Re: Parsing bracket formatted file (third update)

Something like this ?

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11147096
use warnings;

my $testdata = <<END;
sys ecm cloud-provider /Common/aws-ec2 {
    description "The aws-ec2 parameters"
    property-template {
        account { }
        availability-zone {
            valid-values { a b c d }
        }
        instance-type {
            valid-values { t2.micro t2.small t2.medium }
        }
        region {
            valid-values { us-east-1 us-west-1 }
        }
    }
}
END

local $_ = $testdata; # NOTE expr() expect input in $_
my $parse = expr();
use Data::Dump 'dd'; dd $parse;

use List::AllUtils qw( all );
sub fixhash
  {
  local $_ = shift;
  if( ref $_ eq 'ARRAY' )
    {
    all { ref $_ eq 'HASH' } @$_ and return { map { map fixhash($_), %
+$_ } @$_ };
    return [ map fixhash($_), @$_ ];
    }
  elsif( ref $_ eq 'HASH' )
    {
    return { map fixhash($_), %$_ };
    }
  else { return $_ };
  }

sub expr
  {
  /\G\s+/gc;
  my $e = [];
  $e =
    /\G\s+/gc ? $e :
    /\G([^{}\s]+) "(.*?)"/gc ? [ @$e, { $1 => $2 } ] :
    /\G([^{}\s]+) \{/gc ?  [ @$e, { "$1" => (expr(),
        /\G\}/gc || die pos($_), ' b missing }', substr $_, pos($_))[0
+] } ] :
    /\G([^{}\s]+)/gc ? [ @$e, $1 ] :
    return fixhash($e) while 1;
  }
[download]

Outputs:

[
  "sys",
  "ecm",
  "cloud-provider",
  {
    "/Common/aws-ec2" => {
      "description"       => "The aws-ec2 parameters",
      "property-template" => {
                               "account" => {},
                               "availability-zone" => { "valid-values"
+ => ["a" .. "d"] },
                               "instance-type" => { "valid-values" => 
+["t2.micro", "t2.small", "t2.medium"] },
                               "region" => { "valid-values" => ["us-ea
+st-1", "us-west-1"] },
                             },
    },
  },
]
[download]

UPDATE: cleaned up fixhash and added "incomplete parse" check.

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11147096
use warnings;

my $testdata = <<END;
sys ecm cloud-provider /Common/aws-ec2 {
    description "The aws-ec2 parameters"
    property-template {
        account { }
        availability-zone {
            valid-values { a b c d }
        }
        instance-type {
            valid-values { t2.micro t2.small t2.medium }
        }
        region {
            valid-values { us-east-1 us-west-1 }
        }
    }
}
END

sub expr
  {
  /\G\s+/gc;
  my $e = [];
  $e =
    /\G\s+/gc ? $e :
    /\G([^{}\s]+) "(.*?)"/gc ? [ @$e, { $1 => $2 } ] :
    /\G([^{}\s]+) \{/gc ?  [ @$e, { "$1" => (expr(),
        /\G\}/gc || die pos($_), ' missing }', substr $_, pos($_))[0] 
+} ] :
    /\G([^{}\s]+)/gc ? [ @$e, $1 ] :
    return $e while 1;
  }

use List::AllUtils qw( all );
sub fixhash
  {
  local $_ = shift;
  return ref $_ eq 'ARRAY'
    ? ( all { ref $_ eq 'HASH' } @$_ )
      ? { map fixhash($_), map %$_, @$_ }
      : [ map fixhash($_), @$_ ]
    : ref $_ eq 'HASH'
      ? { map fixhash($_), %$_ }
      : $_;
  }

local $_ = $testdata; # NOTE expr() expects input in $_
my $parse = fixhash expr();
pos($_) < length $_ and die "incomplete parse ", substr $_, pos($_);
use Data::Dump 'dd'; dd $parse;
[download]

SECOND UPDATE: eliminating fixhash() by building it into expr()

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11147096
use warnings;
use List::AllUtils qw( all );

my $testdata = <<END;
sys ecm cloud-provider /Common/aws-ec2 {
    description "The aws-ec2 parameters"
    property-template {
        account { }
        availability-zone {
            valid-values { a b c d }
        }
        instance-type {
            valid-values { t2.micro t2.small t2.medium }
        }
        region {
            valid-values { us-east-1 us-west-1 }
        }
    }
}
END

sub expr
  {
  /\G\s+/gc;
  my $e = [];
  $e =
    /\G\s+/gc ? $e :
    /\G([^{}\s]+) "(.*?)"/gc ? [ @$e, { $1 => $2 } ] :
    /\G([^{}\s]+) \{/gc ?  [ @$e, { "$1" => (expr(),
      /\G\}/gc || die pos($_), ' missing } ', substr $_, pos($_))[0] }
+ ] :
    /\G([^{}\s]+)/gc ? [ @$e, $1 ] :
    return ref $e eq 'ARRAY' && ( all { ref $_ eq 'HASH' } @$e ) ?
      { map %$_, @$e } : $e while 1
  }

local $_ = $testdata; # NOTE expr() expects input in $_
my $parse = expr();
pos($_) < length $_ and die "incomplete parse ", substr $_, pos($_);
$Data::Dump::LINEWIDTH = 26;
use Data::Dump 'dd'; dd $parse;
[download]

Outputs:

[
  "sys",
  "ecm",
  "cloud-provider",
  {
    "/Common/aws-ec2" => {
      "description"       => "The aws-ec2 parameters",
      "property-template" => {
                               "account" => {},
                               "availability-zone" => {
                                 "valid-values" => ["a" .. "d"],
                               },
                               "instance-type" => {
                                 "valid-values" => [
                                   "t2.micro",
                                   "t2.small",
                                   "t2.medium",
                                 ],
                               },
                               "region" => {
                                 "valid-values" => ["us-east-1", "us-w
+est-1"],
                               },
                             },
    },
  },
]
[download]

THIRD UPDATE: factoring out a regex and shifting things around a little, maybe making things slightly clearer.

#!/usr/bin/perl

use strict; # https://perlmonks.org/?node_id=11147096
use warnings;

sub expr
  {
  my $val = [];
  $val =
    /\G\s+/gc ? $val :
    /\G([^{}\s]+)/gc ? do { my $key = $1; [ @$val, /\G \{/gc
      ? { $key => (expr(), /\G\}/gc || die ' missing } ')[0] }
      : /\G "(.*?)"/gc ? { $key => $1 } : $key ] } :
    return ref $val eq 'ARRAY' && ( @$val == grep ref $_ eq 'HASH', @$
+val )
      ? { map %$_, @$val }
      : $val while 1
  }

sub parse
  {
  local $_ = join '', @_;
  my $parse = expr;
  pos($_) == length $_ or die "incomplete parse stopped at ", substr $
+_, pos($_);
  return $parse;
  }

my $parse = parse( <DATA> );
$Data::Dump::LINEWIDTH = 26;
use Data::Dump 'dd'; dd $parse;

__DATA__
sys ecm cloud-provider /Common/aws-ec2 {
    description "The aws-ec2 parameters"
    property-template {
        account { }
        availability-zone {
            valid-values { a b c d }
        }
        instance-type {
            valid-values { t2.micro t2.small t2.medium }
        }
        region {
            valid-values { us-east-1 us-west-1 }
        }
    }
}
[download]

Comment on Re: Parsing bracket formatted file (third update) Select or Download Code


Perl: the Markov chain saw
	PerlMonks