Antlr4语法围绕字符串可选括号

Question

I am trying to write a grammar to evaluate an equation which may or may not be surrounded by brackets. ex -

1. (NOT str1 matches 'hello')
2. NOT (str1 matches 'hello')
3. (NOT (str1 matches 'hello'))
4. NOT str1 matches 'hello'

In my rules, I have

clause
: expression EOF
;

expression
 :
   LPAREN expression RPAREN                            #parenExpression
 | isNumeric                                     #isNumericExpression
 | leftSide IS NOT? NULL                               #nullExpression
 | compare                                           #comparatorExpression
 | NOT                                                 #notExpression

;


compare : NOT? LPAREN? NOT? leftSide op=comparator rightSide RPAREN? ;

Now, there are a couple of problems with this.

1. I don't check whether left and right brackets match.
2. There is at most ONE NOT in the clause.

Any help on how to write grammar for these rules would be super helpful.

Thanks in advance!

Answer 1

Conceptually, expression are (typically) either complex, ie , recursive, or simple:

expression
   : LPAREN expression RPAREN                 #parenExpression
   | NOT expression                           #notExpression
   | simpleExpression                         
   ;

simpleExpression
   : isNumeric                                #isNumericExpression
   | leftSide IS NOT? NULL                    #nullExpression
   | leftSide op rightSide                    #comparatorExpression
   ;

The first complex alt allows any level of balanced parenthesis around an expression. The second allows one or more NOT s to precede any expression.

The simple alts represent the most basic forms of an expression.

So, (NOT (str1 matches 'hello')) will be matched by

parenExpression -> notExpression -> parenExpression -> comparatorExpression .

Of course, the two rules could be combined into one.