A simple test grammar for CSV files.

Special thanks to @zoffixznet who took time and pointed me in the right direction on IRC.

csv_grammar.pl6

@@ -0,0 +1,146 @@
+#!/usr/bin/env perl6
+
+#use Grammar::Debugger;
+
+use v6;
+
+grammar CSV::Grammar {
+   token TOP {
+      <record>* % (\n|\r\n)
+   }
+   token record {
+      <field>* % ','
+   }
+   token field {
+      <quoted> | <non_quoted> | <empty>
+   }
+   token quoted {
+      \" (<non_quoted>|<comma>|<newline>) * \"
+   }
+   token non_quoted {
+      ( $<chars>=[ <-[ " , \n ]>+ ] | <double_dquote> ) ** 1..*
+   }
+   token double_dquote {
+      \" ** 2
+   }
+   token comma {
+      ','
+   }
+   token empty {
+      ''
+   }
+   token newline {
+      \n
+   }
+}
+
+class CSV::Record {
+   has @.fields;
+}
+
+class CSV {
+   has CSV::Record @.records;
+
+   method add ( CSV::Record $record is required){
+      @.records.push($record);
+   }
+}
+
+class CSV::Action {
+
+   has CSV $.file = CSV.new();
+
+   method TOP($/) {
+      #say $/<record>.[0].perl;
+      for $/<record> -> $record {
+         #say "  Record: {$record.perl}";
+         self.file.add($record.made);
+      }
+      # $/<record>.map: {
+      #    say "  Record: ", $_.perl;
+      #    $.file.add($_.made)
+      # }
+   }
+   method record($/) {
+      #say "Record: {$/.perl}";
+      my @fields = $/<field>.map: {
+         #say $_.perl;
+         #say "Field: ",$_.made;
+         $_.made;
+      }
+      #@fields.unshift($<field>.made);
+
+      #say @fields.perl;
+      my $csv = CSV::Record.new( fields => @fields );
+      $/.make: $csv;
+      #say $csv.perl;
+   }
+   method field($/) {
+      my $value = $<quoted>.defined ?? $<quoted>.made !!
+         $<non_quoted>.defined ?? $<non_quoted>.made !!
+            $<empty>.made;
+      #say "Value: $value";
+      $/.make: $value;
+   }
+   method quoted($/) {
+      $/.make: join '', $/[0].map: {
+         $_<non_quoted>.defined ?? $_<non_quoted>.made !!
+            $_<comma>.defined ?? $_<comma>.made !!
+               $_<newline>.made
+
+      }
+   }
+   method non_quoted($/) {
+      my $chars = join '', $/[0].map: {
+         $_<chars>.defined ?? ~$_<chars> !! $_<double_dquote>.made
+      }
+      #say $chars;
+      $/.make: $chars;
+      #say $/.made
+   }
+   method double_dquote($/) {
+      $/.make: '"';
+   }
+   method comma($/) {
+      $/.make: ','
+   }
+   method empty($/) {
+      $/.make: ''
+   }
+   method newline($/) {
+      $/.make: "\n"
+   }
+}
+
+my @DATA = (
+    'a,b,c',
+    'This is a test, Only a test, part 2',
+    '"Impossible Dreams","Impossible Sunsets",Pair trees',
+    '"Impossible Dreams","Impossible ""Sunsets""",Pair "" trees',
+    ',,,',
+    'a,b,c
+    d,e,f
+    g,h,i',
+   'a,b,"c
+   d",e,f',
+);
+
+for @DATA -> $line {
+
+   say "Line: $line";
+   my $action = CSV::Action.new();
+
+   my $match_obj = CSV::Grammar.parse($line, :actions($action));
+   #say "\nReturned: ",$match_obj.perl;
+   my $csv = $action.file;
+
+   #say $csv.perl;
+
+   for $csv.records -> $record {
+      say "  Record: {$record.gist}";
+      for $record.fields -> $field {
+         say "    Field: $field";
+      }
+   }
+
+}