QtXML DOM parsing / iTunes Library

Question

I'm trying to get a list of the iTunes albums, by parsing the XML library (iTunes Music Library.xml in the iTunes directory).

#include <iostream>
#include <QtCore>
#include <QFile>
#include <QtXml>

using namespace std;

void parse(QDomNode n) {

    while(!n.isNull()) {

        // If the node has children
        if(n.hasChildNodes() && !n.isNull()) {

            // We get the children
            QDomNodeList nChildren = n.childNodes();

            // We print the current tag name
            //std::cout << "[~] Current tag : <" << qPrintable(n.toElement().tagName()) << ">" << std::endl;

            // And for each sub-tag of the current tag
            for(int i = 0; i < nChildren.count(); i++) {

                // We get the children node
                QDomNode nChild = nChildren.at(i);
                // And the tag value (we're looking for *Album* here)
                QString tagValue = nChild.toElement().text();

                // If the tag isn't null and contain *Album*
                if(!nChild.isNull() && tagValue == "Album") {
                    // The album name is in the next tag
                    QDomElement albumNode = nChild.nextSiblingElement();
                    std::cout << "[-] Album found -> " << qPrintable(albumNode.text()) << std::endl;
                }

                // And we parse the children node
                parse(nChild);
            }
        }

        n = n.nextSibling();
    }
}

int main() {

    QDomDocument doc("Lib");
    QFile file("/Users/wizardman/QtRFIDMusic/Lib.min.xml");
    if(!file.open(QIODevice::ReadOnly))
        return 1;
    if(!doc.setContent(&file)) {
        file.close();
        return 1;
    }
    file.close();

    // Root element
    QDomElement docElem = doc.documentElement();

    // <plist> -> <dict>
    QDomNode n = docElem.firstChild().firstChild();

    cout << endl << "Album list" << endl;
    cout << "------------------------------------" << endl;


    parse(n);

    return 0;
}

The iTunes' XML is not really standart XML, the name of the album is stored in the node next to the <key>Album</key> for each entry. Here is what it looks like. I intentionnaly renamed some nodes for debugging purpose (to see if I reach them in my output).

And here is my output :

Album list
------------------------------------
[-] Album found -> J Dilla - Legacy Vol.1
[-] Album found -> J Dilla - Legacy Vol.2
[-] Album found -> J Dilla - Legacy Vol.1
[-] Album found -> J Dilla - Legacy Vol.2
[-] Album found -> J Dilla - Legacy Vol.2
[-] Album found -> J Dilla - Legacy Vol.2

I can't see why the loop is reparsing the first nodes. Any ideas ?

Answer 1

After running your code under my debugger... it appears that you're iterating through the children too many times. Meaning, you recursively traverse through the entire tree (repeatedly) at <dict>, the inner <dict>, <dict_FOCUS> and also <dict_FOCUS2>.

For me, it was easier just to iterate (without recursion) through the nodes using QDomNode::firstChildElement(QString); I can't guarantee that this is bullet proof.. but it's a start! ;)

// Root element
QDomElement docElem = doc.documentElement();

// <plist> -> <dict>
QDomNode n = docElem.firstChildElement().firstChildElement("dict");

qDebug() << "Album list";
qDebug() << "------------------------------------";

QDomNodeList list = n.childNodes();
int count = list.count();

for(int i = 0; i < count; ++i)
{
  QDomElement node = list.at(i).toElement();
  if(node.tagName().startsWith("dict_FOCUS"))
  {
    node = node.firstChildElement();
    while(!node.isNull())
    {
      if(node.text() == "Album" && node.tagName() == "key")
      {
        node = node.nextSiblingElement();
        if(!node.isNull() && node.tagName() == "string")
        {
          qDebug() << "[-] Album found -> " << qPrintable(node.text());
        }
      }
      node = node.nextSiblingElement();
    }
  }
}