Hi Lukas,
Thanks for the detailed analysis and insight. It was very helpful.
I have rewritten the grammar based on your inputs. And ran into same issue.
After some thinking I identified the issue. My grammar accepts and parses
only one statement!
I modified grammar in the following manner. Agreed that this is not the
best approach. Still for a prototype, it resolved my original issue.
language ::= if_statements action_statements| action_statements
Modified code now is:
===========================================================
use strict;
use warnings;
use warnings qw(FATAL utf8); # Fatalize encoding glitches.
use feature 'say' ;
#use Data::Dumper;
use Data::Dumper::Concise;
#use Log::Handler;
use Marpa::R2;
use PyGen;
#use Moo;
my @flt;
sub flattenArray {
foreach (@_) {
if (ref $_ eq "ARRAY") {
flattenArray(@$_);
}
else {
push @flt,$_;
}
}
}
sub PYGEN_Action::ifAction {
my (undef,@a,@b) = @_;
print Dumper(@a);
print "\n";
}
sub PYGEN_Action::verbAction {
print "verbAction inputs\n";
print Dumper(@_);
}
sub PYGEN_Action::listAction {
#unpack lol. output is global @flt;
flattenArray(@_);
my $pygen = PyGen->new();
shift @flt; #discard empty hash.
$pygen->write_list(\@flt);
}
my $g = Marpa::R2::Scanless::G->new({
default_action => '::array',
source => \(<<'END_OF_SOURCE'),
lexeme default = latm => 1
:start ::= language
# include begin
language ::= if_statements action_statements| action_statements
if_statements ::= if_statement*
if_statement ::= 'If' condition_statement action_statements 'End'
action=>ifAction
condition_statement ::= variable operator value
| variable operator variable
action_statements ::= action_statement*
action_statement ::= action_verb variable action => verbAction
| variable ('List') ('is') comma_list action =>
listAction
comma_list ::= variable
| comma_list (',') variable
:discard ~ ws
ws ~ [\s]+
# include end
action_verb ~ 'remove' | 'Create' | 'Get'| 'Set' | 'CREATE' |'Revise'
variable ~ [\w]+
operator ~ 'has' | '>' | '<' | '='
value ~ 'duplicate'
#str ~ [\w]+
END_OF_SOURCE
});
my $re = Marpa::R2::Scanless::R->new({ grammar => $g });
my $input = <<'INPUT';
If DocumentTitle = A222
Revise Document
Create Author
End
Revise Author
Author List is Wodehouse, Shelley, Hardy, Murakami
INPUT
print "Trying to parse:\n$input\n\n";
print "\n", $re->show_progress(0, -1);
$g->parse(\$input, 'PYGEN_Action');
==========================================================
Output comes includes the List setting.
Output
======
verbAction inputs
{}
"Revise"
"Document"
verbAction inputs
{}
"Create"
"Author"
"If"
[
"DocumentTitle",
"=",
"A222",
]
[
1,
1,
]
"End"
verbAction inputs
{}
"Revise"
"Author"
AuthorList=['Wodehouse','Shelley','Hardy','Murakami',]
On Saturday, April 30, 2022 at 8:45:37 AM UTC-7 Jeffrey Kegler wrote:
>
>
> Good answer, Lukas, thanks! -- jeffrey
>
>
> Sent with ProtonMail secure email.
> ------- Original Message -------
> On Saturday, April 30th, 2022 at 4:27 AM, Lukas Atkinson
> <[email protected]> wrote:
>
>
> > Your grammar does not allow the list in that location. Here is the
> > relevant excerpt:
> >
> > ```
> > :start ::= language
> > #
> > # include begin
> > language ::= if_statement action_statements action => ifAction
> >
> > | action_statements action => verbAction
> >
> > | ('List') variable ('is') comma_list action =>
> >
> > listAction
> > `So you say that a document can EITHER be an if-statement followed by
> any number of action statements, OR any number of action statements, OR a
> List statement. This is why testing a List statement by itself works fine.
> But in your input, you have an if-statement that includes one action
> statement, followed by an unexpected List statement:`
> > If documentTitle = A100
> > remove Document
> > List Author is Dante, Murakami, Ash, Harris
> > `Instead, you would likely want to change your grammar so that you can
> have multiple statements of different type, for example:`
> > language ::= statement*
> >
> > statement ::= if_statement | action | list_statement
> > `But this might become ambiguous. For example, consider this input:`
> > If foo = bar
> > remove Doc1
> > remove Doc2
> > ``Should the `remove Doc2` action be within the if-statement, or should
> it be an unconditional action? In your current grammar, both actions would
> be part of the if-statement. In my suggested grammar, both are allowed and
> the parse is ambiguous. This problem is solved most easily if you can
> change your language to have explicit delimiters, e.g. curly braces:``
> > If foo = bar {
> > remove Doc1
> > }
> > remove Doc2
> > ``or other compound statement terminators, e.g. an `End` like in Ruby or
> Lua:``
> > If foo = bar
> > remove Doc1
> > End
> > remove Doc2
> > `Of course, many languages use the indentation level to indicate
> nesting, for example Python or YAML. Marpa has no built-in features for
> indentation-sensitive parsing. If you are trying to parse an existing
> format or really want to use indentation to indicate the scope of the
> conditional, you would have to use Marpa's event system to handle
> indentation. You could use the same trick that Python uses: having
> special/invisible INDENT and DEDENT tokens that mark where the indented
> block begins and ends. These tokens are emitted whenever the indentation
> changes. In the grammar, these tokens can then be used just like curly
> braces in curly-brace languages. For example, you might then define your
> if-statement like:`
> > if_statement ::= ('If') condition (INDENT) action_statements (DEDENT)
> > ```
> >
> > --
> > You received this message because you are subscribed to the Google
> Groups "marpa parser" group.
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to [email protected].
> > To view this discussion on the web visit
> https://groups.google.com/d/msgid/marpa-parser/71c6eb23-b33d-7003-1342-5781e1dc0dc5%40LukasAtkinson.de
> .
>
--
You received this message because you are subscribed to the Google Groups
"marpa parser" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/marpa-parser/a8889a68-fe18-4f0a-a21f-54bcb69c6025n%40googlegroups.com.
