Quiz

I'm following up on a post that was done in 2014 (to which I am unable to reply, perhaps because it's too old?). 

GIFT (or something like it) is a great format because it requires much less clicking to create a lot of questions for quizzes. You can save GIFT in your own text file, in an email, etc. It's an example of a domain-specific language; they are developed to aid in crafting solutions to problems that are specific to certain domains. Creating quiz questions is one such domain. 

I recently used PEG.js in a project to create a parser for a domain-specific language. It could be useful to consider PEG.js to parse GIFT format. 

The really cool thing about PEG.js is there's a web page that makes grammar development really easy. It's not an IDE, so you have to save your results in a real text file from time to time. You can use sample input, and see the error messages that come from your grammar when it doesn't parse properly.

Here's a start on a grammar that recognizes trivial T/F and MC questions (the input file follows):

Expression
  = Spacing Rule+ EndOfFile

Rule
  = title:QuestionTitle? stem:QuestionStem questionType:SpecificQuestion Space* EndOfLine* 
  { return questionType + ' question:\n ' + title + "\n" + stem }
   / Comment EndOfLine*

SpecificQuestion "(specific question)"
  = TrueFalseQuestion / MCQuestion
 
TrueFalseQuestion "True/False Question"
  = '{' isTrue:('T'/'F') '}' { return isTrue; }

MCQuestion "Multiple-choice Question"
  = '{' EndOfLine choices:(Choice)+ '}' { return "Multiple-choice"; }

Choice "Choice"
  = CorrectChoice / IncorrectChoice
  
CorrectChoice "(correct choice)"
  = '=' correctChoice:RichText EndOfLine { return 'correct choice: ' + correctChoice;}

IncorrectChoice "(correct choice)"
  = '~' incorrectChoice:RichText EndOfLine { return 'incorrect choice: ' + incorrectChoice;}

QuestionTitle
  = '::' title:Title '::' { return ' title: ' + title }
  
QuestionStem
  = stem:RichText { return ' stem: ' + stem; }

Text "(text)"
  = [ A-Za-z0-9.+]

RichText
  = Text* { return 'RichText:' + text() } 

Title
  = Text* { return text() }

_ "whitespace"
  = [ \t]*
  
Spacing 
  = (Space / Comment)*
Comment "(comment)"
  = '//' (!EndOfLine .)* EndOfLine { return 'comment';}
Space "(space)"
  = ' ' / '\t'
EndOfLine "(end of line)"
  = '\r\n' / '\n' / '\r'
EndOfFile 
  = !. { return "EOF"; }

Here's a sample input that parses:

// comment at top
::True statement about Grant::Grant was buried in a tomb in New York City.{T}


::False statement about Grant::Grant was never buried.{F}

// comment in middle

::Math question:: What is 1+2 {
=3
~4
~2
}

I'm pretty sure my grammar above is not perfect - it needs more testing. For example, it doesn't support feedback on answers (e.g. the '#' option). The goal is to show what can be done with PEG and why a grammar for GIFT would be great.