boost regex to match IF statements

Question

I need to write a boost regex to match the following string and seperate it into three tokens depending on the parameters to the IF block

=IF(ISNUMBER(SEARCH("Windows",GETWORKSPACE(1))),ON.TIME(NOW()+"00:00:02","abcdef"),CLOSE(TRUE))

Ideally these should come to

token1 = "ISNUMBER(SEARCH("Windows",GETWORKSPACE(1)))"
token2 = "ON.TIME(NOW()+"00:00:02","abcdef")"
token3 = "CLOSE(TRUE)"

I had originally written a simple regex as "(?<=\=IF\()(.),(.),(.*)(?=\))" which gives out incorrect tokens because the greedy qualifier takes too much of the first token. I am currently getting

token1 =     "ISNUMBER(SEARCH("Windows",GETWORKSPACE(1))),ON.TIME(NOW()+"00:00:02""
token2 =     ""abcdef")"
token3 =     "CLOSE(TRUE)"

Also tried "(?<=\\=IF\\()([A-Za-z(),:\"]*?),([A-Za-z(),.:\"]*?),([A-Z(),:\"]*?)(?=\\))" with no luck. Can someone please suggest a regex ?

Answer 1

You need a simple parser.

Here's one with my favorite Boost swiss-army knife for quick parsers.

I've created a very flexible "token" grammar that honours (nested) parentheses and double-quoted string literals (potentially with embedded escaped quotes and parentheses):

token = raw [ *(
      '(' >> -token_list >> ')'
    | '[' >> -token_list >> ']'
    | '{' >> -token_list >> '}'
    | string_literal
    | lexeme[ + ~char_(")]}([{\"',") ]
    ) ];

Where token_list and string_literal are defined as

string_literal = lexeme [
    '"' >> *('\\' >> char_ | ~char_('"')) >> '"'
];

token_list = token % ',';

Now the parser expression for an =IF(condition, true_part, false_part) is simply:

if_expr
    = '=' >> no_case["if"] 
    >> '(' >> token >> ',' >> token >> ',' >> token >> ')';

For fun I made the IF keyword case-insensitive

DEMO

Live On Coliru

//#define BOOST_SPIRIT_X3_DEBUG
#include <boost/spirit/home/x3.hpp>
#include <boost/fusion/adapted/std_tuple.hpp>
#include <iostream>
#include <iomanip>
namespace x3 = boost::spirit::x3;

namespace parser {
    using namespace x3;

    static rule<struct token_, std::string> const token = "token";

    static auto const string_literal = lexeme [
        '"' >> *('\\' >> char_ | ~char_('"')) >> '"'
    ];

    static auto const token_list = token % ',';

    static auto const token_def = raw [ *(
          '(' >> -token_list >> ')'
        | '[' >> -token_list >> ']'
        | '{' >> -token_list >> '}'
        | string_literal
        | +~char_(")]}([{\"',")  // glue together everything else
        ) ];

    BOOST_SPIRIT_DEFINE(token)

    static auto const if_expr
        = '=' >> no_case["if"] 
        >> '(' >> token >> ',' >> token >> ',' >> token >> ')';
}

int main() {
    for (std::string const& input : {
            R"(=IF(ISNUMBER,ON.TIME,CLOSE))",
            R"(=IF(ISNUMBER(SEARCH("Windows")),ON.TIME(NOW()+"00:00:02","abcdef"),CLOSE(TRUE)))",
            R"(=IF(ISNUMBER(SEARCH("Windows",GETWORKSPACE(1))),ON.TIME(NOW()+"00:00:02","abcdef"),CLOSE(TRUE)))",
            " = if( isnumber, on .time, close ) ",
            R"( = if( "foo, bar", if( isnumber, on .time, close ), IF("[ISN(UM}B\"ER")) )",
        })
    {
        auto f = input.begin(), l = input.end();
        std::cout << "=== " << std::quoted(input) << ":\n";

        std::string condition, true_part, false_part;
        auto attr = std::tie(condition, true_part, false_part);

        if (phrase_parse(f, l, parser::if_expr, x3::blank, attr)) {
            std::cout << "Parsed: \n"
               << " - condition: " << std::quoted(condition) << "\n"
               << " - true_part: " << std::quoted(true_part) << "\n"
               << " - false_part: " << std::quoted(false_part) << "\n";
        } else {
            std::cout << "Parse failed\n";
        }

        if (f!=l) {
            std::cout << "Remaining unparsed: " << std::quoted(std::string(f,l)) << "\n";
        }
    }
}

Prints

=== "=IF(ISNUMBER,ON.TIME,CLOSE)":
Parsed: 
 - condition: "ISNUMBER"
 - true_part: "ON.TIME"
 - false_part: "CLOSE"
=== "=IF(ISNUMBER(SEARCH(\"Windows\")),ON.TIME(NOW()+\"00:00:02\",\"abcdef\"),CLOSE(TRUE))":
Parsed: 
 - condition: "ISNUMBER(SEARCH(\"Windows\"))"
 - true_part: "ON.TIME(NOW()+\"00:00:02\",\"abcdef\")"
 - false_part: "CLOSE(TRUE)"
=== "=IF(ISNUMBER(SEARCH(\"Windows\",GETWORKSPACE(1))),ON.TIME(NOW()+\"00:00:02\",\"abcdef\"),CLOSE(TRUE))":
Parsed: 
 - condition: "ISNUMBER(SEARCH(\"Windows\",GETWORKSPACE(1)))"
 - true_part: "ON.TIME(NOW()+\"00:00:02\",\"abcdef\")"
 - false_part: "CLOSE(TRUE)"
=== " = if( isnumber, on .time, close ) ":
Parsed: 
 - condition: "isnumber"
 - true_part: "on .time"
 - false_part: "close "
=== " = if( \"foo, bar\", if( isnumber, on .time, close ), IF(\"[ISN(UM}B\\\"ER\")) ":
Parsed: 
 - condition: "\"foo, bar\""
 - true_part: "if( isnumber, on .time, close )"
 - false_part: "IF(\"[ISN(UM}B\\\"ER\")"