Finish 0.6.3

Author: gkellogg

README.md

@@ -18,7 +18,7 @@ This is a pure-Ruby library for working with the [Shape Expressions Language][Sh
 
 The ShEx gem implements a [ShEx][ShExSpec] Shape Expression engine version 2.0.
 
-* `ShEx::Parser` parses ShExC and ShExJ formatted documents generating executable operators which can be serialized as [S-Expressions](https://en.wikipedia.org/wiki/S-expression).
+* `ShEx::Parser` parses ShExC and ShExJ formatted documents generating executable operators which can be serialized as [S-Expressions][].
 * `ShEx::Algebra` executes operators against Any `RDF::Graph`, including compliant [RDF.rb][].
 * [Implementation Report](file.earl.html)
 
@@ -183,11 +183,9 @@ Example usage:
 
 
 ## Implementation Notes
-The ShExC parser uses the [EBNF][] gem to generate first, follow and branch tables, and uses the `Parser` and `Lexer` modules to implement the ShExC parser.
+The ShExC parser uses the [EBNF][] gem to generate a [PEG][] parser.
 
-The parser takes branch and follow tables generated from the [ShEx Grammar](file.shex.html) described in the [specification][ShExSpec]. Branch and Follow tables are specified in the generated {ShEx::Meta}.
-
-The result of parsing either ShExC or ShExJ is the creation of a set of executable {ShEx::Algebra} Operators which are directly executed to perform shape validation.
+The parser uses the executable [S-Expressions][] generated from the EBNF ShExC grammar to create a set of executable {ShEx::Algebra} Operators which are directly executed to perform shape validation.
 
 ## Dependencies
 
@@ -260,3 +258,5 @@ see <https://unlicense.org/> or the accompanying {file:LICENSE} file.
 [YARD]:         https://yardoc.org/
 [YARD-GS]:      https://rubydoc.info/docs/yard/file/docs/GettingStarted.md
 [PDD]:              https://unlicense.org/#unlicensing-contributions
+[PEG]:          https://en.wikipedia.org/wiki/Parsing_expression_grammar "Parsing Expression Grammar"
+[S-Expression]: https://en.wikipedia.org/wiki/S-expression

Rakefile

@@ -22,39 +22,41 @@ task default: :spec
 RSpec::Core::RakeTask.new(:spec)
 
 desc 'Create versions of ebnf files in etc'
-task etc: %w{etc/shex.sxp etc/shex.html etc/shex.ll1.sxp}
+task etc: %w{etc/shex.sxp etc/shex.html etc/shex.peg.sxp}
 
 desc 'Build first, follow and branch tables'
 task meta: "lib/shex/meta.rb"
 
 file "lib/shex/meta.rb" => "etc/shex.ebnf" do |t|
   sh %{
-    ebnf --ll1 shexDoc --format rb \
+    ebnf --peg --format rb \
+      --input-format native \
       --mod-name ShEx::Meta \
       --output lib/shex/meta.rb \
       etc/shex.ebnf
   }
 end
 
-file "etc/shex.ll1.sxp" => "etc/shex.ebnf" do |t|
+file "etc/shex.peg.sxp" => "etc/shex.ebnf" do |t|
   sh %{
-    ebnf --ll1 shexDoc --format sxp \
-      --output etc/shex.ll1.sxp \
+    ebnf --peg --format sxp \
+      --input-format native \
+      --output etc/shex.peg.sxp \
       etc/shex.ebnf
   }
 end
 
 file "etc/shex.sxp" => "etc/shex.ebnf" do |t|
   sh %{
-    ebnf --bnf --format sxp \
+    ebnf --input-format native --format sxp \
       --output etc/shex.sxp \
       etc/shex.ebnf
   }
 end
 
 file "etc/shex.html" => "etc/shex.ebnf" do |t|
   sh %{
-    ebnf --format html \
+    ebnf --input-format native --format html \
       --output etc/shex.html \
       etc/shex.ebnf
   }

