SimpleOsdParser.scala

import java.io.InputStream
import java.util.zip.ZipInputStream

import scala.io.Source
import scala.io.Codec
import scala.xml.XML

object SimpleOsdParser {

  type Table = Map[String, Seq[Seq[String]]]

  def parse(bytes: InputStream): Either[String, Table] = {
    val content = {
      def rec(zip: ZipInputStream): Option[String] = {
        Option(zip.getNextEntry) match {
          case Some(entry) if entry.getName == "content.xml" ⇒
            import UnicodeBOMInputStream.BOM
            val withBom = new UnicodeBOMInputStream(zip)
            withBom.getBOM match {
              case BOM.UTF_8 | BOM.NONE ⇒
                val stream = withBom.skipBOM()
                Some(Source.fromInputStream(stream)(Codec.UTF8).mkString)
              case _ ⇒ throw new Exception("Unsupported encoding")
            }
          case Some(entry) ⇒
            rec(zip)
          case None ⇒ None
        }
      }
      val zipStream = new ZipInputStream(bytes)
      rec(zipStream)
    }
    content match {
      case Some(xmlString) ⇒
        val xml = XML.loadString(xmlString)
        val tables = xml \ "body" \ "spreadsheet" \ "table"
        val pairs = tables map { table ⇒
          val name = table \@ "{urn:oasis:names:tc:opendocument:xmlns:table:1.0}name"
          val data = table \ "table-row" map { row ⇒
            val cells = row \ "table-cell" map { cell ⇒
              cell.text
            }
            cells.reverse.filter(_.nonEmpty)
          }
          val res = data.reverse.dropWhile(_.isEmpty).reverse
          (name, res)
        }
        Right(pairs.toMap)
      case None ⇒ Left("Invalid ODS file: content.xml not found")
    }
  }
}