Lucene Query Syntax in Lark

lucene-query.lark

start: clause_default

clause_default: clause_or+

clause_or: clause_and (or_ clause_and)*
clause_and: clause_not (and_ clause_not)*
clause_not: clause_basic (not_ clause_basic)*

clause_basic: modifier? LPAREN clause_default RPAREN term_modifier?
            | atom

atom: modifier? field multi_value term_modifier?
    | modifier? field? value term_modifier?

field: TERM_NORMAL COLON

value: range_term
     | normal
     | truncated
     | quoted
     | quoted_truncated
     | QMARK
     | anything
     | STAR

anything: STAR COLON STAR

two_sided_range_term: (LBRACK|LCURLY) range_value (TO? range_value)? (RBRACK|RCURLY)
range_term: two_sided_range_term

range_value: truncated
           | quoted
           | quoted_truncated
           | date
           | normal
           | STAR

multi_value: LPAREN clause_default RPAREN

normal: TERM_NORMAL
      | NUMBER

truncated: TERM_TRUNCATED

quoted_truncated: PHRASE_ANYTHING
quoted: PHRASE

modifier: PLUS
        | MINUS

term_modifier: boost fuzzy?
             | fuzzy boost?

boost: CARAT NUMBER?
fuzzy: TILDE NUMBER?

not_: AND NOT
    | NOT

and_: AND
or_: OR

date: DIGIT DIGIT? DATE_SEPARATOR DIGIT DIGIT? DATE_SEPARATOR DIGIT DIGIT (DIGIT DIGIT)?


LPAREN: "("
RPAREN: ")"
LBRACK: "["
RBRACK: "]"
COLON: ":"
PLUS: "+"
MINUS: /-|!/
STAR: "*"
QMARK: /\?+/
LCURLY: "{"
RCURLY: "}"
CARAT: "^"
TILDE: "~"
TO: "TO"
DIGIT: /\d/

AND: /(?i:and)|&&?/
OR: /(?i:or)|\|\|?/
NOT: /(?i:not)/

NUMBER: /\d+(\.\d+)?/
DATE_SEPARATOR: ("/"|"-"|".")

ESC_CHAR: /\\./
TERM_START_CHAR: /(\\.|[^ \t\n\r\u3000'"()\[\]{}+\-!:~^?*\\])/
TERM_CHAR: /(\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])/

TERM_NORMAL: /(\\.|[^ \t\n\r\u3000'"()\[\]{}+\-!:~^?*\\])(\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])*/

TERM_TRUNCATED: /(?:\*|\?+)(?:(?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])+(?:\?+|\*))+((?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])*)|(?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])(?:(?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])*(?:\?+|\*))+((?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])*)|(?:\*|\?+)(?:\\.|[^ \t\n\r\u3000'"()\[\]{}!:~^?*\\])+/

PHRASE: /"(\\.|[^"\\?*])+"/
PHRASE_ANYTHING: /"(\\.|[^"\\])+"/

WS: /[ \t\r\n\u3000]+/
%ignore WS

This was created by asking GPT-5 to convert the lucene query syntax from ANTLR to LARK. The ANTLR version is available here.

GPT-5 made an error with the date which I manually corrected.