Import: Lexer for SPARQL (#49)

lib/rouge/demos/sparql

@@ -0,0 +1,6 @@
+SELECT ?item ?itemLabel
+WHERE
+{
+  ?item wdt:P31 wd:Q146.
+  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
+}

lib/rouge/lexers/sparql.rb

@@ -0,0 +1,60 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+module Rouge
+  module Lexers
+    class SPARQL < RegexLexer
+      title "SPARQL"
+      desc "Semantic Query Language, for RDF data"
+      tag 'sparql'
+      filenames '*.rq'
+      mimetypes 'application/sparql-query'
+
+      BUILTINS = Set.new %w[
+        ABS AVG BNODE BOUND CEIL COALESCE CONCAT CONTAINS COUNT DATATYPE DAY
+        ENCODE_FOR_URI FLOOR GROUP_CONCAT HOURS IF IRI isBLANK isIRI isLITERAL
+        isNUMERIC isURI LANG LANGMATCHES LCASE MAX MD5 MIN MINUTES MONTH NOW
+        RAND REGEXP REPLACE ROUND sameTerm SAMPLE SECONDS SEPARATOR SHA1
+        SHA256 SHA384 SHA512 STR STRAFTER STRBEFORE STRDT STRENDS STRLANG
+        STRLEN STRSTARTS STRUUID SUBSTR SUM TIMEZONE TZ UCASE URI UUID YEAR
+      ]
+
+      KEYWORDS = Set.new %w[
+        ADD ALL AS ASC BIND CLEAR COPY CREATE DATA DEFAULT DELETE DESC
+        DISTINCT DROP EXISTS FILTER GRAPH GROUP\ BY HAVING IN INSERT LIMIT
+        LOAD MINUS MOVE NAMED NOT\ EXISTS NOT\ IN OFFSET OPTIONAL ORDER\ BY
+        SELECT SERVICE SILENT UNDEF UNION USING VALUES WHERE WITH
+      ]
+
+      state :root do
+        rule %r("), Str, :string_double
+        rule %r('), Str, :string_single
+        rule %r(#.*), Comment::Single
+        rule %r([$?]\w+), Name::Variable
+        rule %r((\w*:)(\w+)) do |m|
+          token Name::Namespace, m[1]
+          token Str::Symbol, m[2]
+        end
+        rule %r(<[^>]*>), Name::Namespace
+        rule Regexp.union(KEYWORDS.map{ |str| /\b#{str}\b/i }), Keyword
+        rule Regexp.union(BUILTINS.map{ |str| /\b#{str}\b/i }), Name::Builtin
+        rule %r(-?([0-9]+\.[0-9]+|\.[0-9]+|[0-9]+)([eE][+-]?[0-9]+)?), Num
+        rule %r([\]\[(){}.,;=]), Punctuation
+        rule %r([/?*+=!<>]|&&|\|\|), Operator
+        rule %r(\s+), Text::Whitespace
+      end
+
+      state :string_double do
+        rule %r(\\[tbnrf"'\\]), Str::Escape
+        rule %r("), Str, :pop!
+        rule %r(.), Str
+      end
+
+      state :string_single do
+        rule %r(\\[tbnrf"'\\]), Str::Escape
+        rule %r('), Str, :pop!
+        rule %r(.), Str
+      end
+    end
+  end
+end

spec/lexers/sparql_spec.rb

@@ -0,0 +1,18 @@
+# -*- coding: utf-8 -*- #
+# frozen_string_literal: true
+
+describe Rouge::Lexers::SPARQL do
+  let(:subject) { Rouge::Lexers::SPARQL.new }
+
+  describe 'guessing' do
+    include Support::Guessing
+
+    it 'guesses by filename' do
+      assert_guess :filename => 'foo.rq'
+    end
+
+    it 'guesses by mimetype' do
+      assert_guess :mimetype => 'application/sparql-query'
+    end
+  end
+end

spec/visual/samples/sparql

@@ -0,0 +1,41 @@
+#Artists entering in public domain by countries
+#added before 2017-02
+
+SELECT ?item ?itemLabel ?genderLabel (GROUP_CONCAT(DISTINCT ?occupationLabel; SEPARATOR=", ") AS ?occupations) (GROUP_CONCAT(DISTINCT ?countryLabel; SEPARATOR=", ") AS ?countries) ?death {
+   VALUES ?target_country { wd:Q16 wd:Q31 wd:Q39 wd:Q142 } . #countries: Canada, France, Switzerland, Belgium. Remove this line to get worldwide.
+  VALUES ?occupation { wd:Q483501 wd:Q36834 wd:Q639669 wd:Q753110 wd:Q488205 wd:Q49757 wd:Q4964182 wd:Q1281618 wd:Q36180 wd:Q482980 wd:Q1028181 wd:Q6625963 wd:Q28389 wd:Q1930187 wd:Q33999 wd:Q3282637 wd:Q81096 wd:Q201788 wd:Q18939491 wd:Q486748 wd:Q3658608 wd:Q214917 wd:Q11774202 wd:Q205375 } . #occupation: composer, poet, sculptor, writer, artist, painter, etc.
+   ?item wdt:P31 wd:Q5;
+               wdt:P21 ?gender;
+               wdt:P570 ?death;
+               wdt:P27 ?target_country;
+               wdt:P27 ?country;
+               wdt:P106 ?occupation .
+   FILTER( YEAR( ?death ) = 1966 ) .
+   SERVICE wikibase:label {
+       bd:serviceParam wikibase:language "fr,en,ru,el,es,fa" .
+       ?item rdfs:label ?itemLabel .
+       ?gender rdfs:label ?genderLabel .
+       ?occupation rdfs:label ?occupationLabel .
+       ?country rdfs:label ?countryLabel .
+   } .
+} GROUP BY ?item ?itemLabel ?genderLabel ?death ORDER BY ?itemLabel
+
+
+SELECT ?material $VAR (concat("test#\"foo", 'foo\'bar\n') AS ?x_2) (-.123 AS ?num)
+where {
+  ?painting wdt:P31/wdt:P279 wd:Q3305213; # end of line comment
+            p:P186 [ ps:P186 ?material; pq:P518 wd:Q861259 ].
+  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE]". }
+}
+ORDER BY DESC(?count)
+
+
+#illustrates sitelink selection, ";" notation
+SELECT ?item ?itemLabel ?article (-200.23e+3 AS ?exp)
+WHERE
+{
+	?article 	schema:about ?item ;
+			schema:isPartOf <https://species.wikimedia.org/> .
+	SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en" }
+}
+LIMIT +200