The Shape Expressions (ShEx) language describes RDF nodes and graph structures.
A node constraint describes an RDF node (IRI, blank node or literal) and a shape describes the triples involving nodes in an RDF graph.
These descriptions identify predicates and their associated cardinalities and datatypes.
ShEx shapes can be used to communicate data structures associated with some process or interface, generate or validate data, or drive user interfaces.
This document defines the ShEx language.
See the Shape Expressions Primer for a non-normative description of ShEx.
This is an editor's draft of the Shape Expressions specification.
ShEx 2.x differs significantly from the W3C ShEx Submission.
The July 2017 publication included a definition of validation which implied infinite recursion.
This version explicitly includes recursion checks.
No tests changed as a result of this and no implementations or applications are known to have been affected.
If you wish to make comments regarding this document, please raise them as GitHub issues.
There are separate interfaces for specification, langauge and test issues.
Only send comments to public-shex@w3.org (subscribe,
archives) if you are unable to raise issues on GitHub. All comments are welcome.
1. Introduction
The Shape Expressions (ShEx) language provides a structural schema for RDF data.
This can be used to document APIs or datasets, aid in development of API-conformant messages, minimize defensive programming, guide user interfaces, or anything else that involves a machine-readable description of data organization and typing requirements.
ShEx describes RDF graph [RDF11-CONCEPTS] structures as sets of potentially connected Shapes.
These constrain the triples involving nodes in an RDF graph.
Node Constraints constrain RDF nodes by constraining their node kind (IRI, blank node or Literal), enumerating permissible values in value sets, specifying their datatype, and constraining value ranges of Literals.
Additionally, they constrain lexical forms of Literals, IRIs and labeled blank nodes.
Shape Expressions schemas share blank nodes with the constrained RDF graphs in the same way that graphs in RDF datasets [rdf11-concepts] share blank nodes.
ShEx can be represented in JSON structures (ShExJ) or a compact syntax (ShExC).
The compact syntax is intended for human consumption; the JSON structure for machine processing.
This document defines ShEx in terms of ShExJ and includes a section on the ShEx Compact Syntax (ShEx).
2. Notation
The JSON [rfc7159] Syntax serves as a serializable proxy for an abstract syntax.
literals with the datatype http://www.w3.org/2001/XMLSchema#string are represented with the value property, e.g. { "value": "abc" }.
language-tagged strings are represented with an additional language property, e.g. { "value": "hello world", "langague": "en-US" }
datatyped literals are represented with an additional datatype property, e.g. { "value": "123", "datatype": "http://www.w3.org/2001/XMLSchema#integer" }
2.1 JSON Grammar
This specification uses a JSON grammar to describe the set of JSON documents that can be interpreted as a ShEx schema.
ShEx data structures are represented as JSON objects with a member with the name "type" (i.e. an object with a type attribute):
{ "type": "typeName", member0…n }
These are expressed in JSON grammar as typeName { member* }.
RFC7159 Section 2 provides syntactic constraints for JSON — the grammar constraining those to valid ShExJ constructs is composed of:
typeName is the name of the typed data structure.
Types are referenced in the definitions of object members and in the definitions of the semantics for those data structures.
member* is a list of zero or more terminals or references to other typeExpressions.
A typeExpression is one of:
typeName — an object of corresponding type
array: [ typeExpression+ ]— an array of one or more JSON values matching the typeExpression.
choice: typeExpression1 | typeExpression2 | …— a choice between two or more typeExressions.
Cardinalities are represented as by the strings ?, +, * following the notation in the XML specification[XML] or {m,} to indicate a that at least m elements are required.
The following examples are excerpts from the definitions below.
In the JSON notation,
signifies that a NodeConstraint has a nodeKind of one of the four literals followed by any number of xsFacet and an xsFacet is either a stringFacet or a numericFacet.
2.2 References
ShExJ is a dialect of JSON-LD [JSON-LD] and the member id is used as a node identifier.
An object may be represented inline or referenced by its id which may be either a blank node or an IRI.
Not captured in this JSON syntax definition is the rule that every shapeExpr nested in a schema's shapes must have an id and no other shapeExpr may have an id.
The JSON syntax definitition simplifies this by adding id:shapeExprLabel? to every shapeExpr.
