Reputation: 6763
Using Scala parser combinators library is it possible to parse by lines, not but characters? I would like my custom parsers to have signature
(in: Reader[String]) => ParserResult[T]
so that String in.first would be a line of text and in.rest - remaining lines.
I assume I need to subclass Parsers and define Elem type as String. Then implement Reader[String] to split input into lines. I wonder is it the right approach? Are there other solutions?
Upvotes: 0
Views: 806
Reputation: 6763
The following code does that:
import scala.util.parsing.combinator.Parsers
import scala.util.parsing.input.{Position, Reader}
object Main {
def main(args: Array[String]): Unit = CVParser("one\ntwo\nthree")
}
object CVParser {
def apply(in: String): Any = impl.parse(in)
private object impl extends Parsers {
override type Elem = String
def parse(in: String): ParseResult[String~String~String] = {
val reader = new ByLineReader(in.split(System.lineSeparator).toList, 1)
grammar(reader)
}
// private
private val grammar = SampleParser ~ SampleParser ~ SampleParser
private class ByLineReader(lines: List[String], lineNo: Int) extends Reader[String] {
override def first: String = lines.headOption getOrElse ""
override def rest: Reader[String] =
if (!lines.isEmpty) new ByLineReader(lines.tail, lineNo + 1)
else new ByLineReader(lines, lineNo)
override def pos: Position = new Position {
def line: Int = lineNo
def column: Int = 0
def lineContents: String = first
}
override def atEnd: Boolean = lines.size == 1
}
private object SampleParser extends Parser[String] {
def apply(in: Input): ParseResult[String] = {
println(in.first + "\n---------")
Success(in.first, in.rest)
}
}
}
}
Outputs:
one
---------
two
---------
three
---------
Now new parsers which accept a line of text may be written and combined using usual parser combinators, like ~, ~>, <~, |, rep, etc.
Upvotes: 0