source: trunk/gsdl3/src/java/org/greenstone/gsdl3/gs3build/CollectionManager.java@ 6454

package org.greenstone.gsdl3.gs3build;

import java.util.Date;
import java.util.Calendar;
import java.util.List;
import java.util.ArrayList;
import java.util.Map;
import java.util.HashMap;
import java.util.Iterator;
import java.util.GregorianCalendar;

import java.io.File;
import java.io.IOException;

import java.net.URL;

import javax.xml.parsers.*;

import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.NamedNodeMap;
import org.w3c.dom.Node;
import org.w3c.dom.NodeList;
import org.w3c.dom.Text;

import org.xml.sax.SAXException;
import org.xml.sax.SAXParseException;

import org.greenstone.gsdl3.gs3build.collection.*;
import org.greenstone.gsdl3.gs3build.classifier.*;
import org.greenstone.gsdl3.gs3build.indexers.*;

import org.greenstone.gsdl3.gs3build.util.GS3SQLConnection;
import org.greenstone.gsdl3.gs3build.util.GS3SQLConnectionFactory;
import org.greenstone.gsdl3.gs3build.util.DOMUtils;

/**
 *  Store and hold collection-level configuration information for a collection.
 *  This should be used by BuildManager to work out which classes, etc. to load
 *  at build time, and as a repository for the collection-level metadata, and
 *  a means of loading and saving the same to a file or database, as is seen
 *  fit in the final development of gs3.
 */

public class CollectionManager
{
  GregorianCalendar  lastBuildDate;  // pretty obvious
  String             adminEmail;     // the email address of the administrator of the
                             // collection
  int                buildDocNo;     // used to generate document identifiers
  CollectionMetadata metadata;       // collection-level metadata
  GS3SQLConnection   database;       // the database to store everything in
  String             collectionHome;
  String             collectionName;

  class CollectionClassifier
  { File file;
    String type;
    List fields;
    String sort;

    public CollectionClassifier(String type, Node node)
    { this.type = type;
      this.fields = new ArrayList();

      NodeList children = node.getChildNodes();
      for (int c = 0; c < children.getLength(); c ++) {
    Node child = children.item(c);

    if (child.getNodeType() == org.w3c.dom.Node.ELEMENT_NODE) {
      String name = child.getNodeName();
      
      if (name.equals("file")) {
        NamedNodeMap atts = children.item(c).getAttributes();
        Node attribute = atts.getNamedItem("URL");
        String urlString = attribute.getNodeValue();
        if (urlString == null)
          continue;

        this.file = new File(urlString);
      }
      else if (name.equals("field")) {
        String fieldName = DOMUtils.getNodeChildText(children.item(c));
        this.fields.add(fieldName.toString());
      }
      else if (name.equals("sort")) {
        String sortName = DOMUtils.getNodeChildText(children.item(c));
        this.sort = sortName;
      }
    }
      }
    }

    public ClassifierInterface getClassifier()
    {
      if (this.type.toLowerCase().equals("hierarchy")) {
    return new HierarchyClassifier(this.file, this.fields, this.sort);
      }

      return null;
    }
  }
   
  class CollectionIndexer
  { String level;
    String type;
    String name;
    List fields;
    String sort;

    public CollectionIndexer(String type, String indexName, Node node)
    { this.type = type.toLowerCase();
      this.fields = new ArrayList();
      this.name = indexName;
      this.level = IndexerManager.DEFAULT_LEVEL;
    