This example includes a nested shape. Nested shapes are not permitted to have ids.
The shape expressions in a schema are expressed as a list of shape expressions with id attributes because JSON-LD 1.0 has no provision for defining as JSON object as representing a map of URL→object.
Drafts of JSON-LD 1.1 include this expressivity and, once JSON-LD 1.1 exits Candidate Recommendation, future versions of ShExJ will likely adopt it while requiring backard-compatibility with the list of shape expressions with id attributes.
2.3 Document style
JSON examples are rendered in a .json CSS style.
Partial examples include ranges in a .comment CSS style to indicate text which would be substituted in a complete example.
For example { "type": "ShapeAnd", "shapeExprs": [ SE1, … ] } indicates that both SE1 and … would be substituted in a complete example.
In javascript-enabled browsers, schemas with a button can be converted between the JSON representation and the compact syntax by clicking the button.
The button text indicates the currently shown representation.
Selecting the example and pressing "j" or "c" converts the example to the JSON (ShExJ) or compact form (ShExC).
Pressing "shift J" or "shift C" converts all such examples to ShExJ or ShExC.
2.4 Graph access
The validation process defined in this document relies on matching triple patterns in the form (subject, predicate, object) where each position may be supplied by a constant, a previously defined term, or the underscore "_", which represents a previously undefined element or wildcard.
This corresponds to a SPARQL Triple Pattern where each "_" is replaced by a unique blank node.
Matching such a triple pattern against a graph is defined by SPARQL Basic Graph Pattern Matching (BGP) with a BGP containing only that triple pattern.
2.5 Validation process
ShEx validation is defined in this document by the isValid function.
This process takes as input a shapes schema, an RDF graph, and a fixed ShapeMap (abbreviated as "ShapeMap" in this document).
ShEx validation results may be reported as a result ShapeMap [shape-map].
For illustration purposes in this specification, both the fixed ShapeMap input and the result ShapeMap output are represented in a table with four columns: node: a ShapeMap nodeSelector, shape: a ShapeMap shapeLabel, result: "pass" (conformant) or "fail" (nonconformant), and an optional reason: an informal, human-readable explanation.
node
shape
result
reason
<node1>
<Shape1>
pass
<node2>
<Shape1>
fail
no ex:state supplied.
3. Terminology
Shape expressions are defined using terms from RDF semantics [rdf11-mt]:
As well as sections marked as non-normative, all authoring guidelines,
diagrams, examples, and notes in this specification are non-normative.
Everything else in this specification is normative.
The key words MAY, MUST, MUST NOT, and SHOULD NOT are to be interpreted as described in [RFC2119].
Conformance criteria are relevant to authors and authoring tool implementers. As well
as sections marked as non-normative, all authoring guidelines, diagrams, examples,
and notes in this specification are non-normative. Everything else in this
specification is normative.
isValid: For an input graph G, schema Sch and input shape map ism, isValid(G, Sch, ism) indicates that for every RDFnode/shapeExprLabel pair (n, sl) in ism, the node n satisfies the shape expression identified by sl.
The latter is captured by the expression satisfies(n, s, G, Sch, completeTyping(G, Sch)).
The function satisfies is defined for every kind of shape expression.
The validation of an RDF graph G against a ShEx schema Sch is based on the existence of completeTyping(G, Sch).
For an RDF graph G and a shapes schema Sch, a typing is a set of pairs of the form (n, s) where n is a node in G and s is a Shape that appears in some shape expression in the shapes mapping of Sch.
A correct typing is a typing such that for every RDFnode/shape pair (n,s) in typing, matchesShape(n, s, G, Sch, typing) holds.
completeTyping(G, Sch) is a unique correct typing that exists for every graph and every ShEx schema that satisfies the schema requirements.
completeTyping: the definition of completeTyping(G, Sch) is based on a stratification of Sch.
The number of strata of Sch is the number of maximal strongly connected components of the dependency graph of Sch.
A stratification of a schema Sch with k strata is a function stratum that associates with every Shape in Sch a natural number between 1 and k such that:
If s1 and s2 belong to the same maximal connected component, then stratum(s1) = stratum(s2).
If there is a negated reference from s1 to s2, then stratum(s2) < stratum(s1).
