REGEX - 匹配长字符串中的模式

Question

I am trying to match a particular pattern in a lengthy string:

NEW ZEALAND AND (data.operator1:"SHELL AND AMP" AND data.field:"NEW ZEALAND") OR (data.operator:purpose AND data.field:crank) OR (data.operator:REGULATOR AND data.field:HELICOPTOR)

1. I want to select all the below values followed by: but not the AND/OR/NOT operator.
2. I am trying to use look ahead and look after/behind feature in Regex but unable to achieve it

Basically a combination of /(?<?AND)(?<?OR)\s+(:!AND)(?!OR)/g and :" [a-zA-Z ] "

I want to change the strings to title case so that I can clearly distinguish AND/OR/NOT.

New Zealand AND (data.operator1:"Shell And Amp" AND data.field:"New Zealand") OR (data.operator:purpose AND data.field:crank) OR (data.operator:Regulator AND data.field:Helicoptor)

Answer 1

You can easily express lexers using regular expressions with named groups, for example:

const MY_LEXER = String.raw`
    (?<string> "[^"]*")
    |
    (?<operator> and|or|AND|OR)
    |
    (?<word> \w+)
    |
    (?<punct> [().:])
    |
    (?<ws> \s+)
`

The next function gets a string and a lexer and returns a list of pairs [token-type, token-value] :

let tokenize = (str, lexer) =>
    [...
        str.matchAll(
            lexer.replace(/\s+/g, ''))
    ]
        .flatMap(m =>
            Object
                .entries(m.groups)
                .filter(p => p[1]))

The result will be like

  [ 'word', 'NEW' ],
  [ 'ws', ' ' ],
  [ 'word', 'ZEALAND' ],
  [ 'ws', ' ' ],
  [ 'operator', 'AND' ],
  [ 'ws', ' ' ],
  [ 'punct', '(' ],

etc. Now it should be possible to iterate that, transform values as you need and put them back together:

for (let [type, val] of tokenize(myString, MY_LEXER)) {
    if (type === 'string' || type === 'word')
        val = val.toLowerCase();
    output += val;
}