VERSION

@@ -1 +1 @@
-0.6.2
+0.6.3

etc/shex.ebnf

@@ -1,75 +1,3 @@
-# Yacker grammer for (Sh)ape (Ex)pressions (C)ompact language
-# 
-# Copyright 2015, Eric Prud'hommeax, Harold Solbrig, Iovka Boneva, Jose Labra Gayo
-# All rights reserved; please contact copyright holders for any use outside of the ShEx semantics.
-# http://www.w3.org/2005/01/yacker/uploads/ShEx2/bnf
-#
-# yacker: https://www.w3.org/2005/01/yacker/uploads/ShEx3?lang=perl
-# repo: https://github.com/shexSpec/shexTest/
-# branch: TEisSE
-# TODO:
-#   annotations on shapes
-#   look hard at EXTERNAL valueClassDefinitions
-# 
-# Changes from yacker ShEx2:
-# 
-# + '*' in 2nd pos of REPEAT_RANGE so e.g. {2,*} === {2,}
-# ~ code names are iris. this changes existing semantic actions as follows:
-#   <S> { :p1 . %GenX{ blah blah blah %} }
-#   PREFIX GenX: <http://...GenX> <S> { :p1 . %GenX:{ blah blah blah %} }
-#   empty code names %{ ... %} are parsed as empty relative URLs: %<>{ ... %}
-# + ATPNAME_NS, ATPNAME_LN terminals to parse e.g. @ex:foo which looks like a LANGTAG
-# ~ a is an iri (not just for predicates) enabling e.g. <reif> { rdf:predicate (a) }
-# ~ groupShapeConstr can only have ORs (no ANDs) so no <S1> { :p1 @<S2> AND @<S3> }
-# + includeSet parallels inclPropertySet (&<A><B><C>) e.g. <S3> &<S1> <S2> { ... }
-# + factored out (labeled) productions for startActions and semanticActions
-# ~ BNodes can have stringFacets
-# - ANON terminal (not referenced)
-# ~ tripleConstraint: swapped position of annotations and cardinality (EGP 20150727)
-# + valueClass: add xsFacet* to datatype (EGP 20150727)
-# + escapes (\\, \uxxxx, \Uxxxxxxxx) in CODE (EGP 20150730)
-# ~ moved | RDF_TYPE from iri: to predicate: (disables e.g. %a{ ... %}) (EGP 20150807)
-# + tried to intercept non-terminated STRING_LITERAL_LONG{1,2} (EGP 20150807)
-# - unaryShape: id (EGP 20150818)
-# + unaryShape: annotation* (EGP 20150818)
-# + valueClassLabel, valueClassDefinition, valueClassExpr, valueClassOrRef (EGP 20150818)
-# + factored out notStartAction (EGP 20150818)
-# - oneOf (EGP 20150820)
-# ~ split out encapsulatedShape (EGP 20150831)
-# - numericRange takes a numericLiteral, was INTEGER (EGP 20150917)
-# + innerShape ::= multiElementGroup | multiElementSomeOf (EGP 20151022)
-# ~ annotation ::= ';' predicate (iri | literal) -- was iri (EGP 20151022)
-# + generalized AND/OR expressions on valueClasses (EGP 20151031)
-# - groupShapeConstr (shapeOrRef (OR shapeOrRef)*) (EGP 20151031)
-# + stringFacet* on negatableValueClass ::= shapeOrRef (EGP 20151101)
-# ~ vcand and vcor are mutually exclusive (i.e. not recursive) (EGP 20151111)
-# ~ s/valueClassDefinition/valueExprDefinition/ (EGP 20151120)
-# ~ s/valueClassLabel/valueExprLabel/ (EGP 20151120)
-# ~ change valueSet delimiters from ()s to []s (EGP 20160104)
-# ~ codeDecls can have no code (implies code ref) (EGP 20160128)
-# + ()s around valueExprs (EGP 20160520)
-# + shapeDefinition += nonLiteralKind? stringFacet* (EGP 20160520)
-# - shape == "VIRTUAL"? (EGP 20160520)
-# + ';' separator for groups (EGP 20160524)
-# ~ s{;}{//} for annotations (EGP 20160524)
-# + ("AND" shapeDefinition)* (EGP 20160524)
-# + ("OR" shapeDefinition)* (EGP 20160615)
-# ~ generalized shapeExpression to include valueExprDefinition (EGP 20160708)
-# ~ reordered (EGP 20160913)
-# ~ s/ShapeDisjunction/shapeOr/ (EGP 20160913)
-# ~ s/shapeConjunction/shapeAnd/ (EGP 20160913)
-# ~ s/negShapeAtom/shapeNot/ (EGP 20160913)
-# ~ s/inclPropertySet/extraPropertySet/ (EGP 20160913)
-# ~ s/value/valueSetValue/ (EGP 20160913)
-# ~ updated production labels (EGP 20160913)
-# ~ reorder to align with spec (EGP 20160930)
-# ~ s/someOfShape/someOfTripleExpr/ (EGP 20160930)
-# ~ s/innerShape/innerTripleExpr/ (EGP 20160930)
-# ~ s/groupShape/groupTripleExpr/ (EGP 20160930)
-# ~ s/unaryShape/unaryTripleExpr/ (EGP 20160930)
-# ~ s/encapsulatedShape/bracketedTripleExpr/ (EGP 20160930)
-# ~ s/SomeOf/OneOf/g (EGP 20161201)
-
 # Notation:
 #   in-line terminals in ""s are case-insensitive
 #   production numbers ending in t or s are from Turtle or SPARQL.
@@ -87,15 +15,13 @@
 