The existence of a stratification for every schema is guaranteed by the negation requirement.
Given a stratificationstratum of Sch with k strata, define inductively the series of k typings completeTypingOn(1, G, Sch) … completeTypingOn(k, G, Sch).
completeTypingOn(1, G, Sch) is the union of all correct typings that contain only RDFnode/shape pairs (n,s) with stratum(s) = 1;
for every i between 2 and k, completeTypingOn(i, G, Sch) is the union of all correct typings that include completeTypingOn(i-1, G, Sch) and that contain only RDFnode/shape pairs (n,s) with stratum(s) ≤ i;
Then completeTyping(G, Sch) = completeTypingOn(k, G, Sch).
Note
The schema Sch might have several different stratifications but completeTyping(G, Sch) is the same for all these stratifications.
Note
It is not necessary to compute exhaustiveShapeMap in order to decide isValid(Sch, G, m).
Validity can be checked by computing only a portion of exhaustiveShapeMap by an appropriate algorithm presented in a separate document.
A shape expression is composed of four kinds of objects combined with the algebraic operators And, Or and Not:
a node constraint (NodeConstraint) that defines the set of allowed values of a node.
These include specification of RDF node kind, literal datatype, XML String and numeric facets and enumeration of value sets.
a shape constraint (Shape) that defines a constraint on the allowed neighbourhood of a node, that is, the allowed triples that contain this node as subject or object.
an external shape (ShapeExternal) which is an extension mechanism to externally define e.g. functional shapes or prohibitively large value sets.
IRI AND
(
@<http://schema.example/#IssueShape>
OR NOT { … }
)
In this ShapeOr's shapeExprs, "http://schema.example/#IssueShape" is a reference to the shape expression with the id "http://schema.example/#IssueShape".
5.3.2 Semantics
satisfies: The expression satisfies(n, se, G, Sch, t) indicates that a node n and a graphG satisfy a shape expressionse with typingt for schema Sch. notSatisfies: Conversely, notSatisfies(n, se, G, Sch, t) indicates that n and G do not satisfy se with the given typingt.
satisfies(n, se, G,Sch, t) is true if and only if:
For a node n and constraint nc, satisfies2(n, nc) if and only if for every nodeKind, datatype, xsFacet and values constraint value v present in ncnodeSatisfies(n, v).
The following sections define nodeSatisfies for each of these types of constraints:
Note that <issue2> fails not because of a nodeKind violation but instead because of a Cardinality violation described below.
5.4.3 Datatype Constraints
For a node n and constraint value v, nodeSatisfies(n, v) if n is an Literal with the datatype v and, if v is in the set of SPARQL operand data types[sparql11-query], an XML schema string with a value of the lexical form of n can be cast to the target type v per XPath Functions 3.1 section 19 Casting[xpath-functions].
The lexical form and numeric value (where applicable) of all datatypes required by SPARQL XPath Constructor FunctionsMUST be tested for conformance with the corresponding XML Schema form.
ShEx extensions MAY add support for other datatypes.
RDF 1.0 included RDF literals with no datatype or language tag.
These are called "simple literals" in SPARQL11[sparql11-query].
In RDF 1.1, these literals have the datatype http://www.w3.org/2001/XMLSchema#string.
Let len = the number of unicode codepoints in lex
For a node n and constraint value v, nodeSatisfies(n, v):
for "length" constraints, v = len,
for "minlength" constraints, v >= len,
for "maxlength" constraints, v <= len,
for "pattern" constraints, v is unescaped into a valid XPath 3.1 regular expression[xpath-functions-31] re and invoking fn:matches(lex, re) returns fn:true.
If the flags parameter is present, it is passed as a third argument to fn:matches.
The pattern may have XPath 3.1 regular expression escape sequences per the modified production [10] in section 5.6.1.1 as well as numeric escape sequences of the form 'u' HEX HEX HEX HEX or 'U' HEX HEX HEX HEX HEX HEX HEX HEX.
Unescaping replaces numeric escape sequences with the corresponding unicode codepoint.
ex:submittedBy expected to be >= 10 characters, 3 characters found.
Note
Access to blank node identifiers may be impossible or unadvisable for many use cases.
For instance, the SPARQL Query and SPARQL Update languages treat blank nodes in the query, labeled or otherwise, as variables.
