The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.

NAME

Wraith - Parser Combinator in Perl

SYNOPSIS

    use Wraith;

    my ($expn, $term, $factor, $num);
    wraith_rule->makerules(\$expn, \$term, \$factor, \$num);

    $expn = ( (\$term >> $wraith::token->('\+') >> \$expn) ** sub { [ $_[0]->[0] + $_[0]->[2] ] } ) |
            ( (\$term >> $wraith::token->('-') >> \$expn) ** sub { [ $_[0]->[0] - $_[0]->[2] ] } ) |
            ( \$term );

    $term = ( (\$factor >> $wraith::token->('\*') >> \$term) ** sub { [ $_[0]->[0] * $_[0]->[2] ] } ) |
            ( (\$factor >> $wraith::token->('\/') >> \$term) **
                sub { $_[0]->[2] ? [ $_[0]->[0] / $_[0]->[2] ] : [] } ) |
            ( \$factor );

    $factor = ( (\$num) ** sub { my $args = $_[0]; my $val = undef; for my $elt (@$args) { $val .= $elt; } [ $val ] } ) |
              ( ( $wraith::token->('\(') >> \$expn >> $wraith::token->('\)') ) ** sub { my $args = $_[0]; [ $args->[1] ] } );

    $num = $wraith::token->('[1-9][0-9]*');

    print $expn->('2 + (4 - 1) * 3 + 4 -2')->[0]->[0]->[0], "\n";

DESCRIPTION

    Wraith is a simple parser combinator library (not monadic nor memoized) inspired
    by Boost.Spirit. It is not complete as Spirit but the fundamental operators are
    implemented.

    When applied with arguments, all operators/combinators return a function, which 
    takes a string as input sentence(s) and return a reference to a list of pairs:
    [ $pair_1, $pair_2, ..., $pair_n ],
    where each pair is a reference to a two-element list:
    [ ref_to_list_of_results, input_unprocessed ],
    in which ref_to_list_of_results is a reference to a list of analysis results and
    input_unprocessed is a string representing the unprocessed input so far.


    =head2 Basic Operators:

        =head3 reference $succeed

            It is a curried version of operator succeed. The first parameter of succeed is
            the analysis result and the second parameter is the unprocessed input string.

        =head3 reference $fail

            It takes an argument, discards it and return an empty list.

            Those two operators are rarely used. Use them if you need new combinators.

        =head3 reference $literal

            It takes one character as the only argument. The returned function match the first 
            character of its input against the argument character and return (argument, input_left)
            if matched, where input_left is the input without its first character, or return
            an empty list if failed to match.

        =head3 reference $literals

            Almost the same as $literal, but takes a string as the only argument and match
            the first character of input with each character in argument string until matched.

        =head3 reference $token

            Takes a regex string as its first argument. The second and optional argument is a
            regex string of skipped strings. It matches the regex at the beginning of the
            input string, return (token, input_left) if matched or an empty list if failed.

    =head2 Combinators:

        There are four combinators: then for sequence, alt for alternative, many for kleene
        star and using for semantic actions. Except many, the combinators are overloaded
        perl operators which takes at least one operator, combinator, compsite of combinators,
        product, or reference to an instance of those classes as the left-hand-side operand.

        The returned list of function generated by combinators is a list of tokens in the
        order of they appeared in the products.

        =head3 operator >> 

            Sequence combinator. For example, the product S -> T S would be written as
                $S = \$T >> $S;
            where $S and $T are rules, i.e, products.

        =head3 operator | 

            Alternative combinator. For example, the product S -> P | Q would be written as
                $S = \$T | \$Q;
            where $S, $T and $Q are rules.

        =head3 operator **

            Using combinator. It takes a operator, combinator, compsite of combinators, product,
            or reference to an instance of those classes as the left-hand-side operand and a
            subroutine as the right-hand-side operand. The returned value of lhs operand will
            be passed to rhs operand, and the returned value of rhs operand applied with its
            argument will be returned. This combinator is used for semantic actions.
            The returned value must be a reference to a list containing all the results given
            by the semantic subroutine.

        =head3 reference $many

            Kleene star combinator. The argument combinator will be matched at least zero time.
            The returned value is a list of all possible matchings.

    =head2 Rules:

        Rules are products. Products are compsite of operators and/or combinators. To create
        a product, a scalar variable must be declared,
            my $P;
        and then, call Wraith_rules->makerules(\$P) to make it a rule.

        =head3 Wraith_rules->makerules( @list_of_references_to_products )

            It takes a list of references to would-be rules and returned the blessed references.
            However the returned values can be omitted for the contents of the variables are
            already blessed. Thus, the variables are able to use the overloaded operators.

AUTHOR

    Bo Wang <sceneviper@hotmail.com>

COPYRIGHT

    Copyright 2013 - Bo Wang

SEE ALSO

    Parser::Combinators, which implements parsec-like parser combinators.

LICENSE

    This library is free software; you can redistribute it and/or modify it
    under the same terms as Perl itself.