 [8]     statement             ::= directive | notStartAction
 
-[9]     shapeExprDecl         ::= shapeLabel (shapeExpression|"EXTERNAL")
-[10]    shapeExpression       ::= shapeAtomNoRef shapeOr?
-                                | "NOT" (shapeAtomNoRef | shapeRef) shapeOr?
+[9]     shapeExprDecl         ::= shapeExprLabel (shapeExpression | "EXTERNAL")
+[10]    shapeExpression       ::= "NOT"? shapeAtomNoRef shapeOr?
+                                | "NOT" shapeRef shapeOr?
                                 | shapeRef shapeOr
 [11]    inlineShapeExpression ::= inlineShapeOr
-# [12]    shapeOr               ::= shapeAnd ("OR" shapeAnd)*
-[12]    shapeOr               ::= shapeOrA | shapeOrB shapeOrA?
-[12a]   shapeOrA              ::= ("OR" shapeAnd)+
-[12b]   shapeOrB              ::= ("AND" shapeNot)+
+[12]    shapeOr               ::= ("OR" shapeAnd)+
+                                | ("AND" shapeNot)+ ("OR" shapeAnd)*
 [13]    inlineShapeOr         ::= inlineShapeAnd ("OR" inlineShapeAnd)*
 [14]    shapeAnd              ::= shapeNot ("AND" shapeNot)*
 [15]    inlineShapeAnd        ::= inlineShapeNot ("AND" inlineShapeNot)*
@@ -119,9 +45,10 @@
 
 [21]    shapeOrRef            ::= shapeDefinition | shapeRef
 [22]    inlineShapeOrRef      ::= inlineShapeDefinition | shapeRef
-[23]    shapeRef              ::= ATPNAME_LN | ATPNAME_NS | '@' shapeLabel
+[23]    shapeRef              ::= ATPNAME_LN | ATPNAME_NS | '@' shapeExprLabel
 
 [24]    litNodeConstraint     ::= "LITERAL" xsFacet*
+                                | nonLiteralKind stringFacet*
                                 | datatype xsFacet*
                                 | valueSet xsFacet*
                                 | numericFacet+
@@ -137,20 +64,19 @@
 [31]    numericRange          ::= "MININCLUSIVE" | "MINEXCLUSIVE" | "MAXINCLUSIVE" | "MAXEXCLUSIVE"
 [32]    numericLength         ::= "TOTALDIGITS" | "FRACTIONDIGITS"
 