Lexical constraints on blank node identifiers can only be implemented in systems which preserve such labels on data import.
found "\U0001D4B8" instead of "𝒸" (codepoint U+1D4B8).
5.4.5 XML Schema Numeric Facet Constraints
Numeric facet constraints apply to the numeric value of RDF Literals with datatypes listed in SPARQL 1.1 Operand Data Types[sparql11-query].
Numeric constraints on non-numeric values fail.
totaldigits and fractiondigits constraints on values not derived from xsd:decimal fail.
Let num be the numeric value of n.
For a node n and constraint value v, nodeSatisfies(n, v):
for "fractiondigits" constraints, v is less than or equals the number of digits to the right of the decimal place in the XML Schema canonical form[xmlschema-2] of the value of n, ignoring trailing zeros.
vsv is a Wildcard with exclusionsexcls and there is no x in excls such that nodeIn(n, excl).
nodeIn: asserts that an RDF node n is equal to an RDF term s or is in a set defined by a IriStem, LiteralStem or LanguageStem.
The expression nodeIn(n, s) is satisfied if:
A value set can have a single value in it.
This is used to indicate that a specific value is required, e.g. that an ex:state must be equal to <http://schema.example/#Resolved> or the rdf:type of some node must be foaf:Person.
5.5 Shapes and Triple Expressions
Triple expressions are used for defining patterns composed of triple constraints.
Shapes associate triple expressions with flags indicating whether triples match if they do not correspond to triple constraints in the triple expression.
A triple expression is composed of TripleConstraint and tripleExprRef objects composed with grouping and choice operators.
The semantics of the matchesShape function are based on the matches function defined below.
For a noden, shapeS, graphG, a ShExSchema Sch, and typingm, matchesShape(n, S, G, Sch, m) if and only if:
neigh(G, n) can be partitioned into two sets matched and remainder such that matches(matched, expression, m).
If expression is absent, remainder = neigh(G, n).
Let outs be the arcsOut in remainder:
outs = remainder ∩ arcsOut(G, n).
Let matchables be the triples in outs whose predicate appears in a TripleConstraint in expression.
If expression is absent, matchables = Ø (the empty set).
The complexity of partitioning is described briefly in the ShEx2 Primer.
There is no triple in matchables which matches a TripleConstraint in expression.
Let unmatchables be the triples in outs which are not in matchables.
matchables ∪ unmatchables = outs.
There is no triple in matchables whose predicate does not appear in extra.
closed is false or unmatchables is empty.
matches: asserts that a triple expression is matched by a set of triples that come from the neighbourhood of a node in an RDF graph.
The expression matches(T, expr, m) indicates that a set of triples T can satisfy these rules:
expr has semActs and matches(T, expr, m) by the remaining rules in this list and the evaluation of semActs succeeds according to the section below on Semantic Actions.
expr has a cardinality of min and/or max not equal to 1, where a max of -1 is treated as unbounded, and T can be partitioned into k subsets T1, T2,…Tk such that min ≤ k ≤ max and for each Tn, matches(Tn, expr, m) by the remaining rules in this list.
Both the shape expression <PersonShape> and the triple expression <nameExpr> are in scope. schema2's <nameExpr> is referenced in schema1's <EmployeeShape>
Redundant imports are treated as a single import.
This includes circular imports:
It is an error if there are any conflicting labels for shape expressions or triple expressions, or if any imported schema has either a startActs or a start member.
<http://schema.example/#schema1#S1> has conflicting definitions and
an included schema has a start directive and
the reference to <http://schema.example/#schema1#S3> is not resolvable after imports.
5.7 Schema Requirements
The semantics defined above assume three structural requirements beyond those imposed by the grammar of the abstract syntax.
These ensure referential integrity and eliminate logical paradoxes such as those that arrise through the use of negation.
These are not constraints expressed by the schema but instead those imposed on the schema.
satisfies(n, sei, G, Sch, m)
for each shapesei in sen.
5.7.2 Shape Expression Reference Requirement
A shapeExprRefMUST appear in the schema's shapes map (or an imported schema's map) and the corresponding shape expressionMUST be a Shape with a shapeExpr.
The function shapeExprWithId(shapeExprRef) returns the shape expression with an id of shapeExprRef.
