add SPARQL & Turtle highlighter

[VladimirAlexiev]

See https://www.w3.org/TR/sparql11-query/#sparqlGrammar

• http://rawgit2.com/VladimirAlexiev/grammar-diagrams/master/sparql11-grammar.ebnf for plain text version
• http://rawgit2.com/VladimirAlexiev/grammar-diagrams/master/sparql11-grammar.xhtml for railroad syntax diagrams.

I'm thinking of using this as a model:
https://github.com/jgm/skylighting/blob/master/skylighting-core/xml/sql.xml
@jgm, could you give me some pointers? I have a lot of questions (eg what is #stay and #pop, what is the list of reusable itemDatas...)

Guess I should first read https://docs.kde.org/stable5/en/applications/katepart/highlight.html,
then diff https://github.com/KDE/syntax-highlighting/blob/master/data/syntax/sql.xml to the above to see what changes skylighting has compared to KDE/syntax-highlighting?

My next question would be about Turtle: https://www.w3.org/TR/turtle/#h3_sec-grammar-grammar. It has very few keywords and relatively simple structure that consists mostly of terms.

Thanks for all your help!

[VladimirAlexiev]

Emacs does it with regexps:

(defconst sparql-keywords-re
  (regexp-opt
   '("ADD" "ALL" "AS" "ASC" "ASK"
     "BASE" "BIND" "BINDINGS" "BY"
     "CLEAR" "CONSTRUCT" "COPY" "CREATE"
     "DATA" "DEFAULT" "DELETE" "DESC" "DESCRIBE" "DISTINCT" "DROP"
     "EXISTS"
     "FILTER" "FROM"
     "GRAPH" "GROUP"
     "HAVING"
     "IN" "INSERT" "INTO"
     "LIMIT" "LOAD"
     "MINUS" "MOVE"
     "NAMED" "NOT"
     "OFFSET" "OPTIONAL" "ORDER"
     "PREFIX"
     "REDUCED"
     "SELECT" "SERVICE" "SILENT"
     "TO"
     "UNDEF" "UNION" "USING"
     "WHERE" "WITH")
   'symbols))

(defconst sparql-keywords
  `((,(rx "<" (* (not space)) ">")
     0 font-lock-constant-face t)
    (,(rx bol (*? (not (in "#"))) (group "\"" (* (not (in "\""))) "\""))
     1 font-lock-string-face)
    (,(rx bol (*? (not (in "#"))) (group "'" (* (not (in "'"))) "'"))
     1 font-lock-string-face)
    (,(rx (*? not-newline) (group "#" (* not-newline)))
     1 font-lock-comment-face)
    ,sparql-keywords-re
    (,(rx (any "?$") (+ word))
     0 font-lock-variable-name-face)))

Turtle (n3) is basically a subset of sparql:

(setq n3-highlights
  '(("\\(@prefix\\)\\>" 1 font-lock-keyword-face t)
    ("\\(a\\)\\>" 1 font-lock-keyword-face t)
    ("\\(\\S-*?:\\)" 1 font-lock-type-face t)
    (":\\(.+?\\)[ ;.]" 1 font-lock-constant-face t)
    ("\\(<.*?>\\)" 1 font-lock-function-name-face t)
    ("\\(\\\".*?\\\"\\)" 1 font-lock-string-face t)
    ; Bug: some trailing characters are highlighted; restricting comments regexp
    ; ("\\(#.*\\)" 1 font-lock-comment-face t)
    ("^\\s-*\\(#.*\\)" 1 font-lock-comment-face t)
    )
)

[jgm]

Yes, read the KDE docs. skylighting is nearly 100% compatible with existing KDE syntax definitions -- there are just occasionally small tweaks needed due to differences in the regex engine.

[VladimirAlexiev]

There is Turtle & SPARQL highlighting in highlight.js.

• using with pandoc is discussed at https://groups.google.com/g/pandoc-discuss/c/GVJ1XRFp0zE (but highlight.js is again being maintained)
• discussed at https://www.artembutusov.com/pandoc-markdown-syntax-highlighting-with-highlight-js/, but using dynamic JS highlighting
• https://transparency.ontotext.com/spec/#countries is made from markdown using pandoc, and shows SPARQL highlighting, using static hljs spans: I think this is better. @viktor-ribchev can you describe how you do that?

@jgm, what are the disadvantages of using highlight.js instead of skylighting ?

[jgm]

Skylighting will give you coloring support in many of the formats pandoc supports, and the coloring will be baked into the document (not requiring a javascript interpreter). But feel free to use highlighting.js if it meets your needs!