How to ignore any text outside of brackets or parentheses

[Pooideae]

I cannot figure out how to ignore any text outside of brackets or parentheses using langium.
Here is an example of a tiny DSL:

any text {Expr1=[(Aa) any text (Bb)]} any text {Expr2=[ any text (Cc) any text (Dd) any text ]} any text

It consists of one or many commands. A command is enclosed by a simple pair of curly brackets: Expr1, Expr2
A command contains attributes which are enclosed in round brackets within a pair of square brackets: Aa, Bb, Cc, Dd
There can be additional text before, between and after the commands or attributes: any text. How to formulate the grammar file to ignore that additional text (like comments)?

grammar Test

entry Model: (cmd+=Command)*;
Command: '{' param=( Expr1 | Expr2 ) '}' ;
Expr1: 'Expr1' '=' '[' param+=Parameter* ']' ;
Expr2: 'Expr2' '=' '[' param+=Parameter* ']' ;
Parameter: '(' name=('Aa' | 'Bb' | 'Cc' | 'Dd') ')' ;

// Terminals
terminal STRING returns string:   /'(\\.|[^'\\])*'/;
terminal ID     returns string:   /[_a-zA-Z][\w_]*/;

//hidden terminal OUTSIDE_CURLYS: /[^}]+(?![^{]*})/;
//hidden terminal OUTSIDE_PARENS: /[^\)]+(?![^\(]*\])/;
//hidden terminal WORD:             /(?!\.\.)[^{}][^\s:]+/ ;
hidden terminal WS:               /\s+/;

[msujew]

Hey @Pooideae,

The default lexer won't get you very far with this, since the language you're attempting to build needs to encode state during the tokenization phase.

Essentially, you need to start with the free text state, enter the expression state with { and ( and leave it again on [. This is rather non-trivial, but here is a guide that tries to explain this.