Additionally, a shapeExprLabel cannot refer to itself through a shape reference either directly or recursively.
The shapeExprRef closure of a shape expressionse is the set of shape expression labels used as references in se.
The shapeExprLabelsl belongs to shapeExprRefClosure(se) if and only if:
Additionally, a tripleExprLabel cannot refer to itself through a triple expression reference either directly or recursively.
The tripleExprRef closure of a triple expressionte is the set of triple expression labels used as references in te.
The tripleExprLabeltl belongs to tripleExprRefClosure(te) if and only if:
A schema MUST NOT contain any Shape that has a negated reference to itself, either directly or transitively.
This is formalized by the requirement that the dependency graph of a schema MUST NOT have a cycle that traverses some negated reference.
The set of atomic shapes of a shapeExprse contains a Shapes if s or its id appears either directly or by shapeExprRef in se.
That is, s belongs to atomicShapes(se) if and only if
s appears as an atomic shape in se, or
sid is the id of s and sid appears as an atomic shapeExprRef in se, or
s belongs to atomicShapes(se2) for some shape expression se2 such that the id of se2 belongs to the shapeExprRefClosure of se.
The set of atomic triple constraints of a tripleExprte includes every TripleConstrainttc that appears directly or by tripleExprRef in te.
That is, tc belongs to atomicTripleConstraints(te) if and only if:
there is a triple constraint tc that belongs to atomicTripleConstraints(s1.expression), and
tc.valueExpr is present, and
s2 belongs to atomicShapes(tc.valueExpr).
The reference from s1 to s2 is a negated reference if
s2 appears under an odd number of ShapeNot in the shapeExprRef closure of tc.valueExpr, or
tc has predicatep and s1 has extrap.
The dependency graph of the schema Sch is the graph which vertices are all the Shapes that appear in some shape expression in the shapes of Sch, and that has two kinds of edges: negative and positive.
There is a negative edge from s1 to s2 if s1 has a negated reference to s2.
There is a positive edge from s1 to s2 if s1 has a reference but not a negated reference to s2.
Semantic actions serve as an extension point for Shape Expressions.
They appear in lists in Schema's startActs and Shape, OneOf, EachOf and TripleConstraint's semActs.
A semantic action is a tuple of an identifier and some optional code:
The evaluation semActsSatisfied on a list of SemActs returns success or failure.
The evaluation of an individual SemAct is implementation-dependent.
5.8.2 Use - informative
A practical evaluation of a SemAct will provide access to some context.
For instance, the http://shex.io/extensions/Test/ extension requires access to the subject, predicate and object of a triple matching a TripleConstraint.
These are used in a print function.
Annotations provide a format-independent way to provide additional information about elements in a schema.
They appear in lists in Shape, OneOf, EachOf and TripleConstraint's annotations.
Annotations do not affect whether a node conforms to some shape.
Because they are part of the structure of the schema, they can be parsed in one ShEx format and emitted in that format or another.
Here the shape identified by http://schema.example/#IntConstraint is a shape expression consisting of a single NodeConstraint.
Per Shape Expression Semantics, "30"^^<http://www.w3.org/2001/XMLSchema#integer> satisfies IntConstraint.
This document uses this nested tree convention to indicate that the dependency of an evaluation on those nested inside it.
Nesting is expressed as indentation.
Here, the evaluation of satisfies NodeConstraint ("30"^^xsd:integer, S1, G, m) depends on satisfies2 NodeConstraint ("30"^^xsd:integer, S1).
Validate"30"^^<http://www.w3.org/2001/XMLSchema#integer> as IntConstraint:
Validating a shape requires evaluating it's triple expression as well as the variables and functions neigh(G, n), matched, remainder, outs, matchables and unmatchables:
matchables = Ø /* There are no remaining arcs out of <Alice> with predicates appearing in tc1. */
unmatchables = Ø /* There are no other arcs out of <Alice>. */
closed is false /* The Shape's closed paramater has a value of false. */
It is quite common that Shapes will constrain their nested TripleConstraints with NodeConstraints.
Here is an example including that, extra triples and a closed shape:
matchables = Ø, unmatchables = [t5], closed is false
Any mixure of foaf:name with foaf:givenName or foaf:familyName will fail to satisfy the schema as there will be a matchable triple t3 that's not used in the triple expressionte1.