      NodeList children = node.getChildNodes();
      for (int c = 0; c < children.getLength(); c ++) {
    Node child = children.item(c);
    
    if (child.getNodeType() == org.w3c.dom.Node.ELEMENT_NODE) {
      String name = child.getNodeName();
      
      if (name.equals("level")) {
        String levelName = DOMUtils.getNodeChildText(children.item(c));
        this.level = levelName;
      }
      else if (name.equals("field")) {
        String fieldName = DOMUtils.getNodeChildText(children.item(c));
        this.fields.add(fieldName.toString());
      }
    }
      }
      if (this.fields.size() == 0) {
    this.fields.add(IndexerManager.DEFAULT_FIELD);
      }
    }

    public void prepareIndexer(Map indexerMap)
    { IndexerInterface indexer = null;

      if (!indexerMap.containsKey(this.type)) {
    if (this.type.toLowerCase().equals("mg")) {
      indexer = new MGIndexer();
    }
    if (indexer != null) {
      indexerMap.put(this.type, indexer);
    }
      }

      indexer = (IndexerInterface) indexerMap.get(this.type);
      if (indexer != null) {
    indexer.addIndex(this.level, this.fields.get(0).toString());
      }
    }
  }

  /**
   *  Create the collection manager for a given collection
   *
   *  @param <code>String</code> the name of the collection
   */
  public CollectionManager(String collection)
  { String collectRoot = System.getProperty("GSDL3HOME");
  
    this.database = GS3SQLConnectionFactory.createConnection(collection);
    if (this.database != null) {
      this.database.clearCollection(collection);
      this.database = null;
    }
    if (this.database == null) {
      this.database = GS3SQLConnectionFactory.createConnection("test");
      this.database.initCollection(collection);
    }

    this.metadata = new CollectionMetadata();

    if (collectRoot == null)
    { System.out.println("Unable to locate GSDL3HOME");
      //      System.exit(1);
      return;
    }

    if (collectRoot.endsWith(System.getProperty("file.separator"))) 
    { this.collectionHome = collectRoot + "web/sites/localsite/collect" + System.getProperty("file.separator") + collection;
    }
    else
    { this.collectionHome = collectRoot + System.getProperty("file.separator") + "web/sites/localsite/collect" + System.getProperty("file.separator") + collection;
    }
    this.collectionName = collection;

    File buildDirectory = new File(this.collectionHome, "building");
    if (!buildDirectory.exists()) {
      buildDirectory.mkdir();
    }

    File archiveDirectory = new File(this.collectionHome, "archives");
    if (!archiveDirectory.exists()) {
      archiveDirectory.mkdir();
    }

    this.buildDocNo = 1;
  }

  public void configureCollection(BuildManager buildManager)
  { File collectionConfig = new File(collectionHome, "/etc/collectionConfig.xml");
    
    // get the File and read it in
    try
    {
      DocumentBuilderFactory factory = DocumentBuilderFactory.newInstance();
      DocumentBuilder builder = factory.newDocumentBuilder();
      Document document = builder.parse(collectionConfig);

      CollectionClassifier classifier = null;

      Map                  indexerMap = new HashMap();

      // TODO: report an error
      if (document == null)
      { 
      }

      // now parse the manager file...
      Element rootElement = document.getDocumentElement();

      if (rootElement.getTagName() != "collectionConfig")
      { // TODO: throw exception
      }

      NodeList children = rootElement.getChildNodes();
      for (int c = 0; c < children.getLength(); c ++)
      { // assume that non-element children are irrelevant
    if (children.item(c).getNodeType() != org.w3c.dom.Node.ELEMENT_NODE)
    { continue;
    }

    String name = children.item(c).getNodeName();

    // the name is a plugin element
    if (name.equals("index"))
    { NamedNodeMap atts = children.item(c).getAttributes();
      Node attribute = atts.getNamedItem("type");
      String type = attribute.getNodeValue();

      attribute = atts.getNamedItem("name");
      String indexName = attribute.getNodeValue();

      // create the local indexer representation
      CollectionIndexer indexer = new CollectionIndexer(type, indexName, children.item(c));

      // prepare it...
      indexer.prepareIndexer(indexerMap);
    }
    else if (name.equals("browse"))
    {
    }
    else if (name.equals("classifier"))
    { 
      NamedNodeMap atts = children.item(c).getAttributes();
      Node attribute = atts.getNamedItem("type");
      String type = attribute.getNodeValue();
      classifier = new CollectionClassifier(type, children.item(c));

