#!/usr/bin/env groovy

// Usage: ./extractMetadata.groovy [batch directory]

import groovy.xml.DOMBuilder

import java.io.File
import java.io.FileOutputStream

import javax.xml.parsers.DocumentBuilder
import javax.xml.parsers.DocumentBuilderFactory
import javax.xml.transform.sax.SAXSource

import net.sf.saxon.xpath.XPathEvaluator
import net.sf.saxon.xpath.XPathExpression

import org.apache.xml.serialize.OutputFormat
import org.apache.xml.serialize.XMLSerializer
import org.xml.sax.InputSource

// Set up the dom builder factory for later
docBuilderFactory = DocumentBuilderFactory.newInstance()
docBuilderFactory.setNamespaceAware(false)

// The main logic to handle an item directory
// It's most convenient to implement this as a closure
mainAction = { | itemDir | 
  itemDir.eachFile({ | f |
     // This does a regexp match
     if(f.name ==~ "nsc.*_original\.cml") {
	print(".")
 	xpe = new XPathEvaluator(new SAXSource(new InputSource(f.toURL().toString())))
	ichi = xpe.createExpression("//entry/identifier[@version='0.932Beta' and @tautomeric='0']/basic/text()")
 	nsc = xpe.createExpression("//entry/molecule/@id")
	name = xpe.createExpression("//entry/name/text()")

	id = nsc.evaluate().get(0).getValue()
	docBuilder = docBuilderFactory.newDocumentBuilder()
        builder = DOMBuilder.newInstance()

        // Inline method definitions on the builder 'dc' result in 
 	// xml elements and attributes being created.
        dc = builder.dublin_core() {
           dcvalue(element:"identifier", qualifier:"none", id)
	   dcvalue(element:"identifier", qualifier:"ichi", ichi.evaluate().get(0).getValue())
	   dcvalue(element:"title", qualifier:"none", id)
	   dcvalue(element:"creator", qualifier:"none", "US National Cancer Institute")
           dcvalue(element:"publisher", qualifier:"none", "Unilever Center for Molecular Informatics, Cambridge University")
	   dcvalue(element:"date", qualifier:"created", "2003-02-01")
	   dcvalue(element:"type", qualifier:"none", "Other")
	   for(nm in name.evaluate()) {
		dcvalue(element:"title", qualifier:"alternative", nm.getValue())
           }
        }

        fos = new FileOutputStream(itemDir.toString() +"/dublin_core.xml")
	fileserializer = new XMLSerializer(fos, new OutputFormat())
	fileserializer.serialize(dc)
	fos.close()
     }
  })
}

println("--- Extracting metadata for ${args[0]} ---")
new File("${args[0]}").eachFile(mainAction);

println("")