Setting the maximum cardinality of a TripleConstraint with predicatep to zero (i.e. "max": 0 in ShExJ or {0} or {0, 0} in ShExC) asserts that matching nodes must have no triples with predicatep.
The ShEx Compact Syntax expresses ShEx schemas in a compact, human-friendly form.
Parsing ShExC transforms a ShExC document into an equivalent ShExJ structure.
This is defined as a BNF which accepts ShExC followed by instructions for tranlating the rules in the BNF production into their corresponding ShExJ objects.
For example, "shapeExprDecl returns shapeExpression" indicates that the result of matching the shapeExprDecl production is the object produced by parsing the shapeExpression production.
Shape expressions are logical combinations of shape atoms.
Inline variants of shape expressions are used in tripleConstraints and are not permitted to have annotations or semantic actions.
If the right shapeAnd matches one or more times, the result is a ShapeOr object with shapeExprs containing the first shapeAnd followed by the ordered list from the second shapeAnd:
If the right shapeNot matches one or more times, the result is a ShapeAnd object with shapeExprs containing the first shapeNot followed by the ordered list from the second shapeNot:
Shape defintions associate a triple expression with a closed flag and a list of partially constrained (extra) predicates.
Any predicate appearing in a triple expression is fully constrained unless it appears in the list of extras.
If the right groupTripleExpr matches one or more times, the result is a OneOf object with expressions containing the first groupTripleExpr followed by the ordered list from the second groupTripleExpr:
If "." matches and exclusion matches one or more times, all matched items must be consistently iri, literal, or language. valueSetValue returns either a IriStemRange, LiteralStemRange, or LanguageStemRange object with exclusions equal to the set of results of exclusion:
If iri and "~" match and iriExclusion matches one or more times, iriRange returns a IriStemRange object with exclusions equal to the set of results of iriExclusion:
The literal has a lexical form of the first rule argument, String. If the '^^' iri rule matched, the datatype is iri and the literal has no language tag. If the langString rule matched, the datatype is rdf:langString and the language tag is extracted from langTag. If neither matched, the datatype is xsd:string and the literal has no language tag.
The characters between "<" and ">" are taken, with the numeric escape sequences unescaped, to form the unicode string of the IRI. Relative IRI resolution is performed per Turtle Section 6.3.
When used in a prefixDecl production, the prefix is a potentially empty unicode string matching the first argument of the rule and serves as a key into the prefixes map.
A potentially empty prefix is identified by the first token, PNAME_NS. The prefixes mapMUST have a corresponding namespace. The unicode string of the IRI is formed by unescaping the reserved characters [rfc7159] in the second argument, PN_LOCAL, and concatenating this onto the namespace.
A potentially empty prefix is identified by the second token, PNAME_NS. The prefixes mapMUST have a corresponding namespace. The unicode string of the IRI is formed by unescaping the reserved characters [rfc7159] in the third token, PN_LOCAL, and concatenating this onto the namespace.
pattern is a unicode string formed from the characters between the outermost '/'s by unescaping matches of '\\' '/' in the terminal pattern as well as the numeric escape sequences matched by UCHAR.
The remaining escape sequences are included verbatim in pattern, e.g.
^\/\t\\\U0001D4B8$
would become
^/\t\\\U0001D4B8$
. flags is a sequence of the characters [smix] if any were matched. Otherwise no flags attribute is returned.
The characters between the outermost "'''"s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form.
The characters between the outermost '"""'s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form.
The characters between the outermost "'"s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form. The trailing LANGTAG is used to create a language-tagged string.
The characters between the outermost '"'s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form. The trailing LANGTAG is used to create a language-tagged string.
The characters between the outermost "'''"s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form. The trailing LANGTAG is used to create a language-tagged string.
The characters between the outermost '"""'s are taken, with numeric and string escape sequences unescaped, to form the unicode string of a lexical form. The trailing LANGTAG is used to create a language-tagged string.
This section aggregates the JSON grammar rules defined above and includes terminals referenced above.
A ShExJ document is a JSON-LD [JSON-LD] document which uses a proscribed structure to define a schema containing shape expressions and triple expressions.