-[33]    shapeDefinition       ::= (includeSet | extraPropertySet | "CLOSED")* '{' tripleExpression? '}' annotation* semanticActions
-[34]    inlineShapeDefinition ::= (includeSet | extraPropertySet | "CLOSED")* '{' tripleExpression? '}'
+[33]    shapeDefinition       ::= (extraPropertySet | "CLOSED")* '{' tripleExpression? '}' annotation* semanticActions
+[34]    inlineShapeDefinition ::= (extraPropertySet | "CLOSED")* '{' tripleExpression? '}'
 [35]    extraPropertySet      ::= "EXTRA" predicate+
 
 [36]    tripleExpression      ::= oneOfTripleExpr
 
-# First/First Conflicts on "&", "$", "(", "^", :RDF_TYPE , :IRIREF, :PNAME_LN and :PNAME_NS 
+# oneOfTripleExpr and multiElementOneOf both start with groupTripleExpr
 #[37]    oneOfTripleExpr      ::= groupTripleExpr | multiElementOneOf
 #[38]    multiElementOneOf    ::= groupTripleExpr ('|' groupTripleExpr)+
 #[39]    innerTripleExpr       ::= multiElementGroup | multiElementOneOf
 [37]    oneOfTripleExpr      ::= groupTripleExpr ('|' groupTripleExpr)*
 
-# First/First Conflicts on "&", "$", "(", "^", :RDF_TYPE , :IRIREF, :PNAME_LN and :PNAME_NS 
-# First/Follow Conflict on ";"
+# singleElementGroup and multiElementGroup both start with unaryTripleExpr
 #[40]    groupTripleExpr       ::= singleElementGroup | multiElementGroup
 #[41]    singleElementGroup    ::= unaryTripleExpr ';'?
 #[42]    multiElementGroup     ::= unaryTripleExpr (';' unaryTripleExpr)+ ';'?
@@ -179,7 +105,7 @@
 [55]    languageRange         ::= LANGTAG ('~' languageExclusion*)?
 [56]    languageExclusion     ::= '-' LANGTAG '~'?
 
-[57]    include               ::= '&' shapeLabel
+[57]    include               ::= '&' tripleExprLabel
 
 [58]    annotation            ::= '//' predicate (iri | literal)
 [59]    semanticActions       ::= codeDecl*
@@ -188,20 +114,20 @@
 [13t]   literal               ::= rdfLiteral | numericLiteral | booleanLiteral
 [61]    predicate             ::= iri | RDF_TYPE
 [62]    datatype              ::= iri
-[63]    shapeLabel            ::= iri | blankNode
-[16t]   numericLiteral        ::= INTEGER | DECIMAL | DOUBLE
-[129s]  rdfLiteral            ::= langString | string ('^^' datatype)?
+[63]    shapeExprLabel        ::= iri | blankNode
+[64]   	tripleExprLabel       ::= iri | blankNode
+
+[16t]   numericLiteral        ::= DOUBLE | DECIMAL | INTEGER
+[65]    rdfLiteral            ::= langString | string ('^^' datatype)?
 [134s]  booleanLiteral        ::= 'true' | 'false'
-[135s]  string                ::= STRING_LITERAL1 | STRING_LITERAL_LONG1
-                                | STRING_LITERAL2 | STRING_LITERAL_LONG2
-[66]   langString            ::= LANG_STRING_LITERAL1 | LANG_STRING_LITERAL_LONG1
+[135s]  string                ::= STRING_LITERAL_LONG1 | STRING_LITERAL_LONG2
+                                | STRING_LITERAL1 | STRING_LITERAL2
+[66]    langString            ::= LANG_STRING_LITERAL1 | LANG_STRING_LITERAL_LONG1
                                 | LANG_STRING_LITERAL2 | LANG_STRING_LITERAL_LONG2
 [136s]  iri                   ::= IRIREF | prefixedName
 [137s]  prefixedName          ::= PNAME_LN | PNAME_NS
 [138s]  blankNode             ::= BLANK_NODE_LABEL
 
-# Reserved for future use
-#[65]   	includeSet	          ::= "&" (shapeLabel)+
 @terminals
 
 [67]    CODE                  ::= '{' ([^%\\] | '\\' [%\\] | UCHAR)* '%' '}'