file_to_parse.txt

2014-03-20

Line 1
Line 2

2014-03-21

Line 1
Line 2

output.rb

# What I want:

  [{:body=>"Line 1\n\nLine 2", :year=>"2014", :month=>"03", :day=>"20"},
   {:body=>"Line 1\n\nLine 2", :year=>"2014", :month=>"03", :day=>"21"}]

# What I'm getting
  [{:year=>"2014"@0,
    :month=>"03"@5,
    :day=>"20"@8,
    :body=>"Line 1\n\nLine 2\n\n2014-03-21\n\nLine 1\n\nLine 2"@12}]

parser.rb

class Parser < Parslet::Parser
  rule(:integer) { match('[0-9]').repeat(1) }
  rule(:hyphen) { str('-') }
  
  rule(:date)  { integer.repeat(1,4).as(:year) >>
                 hyphen >>
                 integer.repeat(1,2).as(:month) >>
                 hyphen >>
                 integer.repeat(1,2).as(:day) >>
                 str("\n\n") }
  rule(:body) { (any.repeat >> (date.present? | any.absent?)).as(:body) }
  rule(:entry) { date >> body }
  rule(:entries) { entry.repeat }
  
  root(:entries)
end

I believe the problem is that my :body rule is too greedy, but can't figure out how to remedy that.

@r00k Sorry to jump in, mid-conversation, but I'm confused as to why :body has extra \n's in both your desired results and actual results -- shouldn't there be just one, between Line 1 and Line 2? It looks like file_to_parse.txt has one newline between Line 1 and Line 2.

Also, I hope you're coding outside, or at least with the window open. It's amazing, outside, today. 😄

Not familiar with this parser, but I think you want something besides any.

  rule(:body) { (date.absent? >> any).repeat.as(:body) }