A ShExJ document MAY include an @context property referencing http://www.w3.org/ns/shex.jsonld.
In the absense of a top-level @context, ShEx Processors MUST act as if a @context property is present with the value http://www.w3.org/ns/shex.jsonld.
A ShExJ document can also be thought of as the serialization of an RDF Graph using the Shape Expression Vocabulary [shex-vocab] which conforms to the shape defined in B.ShEx Shape.
Processors MAY interpret a ShExJ document as an RDF Graph.
Processors may also transform arbitrary RDF Graphs conforming to B.ShEx Shape into ShExJ using a mechanism not described within this specification.
In ShExJ, the unbounded cardinality constraint is -1, rather than "*".
PREFIX sx: <http://www.w3.org/ns/shex#>
PREFIX xsd: <http://www.w3.org/2001/XMLSchema#>
PREFIX rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
BASE <http://www.w3.org/ns/shex#>
start=@<Schema><Schema> CLOSED {
a [sx:Schema] ;
sx:startActs @<SemActList1Plus>? ;
sx:start @<shapeExpr>?;
sx:shapes @<shapeExpr>*
}
<shapeExpr> @<ShapeOr> OR @<ShapeAnd> OR @<ShapeNot> OR @<NodeConstraint> OR @<Shape> OR @<ShapeExternal><ShapeOr> CLOSED {
a [sx:ShapeOr] ;
sx:shapeExprs @<shapeExprList2Plus>
}
<ShapeAnd> CLOSED {
a [sx:ShapeAnd] ;
sx:shapeExprs @<shapeExprList2Plus>
}
<ShapeNot> CLOSED {
a [sx:ShapeNot] ;
sx:shapeExpr @<shapeExpr>
}
<NodeConstraint> CLOSED {
a [sx:NodeConstraint] ;
sx:nodeKind [sx:iri sx:bnode sx:literal sx:nonliteral]?;
sx:datatype IRI ? ;
&<xsFacets> ;
sx:values @<valueSetValueList1Plus>?
}
<Shape> CLOSED {
a [sx:Shape] ;
sx:closed [true false]? ;
sx:extra IRI* ;
sx:expression @<tripleExpression>? ;
sx:semActs @<SemActList1Plus>? ;
}
<ShapeExternal> CLOSED {
a [sx:ShapeExternal] ;
}
<SemAct> CLOSED {
a [sx:SemAct] ;
sx:name IRI ;
sx:code xsd:string?
}
<Annotation> CLOSED {
a [sx:Annotation] ;
sx:predicate IRI ;
sx:object @<objectValue>
}
# <xsFacet> @<stringFacet> OR @<numericFacet><facet_holder> { # hold labeled productions
$<xsFacets> ( &<stringFacet> | &<numericFacet> )* ;
$<stringFacet> (
sx:length xsd:integer
| sx:minlength xsd:integer
| sx:maxlength xsd:integer
| sx:pattern xsd:string ; sx:flags xsd:string?
);
$<numericFacet> (
sx:mininclusive @<numericLiteral>
| sx:minexclusive @<numericLiteral>
| sx:maxinclusive @<numericLiteral>
| sx:maxexclusive @<numericLiteral>
| sx:totaldigits xsd:integer
| sx:fractiondigits xsd:integer
)
}
<numericLiteral> xsd:integer OR xsd:decimal OR xsd:double
<valueSetValue> @<objectValue> OR @<IriStem> OR @<IriStemRange>
OR @<LiteralStem> OR @<LiteralStemRange>
OR @<LanguageStem> OR @<LanguageStemRange><objectValue> IRI OR LITERAL # rdf:langString breaks on Annotation.object
<IriStem> CLOSED { a [sx:IriStem]; sx:stem xsd:anyUri }
<IriStemRange> CLOSED {
a [sx:IriStemRange];
sx:stem xsd:anyUri OR @<Wildcard>;
sx:exclusion @<objectValue> OR @<IriStem>*
}
<LiteralStem> CLOSED { a [sx:LiteralStem]; sx:stem xsd:string }
<LiteralStemRange> CLOSED {
a [sx:LiteralStemRange];
sx:stem xsd:string OR @<Wildcard>;
sx:exclusion @<objectValue> OR @<LiteralStem>*
}
<LanguageStem> CLOSED { a [sx:LanguageStem]; sx:stem xsd:string }
<LanguageStemRange> CLOSED {
a [sx:LanguageStemRange];
sx:stem xsd:string OR @<Wildcard>;
sx:exclusion @<objectValue> OR @<LanguageStem>*
}
<Wildcard> BNODE CLOSED {
a [sx:Wildcard]
}
<tripleExpression> @<TripleConstraint> OR @<OneOf> OR @<EachOf><OneOf> CLOSED {