      // attach the classifier
      ClassifierInterface classify = classifier.getClassifier();
      buildManager.getClassifierManager().addClassifier(classify);
    }
    // TODO: other elements - make a factory-method approach here...
    else
    {
    }
      }

      Iterator indexers = indexerMap.values().iterator();
      while (indexers.hasNext()) {
    IndexerInterface indexer = (IndexerInterface) indexers.next();

    // attach the classifier
    buildManager.getIndexerManager().addIndexer(indexer);   
      }
    }
    catch (FactoryConfigurationError e) {
      System.out.println(e);
    }
    catch (ParserConfigurationException ex) {
      System.out.println(ex);
    }
    catch (SAXException ex) {
      System.out.println(ex);
    }
    catch (IOException ex)
    {
      System.out.println(ex);
    }

    System.out.println("<<<Obtaining database>>>>");
  }

  public String getEtcDirectory()
  { return this.collectionHome + File.separator + "etc";
  }

  public String getImportDirectory()
  { return this.collectionHome + File.separator + "import";
  }

  public String getBuildDirectory()
  { return this.collectionHome + File.separator + "building";
  }

  public String getArchiveDirectory()
  { return this.collectionHome + File.separator + "archives";
  }

  public GS3SQLConnection getDatabase()
  {
    return this.database;
  }

  public void startBuild()
  { GregorianCalendar today = new GregorianCalendar();
  
    if (this.lastBuildDate != null)
    { // if the build date is different to the last build date, then reset the build
      // document number
      if (today.get(Calendar.YEAR) != this.lastBuildDate.get(Calendar.YEAR) ||
      today.get(Calendar.MONTH) != this.lastBuildDate.get(Calendar.MONTH) ||
      today.get(Calendar.DAY_OF_MONTH) != this.lastBuildDate.get(Calendar.DAY_OF_MONTH))
      { this.buildDocNo = 1;
      }
    }
    this.lastBuildDate = today;
  }

  public void endBuild()
  {
    Date startDate = this.lastBuildDate.getTime();
    Date date = new Date();

    long startTime = startDate.getTime();
    long endTime   = date.getTime();

    long difference = ((endTime - startTime) + 500) / 1000;

    System.out.println("Build completed");
    System.out.println("---------------");
    System.out.println("Total Documents: " + this.getCollectionMetadata("gsdl3", "documentCount"));
    System.out.println("Total Time :     " + (difference / 60) + " min. " + (difference % 60) + " secs.");
  }

  public String getNextDocumentID()
  { StringBuffer ID = new StringBuffer();

    int value;
    ID.append(lastBuildDate.get(Calendar.YEAR));

    // the use of month is a little odd, hence the following
    // code.  Calendar.MONTH yields 0 = January, 1 = February,
    // etc. hence there is a '+1' added to the month to make 
    // it into January = 1, etc., and the padding is altered
    // correspondingly.
    value = lastBuildDate.get(Calendar.MONTH);
    if (value < 9)
    { ID.append("0");
    }
    ID.append(value + 1);
    value = lastBuildDate.get(Calendar.DAY_OF_MONTH);
    if (value < 10)
      ID.append("0");
    ID.append(value);


    value = this.buildDocNo;
    this.buildDocNo ++;
    
    ID.append(":");
    ID.append(Integer.toString(value));
    return ID.toString();
  }
  
  public int getDocumentNumber()
  { this.buildDocNo ++;
    return this.buildDocNo - 1;
  }

  /**
   *  Get the collection metadata item in the given namespace
   *
   *  @param <code>String</code> the namespace
   *  @param <code>String</code> the label of the metadata
   */
  public String getCollectionMetadata(String namespace, String label)
  { return this.metadata.getCollectionMetadata(namespace, label).get(0).toString();
  }

  /**
   *  Set the collection metadata item in the given namespace
   *
   *  @param <code>String</code> the namespace
   *  @param <code>String</code> the label
   *  @param <code>String</code> the value
   */
  public void setCollectionMetadata(String namespace, String label, String value)
  { this.metadata.setCollectionMetadata(namespace, label, value);
  }
}