entecedent graphs supplied. is this right?)

-> RDFGraph

the consequent graph

-> RDFVarBindingModify

is a variable binding modifier value that may impose additional conditions on the variable bindings that can be used for this inference rule, or which may cause new values to be allocated for unbound variables. These modifiers allow for certain inference patterns that are not captured by simple "closure rules", such as the allocation of bnodes corresponding to literals, and are an extension point for incorporating datatypes into an inference process.

If no additional constraints or variable bindings are to be applied, use value varBindingId

-> RDFRule 

Constructs an RDF graph closure rule. That is, a rule that given some set of antecedent statements returns new statements that may be added to the graph.

Create rules using Notation3 statements

makeN3ClosureRule Source #

Arguments

:: Namespace

namespace to which the rule is allocated

-> LName

local name for the rule in the namespace

-> Builder

the Notation3 representation of the antecedent graph. (Note: multiple antecedents can be handled by combining multiple graphs.)

-> Builder

the Notation3 representation of the consequent graph.

-> RDFVarBindingModify

a variable binding modifier value that may impose additional conditions on the variable bindings that can be used for this inference rule, or which may cause new values to be allocated for unbound variables. These modifiers allow for certain inference patterns that are not captured by simple closure rules, such as the allocation of bnodes corresponding to literals, and are an extension point for incorporating datatypes into an inference process.

If no additional constraints or variable bindings are to be applied, use a value of varBindingId, or use makeN3ClosureSimpleRule.

-> RDFRule 

Constructs an RDF graph closure rule. That is, a rule that given some set of antecedent statements returns new statements that may be added to the graph. This is the basis for implementation of most of the inference rules given in the RDF formal semantics document.

makeN3ClosureSimpleRule Source #

Arguments

:: Namespace

namespace to which the rule is allocated

-> LName

local name for the rule in the namepace

-> Builder

the Notation3 representation of the antecedent graph. (Note: multiple antecedents can be handled by combining multiple graphs.)

-> Builder

the Notation3 representation of the consequent graph.

-> RDFRule 

Construct a simple RDF graph closure rule without additional node allocations or variable binding constraints.

makeN3ClosureModifyRule Source #

Arguments

:: Namespace

namespace to which the rule is allocated

-> LName

local name for the rule in the given namespace

-> Builder

the Notation3 representation of the antecedent graph. (Note: multiple antecedents can be handled by combining multiple graphs.)

-> Builder

the Notation3 representation of the consequent graph.

-> RDFVarBindingModify

a variable binding modifier value that may impose additional conditions on the variable bindings that can be used for this inference rule (vflt).

These modifiers allow for certain inference patterns that are not captured by simple closure rules, such as deductions that pertain only to certain kinds of nodes in a graph.

-> RDFVarBindingModify

a variable binding modifier that is applied to the variable bindings obtained, typically to create some additional variable bindings. This is applied before the preceeding filter rule (vflt).

-> RDFRule 

Constructs an RDF graph closure rule that incorporates a variable binding filter and a variable binding modifier.

makeN3ClosureAllocatorRule Source #

Arguments

:: Namespace

namespace to which the rule is allocated

-> LName

local name for the rule in the given namespace

-> Builder

the Notation3 representation of the antecedent graph. (Note: multiple antecedents can be handled by combining multiple graphs.)

-> Builder

the Notation3 representation of the consequent graph.

-> RDFVarBindingModify

variable binding modifier value that may impose additional conditions on the variable bindings that can be used for this inference rule (vflt).

-> ([RDFLabel] -> RDFVarBindingModify)

function applied to a list of nodes to yield a variable binding modifier value.

The supplied parameter is applied to a list of all of the variable nodes (including all blank nodes) in the antecedent graph, and then composed with the vflt value. This allows any node allocation function to avoid allocating any blank nodes that are already used in the antecedent graph. (See makeNodeAllocTo).

-> RDFRule 

Construct an RDF graph closure rule with a bnode allocator.

This function is rather like makeN3ClosureModifyRule, except that the variable binding modifier is a function from the variables in the variables and bnodes contained in the antecedent graph.

makeNodeAllocTo Source #

Arguments

:: RDFLabel

variable node to which a new blank node is bound

-> RDFLabel

variable which is bound in each query to a graph node to which new blank nodes are allocated.

-> [RDFLabel] 
-> RDFVarBindingModify 

This function defines a variable binding modifier that allocates a new blank node for each value bound to a query variable, and binds it to another variable in each query binding.

This provides a single binding for query variables that would otherwise be unbound by a query. For example, consider the inference pattern:

 ?a hasUncle ?c => ?a hasFather ?b . ?b hasBrother ?c .

For a given ?a and ?c, there is insufficient information here to instantiate a value for variable ?b. Using this function as part of a graph instance closure rule allows forward chaining to allocate a single bnode for each occurrence of ?a, so that given:

 Jimmy hasUncle Fred .
 Jimmy hasUncle Bob .

leads to exactly one bnode inference of:

 Jimmy hasFather _:f .

giving:

 Jimmy hasFather _:f .
 _:f   hasBrother Fred .
 _:f   hasBrother Bob .

rather than:

 Jimmy hasFather _:f1 .
 _:f1  hasBrother Fred .
 Jimmy hasFather _:f2 .
 _:f2  hasBrother Bob .

This form of constrained allocation of bnodes is also required for some of the inference patterns described by the RDF formal semantics, particularly those where bnodes are substituted for URIs or literals.

Debugging

graphClosureFwdApply :: GraphClosure RDFLabel -> [RDFGraph] -> [RDFGraph] Source #

Forward chaining function based on RDF graph closure description

Note: antecedents here are presumed to share bnodes.

graphClosureBwdApply :: GraphClosure RDFLabel -> RDFGraph -> [[RDFGraph]] Source #

Backward chaining function based on RDF graph closure description