a [sx:OneOf] ;
sx:min xsd:integer? ;
sx:max xsd:integer? ;
sx:expressions @<tripleExpressionList2Plus> ;
sx:semActs @<SemActList1Plus>? ;
sx:annotation @<Annotation>*
}
<EachOf> CLOSED {
a [sx:EachOf] ;
sx:min xsd:integer? ;
sx:max xsd:integer? ;
sx:expressions @<tripleExpressionList2Plus> ;
sx:semActs @<SemActList1Plus>? ;
sx:annotation @<Annotation>*
}
<tripleExpressionList2Plus> CLOSED {
rdf:first @<tripleExpression> ;
rdf:rest @<tripleExpressionList1Plus>
}
<tripleExpressionList1Plus> CLOSED {
rdf:first @<tripleExpression> ;
rdf:rest [rdf:nil] OR @<tripleExpressionList1Plus>
}
<TripleConstraint> CLOSED {
a [sx:TripleConstraint] ;
sx:inverse [true false]? ;
sx:negated [true false]? ;
sx:min xsd:integer? ;
sx:max xsd:integer? ;
sx:predicate IRI ;
sx:valueExpr @<shapeExpr>? ;
sx:semActs @<SemActList1Plus>? ;
sx:annotation @<Annotation>*
}
<SemActList1Plus> CLOSED {
rdf:first @<SemAct> ;
rdf:rest [rdf:nil] OR @<SemActList1Plus>
}
<shapeExprList2Plus> CLOSED {
rdf:first @<shapeExpr> ;
rdf:rest @<shapeExprList1Plus>
}
<shapeExprList1Plus> CLOSED {
rdf:first @<shapeExpr> ;
rdf:rest [rdf:nil] OR @<shapeExprList1Plus>
}
<valueSetValueList1Plus> CLOSED {
rdf:first @<valueSetValue> ;
rdf:rest [rdf:nil] OR @<valueSetValueList1Plus>
}
C. IANA Considerations
This section has been submitted to the Internet Engineering Steering
Group (IESG) for review, approval, and registration with IANA.
text/shex
Type name:
text
Subtype name:
shex
Required parameters:
None
Encoding considerations:
ShEx Compact Syntax (ShExC) is a text language which is encoded in UTF-8.
Security considerations:
Given that ShExC allows the substitution of long IRIs with short terms,
ShExC documents may expand considerably when processed and, in the worst case,
the resulting data might consume all of the recipient's resources. Applications
should treat any data with due skepticism.
Any programming environment that requires the exchange of
directed graphs. Implementations of ShEx have been created for
JavaScript, Python, Ruby, and Java.
Additional information:
Magic number(s):
Not Applicable
File extension(s):
.shex
Macintosh file type code(s):
TEXT
Person & email address to contact for further information:
Eric Prud'hommeaux <eric@w3.org>
Intended usage:
Common
Restrictions on usage:
None
Author(s):
Eric Prud'hommeaux, Iovka Boneva, Jose Labra Gayo, Gregg Kellogg, Niklas Lindström
Revealing the structure of an RDF graph can reveal information about the content of conformant data.
For instance, a schema with a predicate to describe cancer stage indicates that conforming graphs describe patients with cancer.
The process of testing a graph's conformance to a schema may involve many detailed queries which could draw resources to respond to API calls or SPARQL queries.
ShEx has an extension mechanism which can, in principle, evalute arbitrary code, possibly as some trusted agent.
Such extensions should not be executed if they don't come from a trusted source.
Since ShEx is intended to be a pure data exchange format for validating RDF graphs, the ShExJ serialization SHOULD NOT be passed through a code execution mechanism such as JavaScript's eval() function to be parsed.
An (invalid) document may contain code that, when executed, could lead to unexpected side effects compromising the security of a system.